vvEPA
           United States      Office of Water     EPA 440/5/84-001
           Environmental Protection  Regulations and Standards
           Agency        Washington, D.C. 20460
Water
Lake and Reservoir
Management
                                       •:• f
                                       /

-------
LAKE AND RESERVOIR
     MANAGEMENT
         Proceedings of the
     Third Annual Conference
    North American Lake
    Management Society
         October 18-20, 1983
         Knoxville, Tennessee
0.B, fcrlronnental r-- J  ' " •- ~:V3y'
Jl'-^lon 5, Library i
';.;;•;•,) 3. D&arborn St'«j- ,  . ^/U
Ohloago.,. JJu . 60604       s>
      U.S. Environmental Protection Agency
           Washington, D.C.

              1984

-------
              REVIEW NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policies of the Environmental Protection Agency,
nor does mention of trade names or commercial products
constitute endorsement or recommendations for use.
              EPA 440/5-84-001-
   U.S. Environmental Protection Agency
Office of Water Regulations and Standards
          Washington, D.C. 20460
                 ISSN 0743-8141

-------
FOREWORD
This publication represents the second joint venture, so to speak, of the U.S.
Environmental Protection Agency and the North American Lake Management
Society. This time, another Federal agency, the U.S. Department of
Agriculture, joins EPA in supporting the publication of this, the proceedings
of the International Symposium on Lake and Reservoir Management, held
Oct. 18-20, 1983, in Knoxville, Tenn.
   Once again, this proceedings represents  a partnership between the public
and the private sectors, in this case, a partnership designed to bring together
a variety of papers on lake management. More than 100 papers appear in this
volume: their subject matter ranges from a  section on  agricultural runoff to
one focusing on the role of local lake organizations. In these pages the
reader will find discussions of research  in macrophyte control, fishery
management, acid precipitation.
   Even a casual glance at the contents reveals that this proceedings reflects
most of the concerns of those researching  and working in lake management.
The theme is, of course, the distinct yet similar needs  and problems of
freshwater lakes and reservoirs, with a secondary emphasis on the effects of
nonpoint source pollution on water quality.
  This symposium—also the third annual meeting of the Society—attracted
nearly 600 people  to its sessions, representing the same variety of interests
indicated by the papers presented there. Those involved in lake management
are linked only by their concern for water quality. People who own homes on
lake shores attended those sessions in Knoxville, along with scientists,
teachers, lake  managers, and government people.
  Wise lake management has many facets and involves many people. It is
therefore fitting that this publication covers such a wide spectrum. The first
part of EPA's biennial report to Congress on the state  of this Nation's lake
resources (required under section 304(j) of the Clean Water Act), Lake and
Reservoir Management truly reflects the management  of this continent's
inland waters as it exists in 1983.
                                           Patrick Tobin
                                           Director
                                           Criteria & Standards Division

-------
CONTENTS
Foreword	
Conference Participants.
 IX
NONPOINT SOURCE POLLUTION IN LAKES
EPA's Emerging Nonpoint Source Role	
Rebecca W. Hanmer

USDA Programs and Nonpoint Source Pollution .
Richard D. Siegel

Politics of Acid  Rain	
A.O. Shingler

Rural America: Emphasis on Clean Water	
Willard (Bill) Phillips, Jr.
WATER QUALITY ASSESSMENT
Evaluating Reservoir Trophic Status:
The TVA Approach	
Janice Placke Cox

Wisconsin's Lakes: A Trophic Assessment	
Ronald H. Martin

Nutrients in Canal Flows to Lake Hefner, Oklahoma ...
Dale W. Toetz

Florida Lakes Assessment: Combining Macrophyte,
Chlorophyll, Nutrient, and Public Benefit
Parameters into a Meaningful Lake
Management Scheme	
H.L Edmiston, .V.B. Myers

Tools for Assessing Lake Eutrophication in the
Puget Sound Region, Washington	
Robert J. Gilliom
Surface Runoff Water Quality from Developed Areas
Surrounding a Recreational Lake	
Jay A. Bloomfield, James W. Sutherland,
James Swart, Clifford Siegfried
Computer Assisted Water Quality Data Analysis	
Michael W.  Mullen, Stephen R. Smith,
Richard E. Price, Terry S. Smith
Kentucky Reservoir Assessment of Water Quality and
Biological Conditions	
We/7 £ Carriker, Mahlon P. Taylor
Iron, Manganese, and Sulfide Transformations
Downstream from Normandy Dam	
John A. Gordon, W. Paul Bonner, Jack D.  Milligan
11
17
21
25
32
40
48
53
58
Application of Multispectral Digital Imagery to the
Assessment of Primary Productivity in
Flaming Gorge Reservoir	63
James Verdin, Sharon Campbell, David Wegner
AGRICULTURAL RUNOFF AND WATER QUALITY
Spatial and Seasonal Pattern of Nutrient Availability
in La Plata Lake, Puerto Rico	69
Jorge R. Garcia, Laurence J. Tilly

A Simulation Model for Assessing the Success of
Agricultural Best Management Practices on
Surface Water Quality	 77
James Madigan, Douglas Haith,
Scott O. Quinn, Jay Bloomfield

The Effectiveness of BMP's and Sediment Control
Structures and Their Relationship to
In-lake Water Quality	82
Forrest E. Payne,  Timothy M. Bjork
STATE PROGRAM DEVELOPMENT: PRIORITIES AND
STRATEGIES
Process to Identify, Screen, and Prioritize Rural
Water Resource and Lake Rehabilitation
Projects in Illinois	87
Thomas E. Davenport
Management Planning for 25 New Jersey Lakes	92
John Brzozowski, Stephen J. Souza
Incompatibility of Common  Lake Management
Objectives	  97
Kenneth J. Wagner, Ray T. Oglesby
The History of the Clean Lakes Program
in Tennessee	 101
Fred Van Atta,  Greg Den ton

Proposed Strategies for Management of a Tropical
Eutrophic Reservoir in Puerto Rico	 106
Laurence J. Tilly, Jorge R. Garcia
INTERNAL NUTRIENT CYCLING
Enhancement of Internal Cycling of Phosphorus by
Aquatic Macrophytes, with Implications
for Lake Management	  113
B.C. Moore, H.L Gibbons, W.H. Funk,
T. McKarns, J. Nyznyk, M.V. Gibbons

-------
 Reducing Sediment Phosphorus Release Rates in Long
 Lake Through the Use of Calcium Nitrate	 118
 Peter R.  Willenbring, Mark S. Miller,
 William D. Weidenbacher
 The Role of Internal Phosphorus Loading on the Trophic
 Status of New Jersey's Two Largest Lakes	 122
 Stephen J. Souza, John D. Koppen
 The Importance of  Sediment Release in the Assessment
 of a Shallow, Eutrophic Lake for Phosphorus Control... 129
 Patricia Mitchell
 Long Term Effect of Hypolimnetic Aeration of Lakes and
 Reservoirs, with Special Consideration of Drinking
 Water Quality and Preparation Costs	 134
 Bo Verner
BIOMANIPULATION
The Interactions Among Dissolved Organic Matter,
Bacteria, Suspended Sediments, and Zooplankton	 139
Joseph A. Arruda, G.R. Marzolf

Long Term Grazing Control of Algal  Abundance:
A Case History	 144
Richard A. Osgood

Biological Control of Nuisance Algae by Daphnia Pulex:
Experimental Studies	 1i51
Michael J. Vanni

Spring Daphnia Response in an Urban Lake	 157
Terry A. Noonan
       Socioeconomic and Political Issues Associated with the
       Implementation Phase of the Bear Lake 314
       Clean Lakes Study	219
       Craig Thomas, Vincent Lamarra, V. Dean Adams
       The Nitrogen, Phosphorus, and Carbon Budgets of a
       Large Riverine Marsh, and Their Impact on the
       Bear Lake Ecosystem	223
       flex C. Herron, Vincent A. Lamarra, V.  Dean Adams
       The Effect of Coprecipitation of CaCO3 and Phosphorus
       on the Trophic State of Bear Lake	229
       Paul B/rdsey, Vincent Lamarra,  V.  Dean Adams
       SEDIMENT ANALYSIS
       Sediment Metals Accumulation in a Suburban Lake .,.. 235
       John D. Koppen, Stephen J. Souza

       Sediment Inflows and Water Quality in an
       Urbanizing  Watershed	239
       David F. Brakke

       Sediment Distribution and Quality in a Small
       Wisconsin Reservoir	243
       Robert C. Gunkel, Jr., Robert F. Gaugush,
       Robert H. Kennedy

       Analysis of Surficial Sediment from 63 Illinois Lakes	248
       M. Kelly, R. Hite, K. Rogers

       The Engineering Characteristics of Hydraulically
       Dredged Lake Materials	254
       James E. Walsh, Stanley M. Bemben, Carlos Carranza
MODELING TECHNIQUES AND INNOVATIONS
Use of a Predictive Phosphorus Model to Evaluate
Hypolimnetic Discharge Scenarios for
Lake  Wallenpaupack	
H. Kirk Horstman, Roger S. Copp, Frank X. Browne

Water Quality Simulation of the Proposed
Jordanelle Reservoir, Utah	
David L. Wegner

Time Series Modeling of Reservoir Water Quality
Robert H. Montgomery

Modeling Developments Associated with the
University Lakes Restoration Project	
Ronald F. Malone, Daniel G. Burden,
Constantine E. Mericas
1(55



171


175



1H2




1H6
A Cross-sectional Model for Phosphorus in
Southeastern U.S. Lakes	
Kenneth H. Reckhow, J. Trevor Clements

Phytoplankton-Nutrient Relationships in South Carolina
Reservoirs: Implications for Management Strategies ...  193
Jeffrey Pearse

Relationships Between Suspended Solids, Turbidity,
Light Attenuation, and Algal Productivity	  108
floss Brown

Verification of the Reservoir Water Quality Model,
CE-QUAL-R1, Using Daily Flux Rates	  206
Carol Desormeau Collins, Joseph H. Wlosinski
CASE STUDY: THE BEAR LAKE PROJECT
A Historical Perspective and Present Water Quality
Conditions in Bear Lake, Utah-Idaho	213
Vincent A. Lamarra, V. Dean Adams, Craig Thomas,
Rex Herron, Paul B/rdsey, Victor Kollock, Mary Pitts
COMPARATIVE ANALYSIS OF RESERVOIRS
Regional Comparisons of Lakes and Reservoirs:
Geology, Climatology, and Morphology	261
Kent W. Thornton

Lake-River Interactions: Implications for Nutrient
Dynamics in Reservoirs	266
Robert H. Kennedy

Intermountain West Reservoir Limnology and
Management Options	272
Jerry Miller

Factors Controlling Primary Production in Lakes
and Reservoirs. A  Perspective	277
Bruce L. Kimmel, A/an W. Groeger

Organic Matter Supply and Processing in Lakes
and Reservoirs	282
Alan W. Groeger, Bruce L. Kimmel

Mixing Events in Eau Galle Lake	286
Robert G. Gaugush

Empirical Prediction of Chlorophyll in Reservoirs	292
William W. Walker, Jr
       FISHERY MANAGEMENT
       Effects of Fish Attractors on Sport Fishing Success
       on Norris Reservoir, Tennessee	299
       fl. Glenn Thomas, J. Larry Wilson

       Recent Applications of Hydroacoustics to Assessment
       of Limnetic Fish Abundance and Behavior	305
       Richard E. Thome, Gary L. Thomas

       Use of Columbia River Reservoirs for Rearing by
       Juvenile Fall Chinook Salmon and Some
       Management Implications	310
       Gerard A. Gray, Dennis W. Rondorf
                                                      VI

-------
The Expansion of the White Perch, Morone Americana,
Population in Lake Anna Reservoir, Virginia	314
A.C. Cooke

Catch Composition and Potential Impact of Baited and
Unbaited Commercially Fished Hoop Nets in
Three Central Florida Lakes	320
Marty M. Hale, Joe E. Crumpton, Dennis J. Renfro

Development of Fish Populations and Management
Strategies for the Blenheim-Bilboa Pumped
Storage Reservoirs	 324
David L. Thomas, Quentin Ross,
Alan Milton,  James M. Lynch


URBAN LAKE QUALITY
Fate of Heavy Metals in Stormwater Management
Systems	 329
Harvey H. Harper, Yousef A. Yousef,
Martin P. Wanielista

A Probabilistic Evaluation of Instability in
Hypereutrophic Systems	335
Daniel G. Burden, Ronald F. Malone

Occurrence and Control of Taste and Odor
in Sympson  Lake	340
G.C. Holdren, R. Major Waltman
                                                  345
                                                  356
ACIDIC PRECIPITATION
Calcite Dissolution and Acidification
Mitigation Strategies	
Ha raid U. Sverdrup

Ontario's Experimental Lake Neutralization Project:
Calcite Additions and Short-term Changes in
Lake  Chemistry	
LA. Molot, J.G. Hamilton, G.M, Booth

Adirondack Experimental Lake Liming Program	360
Douglas L. Britt, James E. Fraser

Considerations of Prudence and  Equity for Protecting
Lakes from Acid Precipitation	368
Alfred M. Duda

Studies on the Use of Limestone to Restore Atlantic
Salmon Habitat in Acidified Rivers	374
W.D. Watt, G.J. Farmer, W.J. White

Lake Acidification and the Biology of Adirondack
Lakes: Crustacean Zooplankton Communities	380
James W. Sutherland, Scott O. Quinn,
Jay A. Bloomfield, Clifford A. Siegfried

The Littoral Zooplanktic Communities of  an Acid
and a Nonacid Lake in Maine	
Mike Brett
                                                  385
Soil Liming and Runoff Acidification Mitigation	389
Per Warfvinge, Harald Sverdrup

CASE STUDIES OF WATER QUALITY IMPROVEMENTS
The Improved Water Quality of Long Lake Following
Advanced Wastewater Treatment by the
City of Spokane, Washington	395
Raymond A. Soltero, Donald G. Nichols

Economic Returns and Incentives of Lake
Rehabilitation: Illinois Case Studies	405
Krishan  P. Singh, V, Kothandaraman,
Donna F. Sefton,  Robert P. Clarke

An Historical Overview of a Successful  Lakes
Restoration Project in Baton Rouge, Louisiana	412
Ronald M. Knaus, Ronald F. Malone
                                                         Dredging of Creve Coeur Lake, Missouri	416
                                                         Greg Knauer
                                                         Reservoir Management Planning: An Alternative
                                                         to Remedial Action	423
                                                         Donald W. Anderson
                                                        TROPHIC STATUS
                                                        The Trophic State Concept: A Lake
                                                        Management Perspective	427
                                                        Robert E. Carlson

                                                        Who Needs Trophic State Indices?	431
                                                        Richard Osgood

                                                        Trophic State Indices in Reservoirs	435
                                                        William W.  Walker, Jr.

                                                        Trophic State Indices: Rationale for
                                                        Multivariate Approaches	 441
                                                        Patrick L. Brezonik

                                                        Trophic State Classification of Lakes with
                                                        Aquatic  Macrophytes	 446
                                                        Daniel E. Canfield, Jr., John R. Jones
MACROPHYTE CONTROL
An Overview of Chemicals for Aquatic Plant Control — 453
James C. Schmidt

Effects of Mechanical Control of Aquatic Vegetation
on Biomass, Regrowth Rates, and Juvenile Fish
Populations in Saratoga Lake, New York	456
Gerald F. Mikol

Restructuring Littoral Zones: A Different Approach
to an  Old Problem	 463
Sandy Engel

An Evaluation of Pigmented Nylon Film for Use
in Aquatic Plant Management	467
Michael A. Perkins
                                                         ROLE OF LOCAL LAKE ORGANIZATIONS AND
                                                         PUBLIC EDUCATION
                                                         Volunteer Lake Monitoring: Citizen Action to
                                                         Improve  Lakes	
                                                         Donna F. Sefton, John R. Little,
                                                         Jill A. Hardin, J. William Hammel
Small Lakes Symposia Programs	
Virginia M. Balsamo

Grass Roots Lake and Watershed Management
Organization	
Robert Burrows, John D, Koppen

Lake Associations and Their Role in the
Massachusetts Clean Lakes Program, 1983	
Richard Gelpke

Michigan Lake & Stream Associations, Inc.
Three Rivers, Michigan	
Donald Winne
                                                  473
                                                                                                          478
                                                                                                          482
                                                                                                          487
                                                                                                          491
                                                         RESTORATION TECHNIQUES
                                                         Control of Algal Biomass by Inflow Nitrogen	493
                                                         Eugene B.  Welch, Mark V. Brenner, Kenneth L. Carlson

                                                         Methods and Techniques of Multiple Phase Drawdown-
                                                         Fox Lake, Brevard County, Florida 	498
                                                         Robert J. Massarelli
                                                      VII

-------
 Restoration of Sebasticook Lake, Maine, by
 Seasonal Flushing	  502
 Cher Rock, David Courtemanch, Thomas Hannula

 Minneapolis Chain of Lakes Vacuum Sweeping
 and Runoff Diversion	  508
 John R. Erdmann, Norman B.  Wenck, Perry Damon

 Long-term Evaluation of Three Alum Treated Lakes	  513
 Paul J. Garrison, Douglas R. Knauer
WETLANDS AND LAKE INTERRELATIONSHIPS
Responses of Wetland Vegetation to Water Level
Variations in Lake Ontario	519
Wolf-Dieter N. Busch, Lynn M. Lewis

Limiting Nutrient Flux into an Urban Lake by
Natural Treatment and Diversion	525
William D. Weidenbacher, Peter R. Willenbring

The Effects of Shorezone Development on the Nature
of Adjacent Aquatic Plant Communities in
Lac St. Louis, Quebec	!529
7.C. Meredith
DESTRATIFICATION TECHNIQUES
Prediction of Lake Response to Induced Circulation	!>31
Robert A. Pastorok, Thomas M. Grieb

Thoughts on Selection and Design of
Reservoir Aeration Devices	837
Perry L. Johnson

Effects of Aeration on Lake Cachuma,
California, 1980-82	542
John R. Boehmke

Review of Design Guidance on Hydraulic
Destratification	 «>49
Jeffrey P. Holland

Enhancement of Reservoir Release Quality
with Localized Mixing	Ji52
Jeffrey P. Holland
WATERSHED MANAGEMENT
Illinois Soil and Water Conservation Districts
Action Program for Lake Watershed Improvement	555
Harold Hendrickson, Warren Fitzgerald, Roger Rowe

Watershed Management: Modifications in
Project Approach	553
Donald R. Urban, Walter Rittall

Watershed Management: Cooperation and
Compromise	
William K. Norris

A Screening Methodology for the Selection of
Urban Lakes'  Enhancement	
Car/a N. Palmer, Martin P. Wanielista,
Russel L, Mills,  Gilbert Nicholson, Robert Haven

Comprehensive  Monitoring and Evaluation of
the Blue Creek Watershed	
Thomas E. Davenport
                                                                                                          561
564
570
SEDIMENT PROBLEMS AND MANAGEMENT TECHNIQUES
Can a Microcomputer Help the Manager of a
Multipurpose Reservoir? The Experience of
Lake Como	575
R. Guariso, S. Rinaldo, R. Soncini-Sessa

Vancouver Lake: Dredge Material Disposal and Return
Flow Management in a Large Lake Dredging Project	580
Richard Raymond, Fred Cooper

Dredging and Dredged Material Disposal Techniques
for Contaminated Sediments	586
Raymond L. Montgomery

Dredging for Controlling Eutrophication of
Lake Kasumigaura, Japan	592
Ken Murakami

Gibraltar Lake Restoration Project—a Research and
Development Program for Evaluation of the
Transportation (Dredging) of Contaminated
Sediments	 599
Raymond E. Spencer
                                                    viii

-------
North American  Lake  Management  Society



International  Symposium on Lake & Reservoir Management

Third Annual Conference
October 18-20, 1983
Knoxville, Tennessee


President
Robert J. Johnson
Tennessee Valley Authority
Knoxville, Tennessee

Conference Committee Chairman
Wayne Poppe
Tennessee Valley Authority
Chattanooga, Tennessee

Program Chairman
Lowell Klessig
College of Natural Resources
University of Wisconsin—Extension
Steven's Point, Wisconsin

Exhibits  Chairman
Kent Thornton
Ford, Thornton, Norton & Associates
Little Rock, Arkansas

Additional Committee Members
Alfred Duda
Tennessee Valley Authority
Knoxville, Tennessee
Ronald  Raschke
U.S. Environmental Protection Agency
Athens, Georgia

Publications and Proceedings Editors
Judith Taggart, Editor
Lynn Moore, Associate
JT&A, Inc.
Washington, D.C.
Kenneth M. Mackenthun, Technical Editor
Environmental Consultant
Greenville, South Carolina

Session  Chairmen and Co-chairs
Opening Plenary: Mohamed T. EI-Ashry, World Resources Institute, Washington, D.C.

Water Quality Assessment Methods I: John Grossman, Tennessee Valley Authority, Knoxville, Tenn.; Michael
  Mullen, Engineering Analysis, Inc., Huntsville, Ala.

Water Quality Assessment Methods II: Lowell Keup, U.S. Environmental Protection Agency, Washington, D.C.;
  Terry Anderson, Kentucky Department for Natural Resources & Environmental Protection, Frankfort.

Agricultural Runoff and Water Quality: Tim Bjork, South Dakota Department of Water and Natural Resources,
  Pierre; Donna Sefton, Illinois Environmental Protection Agency, Springfield.

State Program Development: Priorities and Strategies: Jean Gregory, State Water Control Board, Richmond,
  Va.; Ron Manfredonia, U.S. Environmental Protection Agency, Boston, Mass.

Internal Nutrient Cycling: Eugene Welch, University of Washington, Seattle; Richard Harvey, South Carolina
  Department of Health & Environmental Control, Columbia.

Biomanipulation: Joel Schilling, Consultant, St. Paul, Minn.; Richard Osgood,  Metropolitan Council—Minneapolis,
  St. Paul, Minn.


                                             ix

-------
 Modeling Techniques and Innovations I: Kenneth Reckhow, Duke University, Durham N C • Todd Harris
    Southeast Wildlife Service, Athens, Ga.                                           "


 Modeling Techniques and Innovations II: Jack Waide, U.S. Army Corps of Engineers Vicksburq Ms • Steve
    Chapra, Texas A&M University, College Station.

 Case Study: The Bear Lake Project: G. Chris  Holdreri, University of Louisville, Louisville, Ky.

 Sediment Analysis: Vernon Myers, Florida Department of  Environmental Regulation, Tallahassee- Robert
    Gaugush, U.S. Army Corps of Engineers,  Vicksburg, Ms.

 Workshop: Clean Lakes Program: Frank Lapensee,  Michael Moyer, U.S. Environmental Protection Agency
    Washington, D.C.


 Comparative Analysis of Reservoirs: Jerry Miller, U.S. Bureau of Reclamation, Salt Lake City  Utah- Robert
    Kennedy, U.S. Army Corps of Engineers, Vicksburg, Ms.

 Fishery Management I: Paul Frey, U.S. Environmental Protection Agency, Athens, Ga.; Robert Gilliom  US
    Geological Survey, Reston, Va.


 Fishery Management II: Ronald L Raschke, U.S. Environmental Protection Agency, Athens  Ga • Al Brown
   Tennessee Valley Authority, Knoxville, Tenn.

 Urban Lake Quality:  John Jones, University of Missojri, Columbia; Yousef Yousef, University of Central
    Florida, Orlando.

 Acidic Precipitation  I: Harvey Olem, Tennessee Valley Authority, Chattanooga, Tenn.; Douglas Britt  International
   Science & Technology, Reston, Va.

 Acidic Precipitation II: Douglas Britt, International Science & Technology, Reston, Va.; Harvey Olem  Tennessee
   Valley Authority, Chattanooga, Tenn.

 Case Studies of Water Quality Improvements: Michael Moyer, U.S. Environmental Protection Agency
   Washington, D.C.; Dale Toetz, Oklahoma State University, Stillwater, Okla.

 Trophic Status: G. Dennis Cooke, Kent State Univers ty, Kent, Ohio.

 Macrophyte Control: Michael  Dennis, Breedlove Associates, Orlando, Fla; Leon Bates, Tennessee Valley
   Authority, Muscle Shoals, Ala.

 Role of Local Lake Organizations and  Public Education: Thomas U. Gordon, Winthrop Lakes Environmental
   Center, Winthrop, Maine; Virginia Balsamo,  Barrinciton, III.

 Restoration Techniques: Gareth Goodchild, Ministry of Natural Resources, Ontario, Canada- Spencer Peterson
   U.S. Environmental Protection Agency, Corvallis, Ore.

 Wetlands and Lake Interrelationships:  Richard McVoy, Massachusetts Division of Water Pollution  Control,
   Westboro; Richard Ruane, Tennessee Valley Authority, Chattanooga, Tenn.

 Destratification Techniques: Richard Ruane, Tennessee Valley Authority, Chattanooga, Tenn.

 Public Pressures for  Lake Management/Clean Water: Alfred Duda, Tennessee Valley Authority, Knoxville,
   Tenn.;  Lowell Klessig, University of Wisconsin—Extension, Steven's Point, Wis.

 Watershed Management:  Martin Wanielista,  University of South Florida, Orlando; Fred Davis, South Florida
   Water Management District, West Palm Beach.

 Microbiological Ramifications of Multiple Uses of Lakes and Reservoirs (oral presentation only): George
   Gibson, University  of Maryland, College Park.

 Sediment Problems and Management Techniques: Spencer Peterson, U.S. Environmental Protection Aaency
   Corvallis, Ore.

 Small Pond Management (oral presentation only): Thomas  Forsythe, Tennessee Valley Authority Golden
   Pond, Ky.

 Research Needs (oral presentation only): Richard Ruane, Tennessee Valley Authority, Chattanooga, Tenn.

 Cosponsors

U.S. Environmental Protection Agency,  Office of Water

U.S. Department of Agriculture, Office of Rural Development Policy, Soil Conservation Service

-------
Tennessee Valley Authority
Electric Power Research Institute
City of Knoxville
Delta Airlines
A list of conference attendees is available upon request from the North American Lake Management Society,
  P.O. Box 217, Merrifield, Va. 22116.
Proceedings Book
Production by Stephen J. Downs III and John M. Frazier
Cover art by Patricia J. Perry
Copies of this book may be ordered from the North American Lake Management Society, P.O. Box 217,
  Merrifield, Va. 22116. Cost: $10, postage and handling.
                                                  XI

-------
                                               Nonpoint  Source
                                             Pollution   in   Lakes
ERA'S  EMERGING NONPOINT SOURCE ROLE
REBECCA W. HAMMER
U.S. Environmental  Protection Agency
Washington, D.C.
I think there are two issues that we can all agree on:
We favor clean lakes in this country, and we know they
are continually threatened by nonpoint source pollu-
tion—be it excess sedimentation or an overloading of
nutrients.
  As you know, Section 314 of  the Clean Water Act
authorized the Clean Lakes Program, which got under-
way in 1975. In a demonstration program, we learned
quite a bit about degraded lakes  and became aware of
how important lake preservation is to the people of
this country. Since 99 percent of our population lives
within 50 miles of one of our approximately 37,000
publicly owned freshwater lakes, the public has  a
demonstrated  stake in clean lakes.
  The Clean Lakes Program was a sequence of assis-
tance grants to help States build programs to address
lake  problems at  the State  and local  levels.  EPA
funded over 100 projects in its initial Phase I study and
spent $78.6  million to diagnose and  implement
measures to restore degraded lakes.
  Admittedly, this was only a beginning. An estimated
85 percent of  this Nation's lakes are degraded  and
approximately one  third  need  protective   and
restorative measures. The job is not over in any sense,
but one way we  can move toward improvement is
through control of nonpoint source pollution. Here is
what EPA is trying to do.
  When EPA  first attempted to  quantify  nonpoint
source problems  in the early 1970's,  we noted that
there were billions of tons per year of soil erosion in
the United States  and a very high volume  of urban
stormwater  runoff.  As  our  understanding   has
developed, we have attempted to go beyond the focus
on total  loads and concentrate  instead on those
specific cases where nonpoint  sources  must be ad-
dressed to meet water quality goals, protect aquatic
resources, and realize the full benefits of point source
control investments.
  With the establishment of technology-based stan-
dards well underway for point sources, pollution from
nonpoint sources represents a continuing water pollu-
tion control problem in some areas. The Aquatic Life
Survey we are just completing shows that only 20 per-
cent of the Nation's streams are  affected by point
sources of municipal and industrial  pollution. Any
degradation of the remaining 80 percent is then non-
point source in origin. The survey results show 34 per-
cent of stream miles are adversely affected by non-
point sources.
  The 1982 State water quality assessments report
prepared under Section 305(b) of the Act describes the
nature of this problem: (1) in one fifth of the States,
nonpoint sources  are now the most important cause
of water degradation in  those waters not  meeting
designated uses; (2) agricultural nonpoint sources are
the major nutrient sources in  lakes with eutrophica-
tion problems; and (3) of the 20 States which specifi-
cally quantified their progress towards the 1983 Clean
Water Act goals,  15 (75 percent) listed nonpoint
sources as significant sources in their remaining pro-
blem waters.
   Over the last decade, our understanding and experi-
ence with effective control has increased significant-
ly.  Now implementation should  proceed in  water
bodies identified as being seriously affected by non-
point source pollution. The major question facing us is
how to accomplish this in the most effective and effi-
cient way possible. In June of this year, EPA met with
representatives of Federal agencies,  State and local
governments, conservation districts, forest industry
representatives, and environmentally concerned  in-
terests to discuss how we should  all proceed. There
appeared to be a consensus on two major points.
   First, State and local governments should have the
major implementation responsibility. They unques-
tionably have a broader range of authorities and more

-------
 LAKE AND RESERVOIR MANAGEMENT

 detailed knowledge of site-specific conditions than
 exists at the Federal level. They are able therefore to
 tailor nonpoint source control programs to meet their
 specific needs. Existing State programs have included
 both voluntary  and regulatory  approaches and  ap-
 propriate financing mechanisms including fees and
 cost sharing.
   Second, while States must have the primary respon-
 sibility for  nonpoint source program development and
 implementation,  EPA  must review its emphasis 01
 nonpoint sources. The participants in our meeting em-
 phasized that EPA should provide a Federal overview.
 It should assure that States address nonpoint source
 problems, particularly in interstate waters, and assist
 the States  in this effort. The Agency also has a major
 responsibility to  ensure that the technical  expertise
 and knowledge gained at the Federal, State, and local
 level, through working on specific problems  in dif-
 ferent  parts of the country, are targeted and made
 available to other communities with similar problems.
   The Administrator is reviewing our present situation
 and believes that we can take a number of steps to en-
 sure that  the Agency effectively  carries  out  this
 responsibility. First, under existing Clean Water Act
 authorities, the Agency  can issue a strong  policy
 statement requiring States to update or develop non-
 point source programs where nonpoint sources cause
 or contribute significantly to violations of water quali-
 ty standards  and  impede  designated  water uses.
 States will  be able to  use the existing water quality
 management process  to accomplish  this.  EPA  will
 also work with the Department of Agriculture to target
 technical assistance to State-identified agricultural
 nonpoint source problems.
  We believe the information available clearly shows
 that agricultural sources are the major area to address;
 if we are to implement an effective nonpomt source;
 program. EPA needs the USDA technical assistance
 delivery systems to make sure that the many farmers;
 in this country understand the impact their activities
 can  have on water quality and how they can modify
 their own farming practices to protect water quality
 USDA has a long and successful history of working al
 the grass roots to develop conservation practices, to
 educate the farmers on their practices and to provide1
 technical assistance in applying  them. At the county
 level, Agricultural Stabilization and Conservation Serv-
 ice, the Soil Conservation Service, and the Extension
Service all work together through various commitees
and  agents to provide support  and assistance  to
 farmers needing or requesting help. By supporting ad-
 ditional funding to the Department of Agriculture, EPA
 intends that they will be able  to focus their delivery
 systems on the areas with critical water quality prob-
 lems and measurably improve water quality  condi-
 tions.
   In term of resources, EPA will encourage States to
 use funds available under Section 205(j) of the Act for
 nonpoint source program development and will pro-
 vide Agency technical and program support. In addi-
 tion, EPA will use $2 million in appropriated FY 1984
 extramural funds for a nonpoint source report to Con-
 gress, guidance on alternative control strategies, and
 technical  transfer efforts including urban  runoff in-
 itiatives with outside organizations. Agency positions
 will  be  allocated at Headquarters and in the EPA
 Regional offices for nonpoint source control  in FY
 1984. The Agency is presently examining its FY 1985
 needs, including  the  possibility of providing  funds
 under  Section  304(k)  for USDA  nonpoint  source
 technical assistance to be targeted on State-identified
 needs. In  addition, EPA will work with  USDA to see
 whether  Agricultural  Conservation  Program  grant
 funds can be  used to  support approved  nonpoint
 source control programs.
  Some nonpoint source control proposals provide for
 Federal  financial support for the implementation of
 needed controls. In his testimony before the Senate,
 Administrator Bill Ruckelshaus has expressed a con-
 cern about this  because he believes that such an ap-
 proach would create a disincentive to voluntary action
 by anyone not receiving Federal funds. We are hopeful
 that  the approach  EPA is following will  yield signifi-
 cant water quality benefits without the serious draw-
 backs of creating  a queue for Federal  cost-sharing
 funds.
  This country, according to the Conservation  Foun-
 dation, annually spends  $2.5 to $3 billion  for water
 treatment, dredging, loss of reservoir capacity, flood
 control,  channel maintenance, and other  damages
 from sedimentation and  pollutants  associated with
 erosion.  The  Administration,  Congress,  and  the
 Federal agencies recognize that the nonpoint source
 problem can no longer be neglected. There is  broad
 agreement to take action, evidenced by the legislative
 initiatives being discussed in  conjunction  with the
 reauthorization  of the Clean Water Act. We in EPA
 have renewed our commitment to implementing a non-
 point source control program under the Clean Water
Act. Such a program is vital for clean lakes.

-------
USDA PROGRAMS AND NONPOINT SOURCE POLLUTION
RICHARD D. SIEGEL
Deputy Assistant Secretary for Natural Resources and Environment
U.S. Department of Agriculture
Washington, D.C.
USDA has been concerned for some time about non-
point source pollution of streams by various types of
runoff from cropland and pastureland, including soil
erosion and runoff of fertilizer, pesticides and animal
waste.
  USDA has estimated, in its 1977 Natural Resources
Inventory, that 5.3 billion tons of  soil erodes on non-
Federal land each year and that a percentage (25-50
percent on the average for individual fields)  reaches
streams.
  Most cropland today is heavily fertilized, and 15 per-
cent of fertilizers that are applied wash into surface
waters. Pesticide use has increased. These chemicals
are reaching streams. Twenty-five percent of livestock
operations have the potential to degrade water quali-
ty. Finally, 100 million tons of salt  return to surface
water through irrigation.
  So agriculture can and does cause nonpoint source
pollution. Even before the 1972 Federal Water  Pollu-
tion Control Act, with its landmark Section 208 dealing
with nonpoint source pollution,  several  States had
begun to regulate nonpoint source  pollution by agri-
culture. But  it is extremely hard to  generalize  about
nonpoint source pollution from agriculture. Above all,
it is site specific. The sheer volume of pollutant is not
a measure of whether nonpoint source pollution oc-
curs. And  in  many situations  where  agricultural
pollutants enter a stream  and the stream  itself
violates  water  quality standards,  there are  other
sources feeding pollutants into  the stream at  the
same time. Failing  septic systems leak nutrients and
bacteria.  Streambanks erode and supply sediment.
Road maintenance work and the air itself are other
sources.  Finally, streamflow conditions are a further
variable in determining whether pollution occurs. Dam
construction and irrigation can affect streamflow.
  In short, even in the most rural  setting where pollu-
tion is found, pollutants from agricultural land may
not be the only culprits—even though they might be
the most visible.
  Since 1972 we have been in the era of Section 208.
Planning activities  by States under  Section 208 have
placed USDA and EPA in a cooperative posture to ex-
amine the  problem and  address those  sources of
pollution that originate from farmland. In 1983 we have
found renewed interest in nonpoint  source pollution.
There has been general interest on Capitol Hill, and
EPA's Chesapeake Bay study has brought this matter
close to official Washington's backyard. This partner-
ship of EPA and USDA to address nonpoint source
pollution is reviving.
  The partnership is appropriate because USDA, in its
50  years of soil conservation programs, has been
educating farmers  and ranchers on soil-conserving
practices and assisting them financially  in installing
these practices. The technology of conserving soil on
farmland by  preventing runoff  is  USDA's stock-in-
trade. The purpose  of these USDA efforts has been to
preserve  the soil's productivity, but an incidental
benefit from  a  good soil conservation system on a
farm or ranch is better water quality off the property.
At the forefront  of  USDA's  work with  individual
farmers  and  ranchers have  been  two  important
delivery mechanisms; first, 3,000 soil and water con-
servation  districts throughout  the  Nation.  These
districts are units of local  government established
under State law.  It  is through  these districts that
USDA provides technical assitance personnel to work
with private landowners.
  Second, the ASCS  county  committees dispense
cost-sharing funds for conservation practices.
  The aim of these USDA efforts, however, has been
voluntary action by farmers and ranchers, not coercive
regulation. And this has led to a certain amount of ten-
sion and impatience, especially since the 1972 Act,
when Section 208 gave nonpint source pollution more
visibility as a national concern. The early version of
Senator Durenburger's amendment to the Senate ver-
sion of the Clean Water Act extension this year would
have moved in the direction of more regulatory teeth in
the Federal law for nonpoint source pollution control.
This made USDA uneasy. The  history  of nonpoint
source policy in the Federal Government since 1972
has been,  then, a "shotgun marriage" between EPA
and USDA.
  The way has been slower than many have wanted.
After Section 208 planning activity began in 1972 and
was underway for a few  years,  based as it  was on
voluntary actions of farmers and ranchers, there came
a new round of emphasis by Federal agencies. In early
1977 came the Model Implementation Program, an at-
tempt by  USDA and EPA to  use existing program
authorities to address the  nonpoint source problem in
seven project areas. Four of these were lake projects:
   Lake Herman in South Dakota
   Indiana  Heartland in Indiana (Eagle Creek Reservoir)
   Broadway Lake in South Carolina
   West  Branch of Delaware River in New York (Can-
   nonsville Reservoir)
These projects used EPA funds from both Section 208,
Water Quality Management, and Section  314, the
Clean Lakes Program. In  addition, there were USDA
Agricultural  Conservation Program (ACP) incentive
payments, technical assistance by the Soil Conserva-
tion  Service, research  funds from both EPA and the
Agricultural Research Service, and funds from a varie-
ty of other State and local sources.
  The success of this initial effort led USDA to ear-
mark $20 million of the ACP program in 1979 for water
quality projects. Congress then responded in 1980 and
1981 with $70 million for the Experimental Rural Clean
Water Program which will  continue through 1991. The
terminology of "model program," "experimental pro-
gram," "demonstration program" has been pervasive.
It indicates that the Federal role in the control of non-
point source water quality problems has not yet been
clearly defined. Yet what has been achieved has been
an  awareness  of  the   problem  by  conservation
districts, ASCS, and SCS. Water quality is part of the

-------
LAKE AND RESERVOIR MANAGEMENT

conservation agenda in many local areas. A local and
State constituency exists that did not before.
  The USDA's RCA Appraisal of 1980 was a sweeping
examination of  a variety of resource  and environ-
mental problems fitting under the heading of "Soil and
Water Conservation." Its purpose  was to form the
basis of a program, issued in  1982, on how USDA
could strengthen its conservation role. The conclusion
in our program, frankly, was that USDA nationally had
been trying to do too many things for too many pro-
blems all over  the United States. It had only so much
to spend. More was not  expected. It was time to set
clear-cut priorities and devote the bulk of the program
resources in USDA to them. We selected erosion con-
trol to ensure  productivity, water conservation,  and
upstream flood damage prevention. And we are selec-
ting  specific areas  for  targeting of financial  and
technical assistance.
  One might say that this was bringing the USDA con-
servation  effort  "back to basics."  While I  did  not
elevate water quality to a specific priority for national
emphasis, our national program is designed to foster
the implementation of soil conservation practices for
the reduction of erosion. Reductions in soil runoff are
necessary for an effective water quality management
program. Our basic strategy, therefore, benefits water
quality concerns in  a major way. We are again in an
active dialogue with EPA over how USDA can assist in
the new efforts to combat nonpoint source pollution. If
Congress  gives  EPA a stronger mandate for water
quality, then USDA stands ready to assist in this effort
with its delivery  system and expertise on  farmer at-
titudes, soils, crops, erosion problems, and conserva-
tion practices themselves.

-------
POLITICS OF ACID RAIN
A. O. SHINGLER
Executive Director
The Federation of Ontario Cottagers' Associations, Inc.
Scarborough, Ontario, Canada
We have made substantial progress in our war against
acid rain. Our determination to win it remains undimi-
nished.
  We have signed cooperative agreements with the
State of New York and the Federal Republic of Ger-
many, and we have new reason for optimism in events
that are  reshaping  the  Environmental  Protection
Agency of the United States.
  It's been a busy and productive year, with the On-
tario Ministry of the Environment well in the forefront
of activity.
  In June  of  last year the acid  rain  issue was
thoroughly examined at an international conference in
Stockholm. Experts from 21  European nations, the
United States, and Canada reached 29 conclusions,
which  can  be summarized in one short  statement:
"Unless  we  reduce our  emissions  of  sulfur  and
nitrogen oxides, more lakes and streams, more ground
water,  more soils and forests will become acidified
and we will be adding to the  economic and esthetic
damage we have already done."
  This  was  a profound  conclusion,  and  we are
grateful for it. It reflects with great accuracy the posi-
tion  Ontario  has  advanced   consistently,  and
sometimes against powerful opposition, since we first
defined the  problem and began working toward its
solution in the mid-1970's.
  As part of its acid rain program, Ontario has been
operating for more than 2 years now an extensive acid
deposition monitoring network,  one of the  most ad-
vanced of its kind in North  America. Both wet and dry
deposition  of atmospheric acids and related sub-
stances are measured since acidity comes down not
only with the rain, but also  through absorption  of
gases  and particles by vegetation, waterbodies, and
other surfaces.
  To monitor wet  deposition  at  more than 50 sites
across the Province, we are using special samplers,
which open only when it is raining or snowing. The
precipitation is analyzed for acids and related sub-
stances (such as sulfates and nitrates), neutralizers  of
atmospheric  acidity (ammonia,  and  calcium), and
various trace metals.
  Exploratory  experiments are  also underway   to
determine the  deposition  of mercury,  pesticides,
PCB's, and other organic contaminants by precipita-
tion. An analysis of the available data has  shown that
wet deposition of acidity,  sulfates,  and  nitrates  is
greatest in southern Ontario.
  In 1981, for example, wet sulfate  loadings ex-
ceeding 20 kg per hectare per year, which is thought
by  our scientists to be critical  for  sensitive water-
bodies, were occurring in all of the southern portion  of
the Province (south of 46° N). In this same  area, preci-
pitation pH values were generally less than 4.6, in-
dicating an acid content more than 10 times that ex-
pected for clean water in equilibrium with atmospheric
carbon dioxide.
  A meteorological analysis of the data indicates that
at least 50 percent of the wet  deposition of acidic
substances in 1981  was associated with air flows
from one quadrant—that between the south and west
compass directions. Air masses reaching southern
and central Ontario from this quadrant have passed
over heavily industrialized areas in the United States
and Ontario that have high emission rates of sulfur
and nitrogen oxides.
  Dry deposition across the Province is inferred from
the air  concentration  of sulfur and nitrogen  com-
pounds, as well as a number  of trace metals. An
especially designed air monitoring network measures
these substances at 27 sites across  the Province.
  Interpretation of the data from this network is still
at an early stage, but preliminary results indicate that,
in southern Ontario,  dry deposition of acidic  sub-
stances,  such as sulfates, is comparable in magni-
tude to wet deposition, while in northern Ontario, most
of the atmospheric acidity is delivered by precipita-
tion.

  Analysis of the data from these networks is an on-
going activity of the Ministry's scientists. A number of
reports have already been published. Several more are
in preparation including a joint project with Environ-
ment Canada and the Ministry of the Environment of
Quebec to assess the impact of the Sudbury smelters
on acidic deposition by comparing data obtained dur-
ing  the recent period when the smelters were shut
down, with corresponding data  when the smelters
were operating.
  Such deposition monitoring and data analysis ac-
tivities are expected to continue for  several years to
come, to determine changes accompanying emission
controls that  will be instituted in Ontario and all of
Canada and which we hope  will be  instituted in the
United States.
  In Nova Scotia, another  Province affected  by acid
rain, many rivers no longer support salmon. It is esti-
mated that in the United States some 36,000 square
kilometers of surface water  are  receiving excessive
amounts of acid  rain.
  Elsewhere,  there is evidence that acidic deposition
leads to the removal of important plant nutrients and
the release of toxic metals from the soils, thus threat-
ening forests. Toxic metals  have been traced  from
soils to ground water and eventually  to streams.
  In  Germany, scientists believe the mobilization of
metals in forest soils resulting from acid precipitation
is causing dieback in their forests.
  The recently released final reports of the Canada-
United States Work Groups,  established under the
1980 Memorandum of Intent, provide up-to-date scien-
tific  information on acid rain. While not all of the
members of the Work Groups agreed on all points, a
number of conclusions  can  be drawn from  the
Memorandum of Intent reports:

-------
LAKE AND RESERVOIR MANAGEMENT
   • Sulfur deposition causes both short- and  long-
 term damage in areas vulnerable to acid rain.
   • Wet  sulfate deposition above 20 kg per ha per
 year (or  18  Ibs per acre) in vulnerable  areas  is
 associated with damage. Areas with deposition less
 than 17 kg per ha per year have no recorded damage.
   • The damage is caused by sulfur deposition and
 the solution  is to reduce it.
   • Acid rain falls on eastern North America in and
 downwind from the major industrial regions.
   • Technology exists  to reduce emissions by sub-
 stantial amounts.
   To determine the threat  of acid rain  to  surfeice
 waters in  Ontario, our Ministry is continuing with Ihe
 survey of acid sensitivity status. The third annual  sum-
 mary of this program is  available as fact sheets.
   These ongoing surveys of the susceptibility of lakes
 in Ontario to acid rain are based on chemical analyses
 of water samples taken from each lake. Our data  base
 has now  increased to 4,016 lakes, up from  2,619 in
 1982.
   The primary factor in  determining the sensitivity of
 a lake to acidification is  its alkalinity; we  have
 classified lakes into five categories.
   Level One lakes  have zero  or negative alkalinity.
 They have already become acidic and many or all fish
 species may be absent from these lakes. Of the  more
 than 4,000 lakes actually tested, 155 or 4 percent  were
 in this category.
   Level Two lakes have very low alkalinity (<40 jceq
 l~1)  and  are extremely sensitive  to  heavy  acid
 loadings.  Fishkills and other biological damage may
 occur in these lakes during spring runoff. Thirteen pier-
 cent of the surveyed lakes were in this category.
   Forty-one percent of the lakes were moderately ssn-
 sitive (40-200 ^eq  M), being less at risk  in  com-
 parison to Level Two lakes, and 18 percent were class-
 ed as having low sensitivity. (200-500 ^eq l~1). These
 lakes are  likely to experience biological damage only
 under extreme snowmelt conditions during spr ng
 runoff.
   So, a total of 72  percent of  the 4,000 survey lakes
 showed some sensitivity to acidification. The remain-
 ing lakes are not considered sensitive to acid loadings
 because they contain sufficient buffering  capacfty to
 neutralize acid rain  for an indefinite period.
   Our Ministry has also  completed surveys on !he
 acidity of ground water in the Muskoka-Haliburton,
 Sudbury,  North Bay, and Timmins areas  of  the Pro-
 vince, sampling over 350 domestic wells. Results in-
 dicate that well water was acidic (pH 6.0) in 5 percent
 of  the Sudbury wells  and  12 percent of those  in
 Muskoka-Haliburton.
  However,  while the acidity  of surface waters is
caused  largely by  acidic deposition,  groundwaler
acidity is more commonly the  result of naturally ac-
cumulating carbonic acid—formed from the reaction
of carbon dioxide in  the soil with water. However, this
may  change  in  the  future.  Suggests  data  from
Sweden,  there, some areas have  groundwater pH
values of  less than  5, which has been  attributed to
acidic precipitation.
  In acidic well water in both Muskoka-Haliburton
and  Sudbury, the drinking water objectives for  lead
and copper have been exceeded. This occurs because
metals from  piping and joints are released in the
acidic water. While medical advisers have not  sug-
gested the presence  of any  major  or widespread
danger to public health, they have advised discretion.
For instance, it is not  advisable to mix baby formula
with the first water taken from a tap that has not bean
 turned on for several days. The Ministry has circulated
 to newspapers and health units in affected areas a
 notice urging cottagers drawing water from lakes and
 wells to flush the taps before use as this reduces the
 metal levels below water quality objectives.
   In addition to determining the extent of the acidifi-
 cation problem in  surface and ground water, the
 Ministry  has undertaken a comprehensive research
 program  into the mechanisms of acidification. This in-
 formation  is obtained  from  intensive sampling of
 lakes, streams, and ground water in a  relatively small
 (about 10) number of representative systems called
 calibrated  watersheds.  Two to 7 years of data are
 available on these systems and a detailed picture of
 the physical, chemical, and biological nature of acid-
 stressed lakes  is emerging.
   For example, several  important  results arise from
 this work: sulfur deposition is of greater importance
 than nitrogen deposition in the  acidification of lakes;
 most of the acidic input to lakes and streams occurs
 during the period of spring melt with resultant short-
 term depressions of pH; elevated levels of aluminum
 in surface waters are associated  with low pH, and
 aluminum can reach levels shown to be lethal to fish
 in laboratory experiments.
   Information from the calibrated watersheds is being
 used to develop mathematical  equations linking the
 deposition of acidic compounds to the chemistry of
 surface waters. The relationships established by the
 detailed studies can be extrapolated to large numbers
 of lakes from which less complete data are available.
  Of course, it is the biological damage caused by
 acidification and the resultant high metal levels that
 are of uppermost importance. As a result of our in-
 house research and that done by universities, we are
 getting a clearer picture  of the nature of this damage.
 We now know that fishkills have been observed in one
 lake in Muskoka during spring runoff when the lake pH
 is low. Here are some other findings:
  •  Complete loss of fish populations has been ob-
 served in lakes  in the Sudbury area concurrent with a
 decline in lake pH.
  • The concentrations  of trace metals such as mer-
 cury, lead, and  cadmium are elevated  in fish in lakes
 of low pH.
  • A  decline  in the breeding population  of  some
 types of  amphibians has been  observed in streams
 with low pH.
  • Changes have been observed  in the occurrence
 and abundance of zooplankton in acid-stressed lakes.
  • Changes have been observed  in the occurrence
 of algal  species  in acid-stressed  and acidic  lakes
 which may be detrimental to the recreational use of
 the lakes. For example, lakes with reduced pH support
 more filamentous algae  attached to the lake bottom.
 In other  acid-stressed lakes, an alga is appearing
 which causes "rotten cabbage" odors.
  The research program that our Ministry has under-
taken complements that  of the Canadian  Federal
 Government and  is similar to  those  of the  U.S.A.,
Sweden, and Norway.  The results are well-respected,
 and the work has been  presented at numerous con-
 ferences  and appeared  in  many scientific publica-
tions.
  Taken together with the survey  data, information
 from the calibrated watersheds will  be  used as a data
 base from which to develop abatement strategies to
 halt acid rain. For example, the findings on the relative
contribution of sulfur and nitrogen deposition to acidi-
 fication have definite  implications for abatement
strategies.

-------
                                                                     NONPOINT SOURCE POLLUTION IN LAKES

  Ontario's position has been that reductions in acid    Sudbury and two in the Parry Sound area. The lakes
rain should be carried out by emission control at the    have been monitored to obtain background data and
source.  However, we  have also undertaken  joint  in-    one was treated with neutralizing chemicals beginn-
vestigations with the  Ministry of Natural Resources    ing in August 1983.
regarding  the  feasibility and  effects  of  artificial      The best answer, we have always contended is ob-
neutralization of acidic and acid-stressed lakes. Three    vious: Cut off the source.
lakes have been selected  for the experiment, one near

-------
RURAL AMERICA: EMPHASIS ON CLEAN WATER
WILLARD (BILL) PHILLIPS, JR.
Director Office of Rural Development Policy
U.S. Department  of Agriculture
Washington, D.C.
Nonpoint pollution is a serious problem. Clean lakes
are essential to rural Americans' way of life. Part of rny
mission at the Department of Agriculture is to see that
rural America is improved. All your work with the study
of lakes, the relationship of pollutants, management
of lakes and  related projects, can only benefit rural
Americans.
  The North American Lake Management Society and
the Office of  Rural Development Policy have a great
deal in common:
  • We are very concerned with improving the quality
of life in rural America with your basic concern of how
polluted lakes reflect on rural life.
  • Both of our organizations are newly formed and
with each  others'  help we will  continue  to  gein
momentum.
  Though the Office of Rural Development Policy may
be  relatively  new, its  leadership and coordination
responsibility is not. The Agriculture Act of 1970 made
rural development a major mission of the Department
of Agriculture. This continued recently with the Ruial
Development Act of 1980, which established the posi-
tion  of  Under Secretary for Small Community and
Rural Development and directed us to prepare a na-
tional rural development strategy—a  game plan for
rural America.
  Congress has recognized for many years that it is
essential for the Federal Government to have a ruial
development policy focus. Rural America needs an ad-
vocate to coordinate this work, and the Office of Ruial
Development Policy is that advocate.
  One of the specific tasks that Congress required
our Office to do was to deliver a national rural develop-
ment strategy.
  Secretary Block defined what he expected from that
strategy:

  What  I am  looking for is a practical strategy  for
  responding positively to the diverse problems and  op-
  portunities in Rural America. To be more specific, I
  want a strategy  developed to (1)  identify  emerging
  rural issues  and needs on an ongoing basis;  (2)
  strengthen the State and local government role in rural
  development; (3) contribute to strengthening local
  economic  viability  and   improving  community
  resources through encouraging the private sector to
  expand its role in rural development; and, (4) develop
  and implement policy guidelines  that can provide
  sound government program direction  for service  to
  rural America.

  We did just as  the  Secretary asked  when we
delivered  "Better  Country: A Rural  Development
Strategy for the 1980's" to Congress in January 1983.
  The goal of the Strategy was to come up with affor-
dable and practical recommendations which address
the most pressing  needs in rural America. Those areas
in which the Federal Government can be most useful.
Where its help can be most beneficial and  least ob-
trusive. This first Strategy did not  address every rural
issue, but  the first annual  strategy update is under-
way. The update will address new rural issues and ex-
pound on recommendations from last year.
  A new rural America lies beyond the farm, home to
nearly 60 million people of which 5 million are farmers
and their families and 54 million are nonfarmers. This
is rural America whose economy must diversify to sur-
vive, and where a majority of Americans say they
would prefer to  live.
  In recent years, dramatically increasing numbers of
Americans have been moving from cities to rural com-
munities. These new rural  pioneers are not farmers;
rather, they are doctors, engineers, teachers, business
executives, and laborers, people of every race, age,
and station of life. The fact is rural America is chang-
ing  rapidly and those changes are making  great
demands on rural lands and water. We have to plan for
these changes,  we  have to sit down and forge strong
policy or we will lose  the natural resources of this
frontier.
  The rural areas  of America  are blessed with an
abundance of  renewable and  nonrenewable natural
resources. Good soil and good water are needed to
produce the food,  fiber,  and timber that  benefit all
Americans—both urban and rural. Natural resources
are a productive base that when developed can mean
more  jobs in economically distressed  areas where
they are most needed. Investments in clean, beautiful
lakes allow  local  communities  to  reap economic
benefits from enhanced tourism and recreation use.
  USDA has estimated that 5.3  billion tons of erosion
occur on non-Federal lands each year and that 25 to
50 percent reach aquatic systems; 15 percent of ap-
plied fertilizers and 5 percent of pesticides wash into
surface waters; 100 million tons of salt enter surface
waters through irrigation; 25  percent  of livestock
operations may degrade water quality.
  These  statistics  are a statement of  serious prob-
lems. The National  Advisory  Council  on  Rural
Development,  a  23-member   body appointed  by
Secretary Block, chose to address natural resource
preservation and conservation  as a  priority issue in
the strategy update this year.
  Even though I am not directly involved with the ac-
tual administration of any  EPA or USDA clean  water
programs or projects, I can give you the rural perspec-
tive of how these projects help develop rural areas.
For example:  the Rural Clean Water Program  has a
strong economic impact on Tillamook  County, Ore.
The $2.3  million Tillamook bay  drainage basin, Rural
Clean Water Program will generate an estimated $8.6
million in local revenues. Revenues will be generated
in wages, salaries, taxes, and  the  fishing  industry
along with other sectors  of the  economy.
  Other spinoff benefits  from this project associated
with the Improvement of  rural life are:
  « Reduced closures oi Tillamook  Bay to  shellfish
harvesting because of high fecal coliform count;
  • Improved  salmon  fish spawning  and rearing
areas;

-------
                                                                    NONPOINT SOURCE POLLUTION IN LAKES
  • Reduced Food and Drug Administration threat of
withdrawing  certification  of the shellfish growing
areas for use in interstate  commerce;
  • A sustained  healthy dairy  industry and related
food  processing  facilities that  provide jobs  in  the
county because dairy farms contributed to the non-
point  pollution of  the bay area.
  People must realize that polluted lakes aren't a pret-
ty sight in many ways. And clean lakes make  for
economic as well  as attractive resources.
  Prairie Rose Rural Clean Water  Project in Shelby
County, Iowa, is another Clean Lake and rural develop-
ment  success story. The Prairie Rose Lake's fishing,,
boating, swimming, and other recreational activities
were  seriously impaired by  sediment and nutrients
from agricultural nonpoint  problems. After 3 years of
best management projects (BMP) installed on farms in
the watershed the lake has become quite functional
once again, restocked with fish, and enjoyed by many
rural and urban residents. The lake was resurrected by
controlling nonpoint pollution.
  It  is  quite clear to me that a  fine balance  of
agriculture, rural  enterprises, clean lakes,  resource
conservation, and strong infrastructure make up the
total rural community. They need each other to work
effectively. They are two sides of the same blade.
  The new rural America sustains the farming com-
munity at least as much as it is sustained in return. Its
expanding requirements for land and water and other
natural resources are matters of legitimate concern to
the farmer, nonfarmer, and urban resident.  Further,
the emergence of this new  rural  America as the
residence of choice for most Americans may have pro-
found social, economic, and political consequences
for our country as we approach  a new century.
  Nonpoint pollution must be controlled if we want to
have a healthy rural America because, as Will Rogers
said, "Even if you're on the right track, you'll get run
over if you just sit still." The  Department of  Agricul-
ture, the Office of Rural Development Policy,  and the
North American Lake Management Society are on the
right track—and moving!

-------
                                                         Water  Quality
                                    Assessment   Methods
 EVALUATING RESERVOIR TROPHIC STATUS: THE TVA APPROACH
 JANICE PLACKE COX
 Division of Services and Field Operations
 Tennessee Valley Authority
 Chattanooga, Tennessee


            ABSTRACT
            TVA tributary and mainstem reservoirs show generalized differences in morphometry, hydraulics,
            nutrient loads, and response to nutrient conditions. Neither type of reservoir is comparable to the natural
            lakes on which classical eutrophication studies have been based The majority of published trophic
            state indices and standards (e.g., hypolminetc dissolved oxygen depletion, Secchi depth, areal nutrient
            loading rates, m-reservoir phosphorus concentrations) are inappropriate for evaluation of the trophic
            status of some or all TVA reservoirs. Relative trophic state indices were developed for mainstem and
            tributary reservoirs using relevant potentiating and response variables. Ranking of the mainstem reser-
            voirs is based on planktonic chlorophyll, macrophyte coverage, hydraulic retention time, reservoir area
            less than 5 feet deep, annual pool elevation drawdown, and Secchi depth. Ranking of the tributary
            reservoirs is based on planktonic chlorophyll, total phosphorus and total nitrogen weighted by the
            N-P ratio, and bioavailable inorganic carbon levels.
INTRODUCTION

The term trophic status has been used indiscriminant-
ly to describe both the abundance of nutrients and the
intensity of biological productivity supported by the
nutrient flux. Realizing that reservoir eutrophication is
a  complex concept involving  the  interaction of
physical and chemical driving forces and biological
responses, it is useful to draw a distinction between
trophic potential and trophic response. Trophic poten-
tial can be defined as the theoretical carrying capacity
of the aquatic ecosystem and is a function of nutrient
concentrations, light penetration, climate, hydraulic
regime, and so on.  Trophic response refers to the
amount, type and rate of biomass production, and the
water quality variations that occur during ecosystem
assimilation of that biomass. Trophic  response is
realized  within the hypervolume of physical  and
chemical trophic potential factors as well as complex
biological  interactions including interspecific com-
petition, grazing, parasitism, allelopathy, and artificial
manipulation by herbicide application and biomass
harvesting.
  The prediction of trophic response and associated
water quality  from  a  limited  number  of  easily
measured trophic potential factors is a worthy  lake
management goal. It requires an identification of the
trophic potential factors most strongly correlated  with
trophic response  variables for the particular water-
bodies in question. These relationships can then be
summarized in a multivariate trophic state index  that
allows a relative classification of a group of water-
bodies or monitoring of eutrophication of a single
waterbody over a period of time.
  However,  because the  functional  dynamics of
aquatic ecosystems and the symptoms by which ex-
cessive productivity is manifested are highly depen-
dent on lake type and geographical region, the search
for a universal trophic state index is futile. No purpose
is served in  loyalty to a single standardized index if
                                             11

-------
LAKE AND RESERVOIR MANAGEMENT
the index does not represent the mechanics and mani-
festations of eutrophication accurately and helpfully.
Because TVA reservoirs are not functionally  com-
parable to the natural lakes on which most classical
eutrophication studies have been based, new indices
were developed for evaluating reservoir trophic status.
CHARACTERISTICS OF TVA RESERVOIRS

There are two types of reservoirs in the TVA system:
mainstem reservoirs, which are impoundments of the
Tennessee River, and tributary reservoirs, which are
impoundments of the  tributaries of  the Tennessee
River. While the hydraulic  retention time of natural
lakes is often measured  in years,  complete water
replacement generally takes less than 10 days in the
mainstem reservoirs and less  than  150 days in the
tributary reservoirs.
   Like natural lakes, the tributary reservoirs undergo
thermal stratification during the summer and autumn.
Thermal stratification  in  the  mainstem  reservoirs,
however, is intermittent or nonexistant. The mainsteri
reservoirs and most of the tributary reservoirs are fair-
ly shallow (z less than  20 m). Both types of reservoir
have large shoreline development indices (DL up to 30)
and extensive overbank areas forming large littoral
zones. In the tributary reservoirs, winter drawdown of
pool elevation for flood control results in an annual
drying and freezing of the littoral zone, which hinders
macrophyte colonization.  Mainstem  reservoir eleva-
tions, however, are kept more constant for navigation
with the result that shallow embayment areas are fre-
quently  infested  with  Myriophyllum,  Najas,
Potamogeton,  and floating mats of  Oedogonium,
Mougeotia, and Lyngbya.
   TVA reservoirs are rather turbid in comparison with
most natural  lakes, with Secchi depths averaging 2-3
m in the tributary reservoirs and approximately 1 m in
the mainstem reservoirs. The  settling of suspended
particles  as water flows through the tributary reser-
voirs often gives rise to longitudinal gradients with in-
creasing  water clarity  and decreasing nutrient con-
centrations  in a  downstream  direction.  Strong
longitudinal gradients  are not  typically found in the
mainstem reservoirs.
others). These  models assume steady-state  condi-
tions  and  continuous  stirred  tank reactor (CSTR)
behavior (that is, lake phosphorus concentration, in-
flows,  and  outflows  are constant;  complete instan-
taneous mixing occurs so that lake concentration is
uniform and lake concentration equals outflow con-
centration).  The performance  of  such models  in
predicting phosphorus  concentrations in TVA reser-
voirs has been disappointing because TVA reservoirs
do not meet the model assumptions. Most TVA reser-
voirs have  an  areal hydraulic  loading (mean depth/
hydraulic retention time) in  excess  of 50 m/yr. Reck-
how (1979)  and Chapra (1975) have noted that phos-
phorus budgets of such waterbodies may be governed
by  mechanisms different  from those  operating  in
natural  lakes.  Phosphorus  inflows and phosphorus
concentrations in TVA reservoirs do not show steady-
state behavior. Hydraulic loading and phosphorus in-
flows  tend to vary seasonally in TVA reservoirs (Hig-
gins and Kim, 1981; Placke,  1983). Because  the reser-
voirs  have  short hydraulic  retention times, annual
averages of phosphorus  loading  fail  to accurately
reflect the  seasonal  variability  of reservoir  phos-
phorus concentrations. Seasonal variability in nutrient
loading is further complicated  in the tributary reser-
voirs where nutrient-laden inflow may plunge beneath
the epilimnion  during  stratification. TVA  tributary
reservoirs  also  violate  the  assumption of  complete
mixing. Phosphorus  concentrations  are  generally
higher in the upstream portions of the tributary reser-
voirs,  leading Higgins and  Kim (1981) to propose a
plug flow reactor (PFR) model.  Vertical heterogeneity
also exists in the  tributary reservoirs during thermal
stratification.Epilimnetic phosphorus concentrations
in these reservoirs tend to be less than hypolimnetic
(and outflow) concentrations.
  One final problem  in trying  to relate phosphorus
loading to trophic status, is uncertainty of the form of
phosphorus to be  modeled.  Available models predict
total  phosphorus  concentrations  using total  phos-
phorus loads, but there is no evidence that  the entire
phosphorus load  is  biologically available. Further-
more,  the  settling behavior and retention of  par-
ticulate and dissolved forms of phosphorus can be ex-
pected to differ. This may have particular significance
in  reservoirs with large inorganic suspended  solids
loads.
EVALUATING THE APPLICABILITY OF
TRADITIONAL TROPHIC STATE
INDICATORS TO TVA RESERVOIRS

The merit of a particular trophic state index depends
on  the appropriateness of the trophic potential and
trophic response indicators used in the index and, to a
lesser extent, on the numerical methods used to syn-
thesize  index values.  A wide variety of  indicater
variables  and  standards  have been  proposed  fcr
natural lakes but their applicability to TVA reservoirs
must be evaluated individually.

Trophic Potential Indicators
Empirical Phosphorus Loading Models

Several   empirical  models  and  theoretical  mass
balance models have been developed to predict in-
lake  phosphorus  concentration as  a  function of
phosphorus load (Dillon and Rigler, 1974; Dillon and
Kirchner,  1975;  Kirchner and Dillon,  1975; Volleri-
weider, 1975, 1976; Jones and Bachman, 1976; Larsen
and Mercier, 1976; Chapra, 1977; Reckhow, 1979; and
Phosphorus and Nitrogen Concentrations

Phytoplanktonic  chlorophyll concentrations in the
mainstem reservoirs are not significantly correlated
with either phosphorus  or  nitrogen concentrations.
Rather, the mainstem reservoir phytoplankton appear
to be limited by the shallow light penetration relative
to mixed depth, and by hydraulic washout. Studies on
the embayment  macrophytes Myriophyllum, A/a/as,
and Potamogeton have shown that nitrogen and phos-
phorus levels found in mainstem reservoir waters and
sediments  are   not  limiting to  growth and  that
macrophyte infestation  is  a function  of  available
substrate, light  penetration, pool elevation  fluctua-
tion, and turbulence (Martin et  al.  1969; Peltier and
Welch, 1968). Because nitrogen  and phosphorus are
not limiting, classification of the trophic status of the
mainstem reservoirs based on nutrient concentrations
would be misleading.
  Chlorophyll concentrations in the  tributary reser-
voirs are weakly but  statistically  significantly cor-
related with total phosphorus. The regression equa-
tion  is  similar to thoc-'- reported in the literature (Ed-
                                                 12

-------
                                                                      WATER QUALITY ASSESSMENT METHODS
mondson, 1972; Schindler, 1978; Carlson, 1977; Dillon
and Rigler, 1974; Jones and Bachman, 1976; Past and
Lee, 1978; Williams et al. 1978).

           log (C) = 1.298 log (P)  - 0.670
                (R2 = 0.59, n = 33)

 where C  = chlorophyll in mg/m3 (range 1-15 mg/m3)
 P  =  total phosphorus in mg/m3 (range 10-30 mg/m3)

The correlation coefficient between  chlorophyll and
total soluble inorganic nitrogen (TSIN) for the tributary
reservoirs as a group is 0.005. However, if only obser-
vations with total N:P ratios (mg/mg)  less than 25 are
used, a highly significant regression is generated.

         log (C) = 0.883 log (TSIN) - 1.032
                (R2 = 0.87, n = 15)

 where C  = chlorophyll in mg/m3 (range 2-15 mg/m3)
  TSIN =  total soluble inorganic nitrogen in mg/m3
               (range 30-350 mg/m3)

Thus, TSIN is a good predictor of chlorophyll concen-
trations under conditions traditionally considered to
be  nitrogen  limited.  Potentially  N2-fixing  hetero-
cystous cyanophytes are found in tributary reservoirs
with both  high and low total N:P ratios.
  The tributary reservoirs show large seasonal and
spatial variations  in  N:P  ratios.  It  is unlikely that
phytoplankton growth in any of these reservoirs is
always nitrogen limited or always phosphorus limited,
but rather that the importance of a particular nutrient
depends on the N:P ratio and other limiting variables
(light, temperature,  mixing depth, etc.) as well as the
seasonal succession of dominant algal types.

Inorganic Carbon

Readily assimilable inorganic carbon concentration
(H2CC>3 +  free CO^ is not often viewed as a  trophic
potential factor. The significance of inorganic carbon
availability is probably minimal in the  mainstem reser-
voirs  where alkalinity is high  and pH is consistently
near neutral. However, in certain tributary reservoirs
where the alkalinity is low (less than 10 mg/l) or where
algal  photosynthesis drives the  pH  to 10 or above,
readily assimilable  inorganic  carbon concentrations
have been observed to decline to less than 5 ^M. While
it seems  unlikely that the  availability of carbon ever
limits the  ultimate size of  the algal standing crop, it
may exert  qualitiative effects. King  (1970) provides
evidence that low carbon availability (10 iM) is one of
several factors that selects in favor of cyanophytes.
Additionally, transient carbon limitation of photosyn-
thetic rates  may  interfere  with   the  ability of
cyanophytes to regulate their buoyancy, leading to ob-
jectionable blooms  on the water surface (Paerl and
Ustach, 1982; and others).

Trophic Response Indicators

Secchi Depth

Use of Secchi depth as a trophic  response indicator
rests  on the assumption that  transparency is a func-
tion of chlorophyll concentration. Secchi depth in the
mainstem reservoirs is not correlated with chlorophyll
concentrations, indicating that light  penetration is
limited by inorganic turbidity and  color.  Because
phytoplankton production  in  the main channel and
macrophyte infestation in  the embayments  of  the
mainstem reservoirs are limited  to some degree by
 light, Secchi depth can be used as a trophic potential
 indicator with the assumption that mainstem reser-
 voirs with greater Secchi depths have the potential for
 developing larger phytoplankton crops and more ex-
 tensive macrophyte infestations.
   Secchi depth is weakly but statistically significantly
 correlated with  chlorophyll  concentration  in  the
 tributary reservoirs.

 log (S) = -0.315 log (C) + 0.596 (R2 = 0.36, n =  33)
   where S = Secchi depth in m
         C = chlorophyll in mg/m3

 The regression equation is similar to those of Carlson
 (1977) and Rast. and  Lee  (1978)  but  predicts lower
 transparency at low chlorophyll  concentrations, in-
 dicating that inorganic turbidity and color play a light-
 limiting role in the tributary reservoirs as well.

 Hypolimnetic Dissolved Oxygen (DO)
 Depletion

 The rate at which hypolimnetic oxygen is depleted and
 the  ultimate oxygen  deficit  have  frequently been
 employed as trophic response indicators (Edmondson
 et al.  1956; Hutchinson, 1957; Fruh et al. 1966; Hooper,
 1969; Bazin and Saunders, 1971;  U.S.  Environ. Prot.
 Agency,  1974;  Rast and  Lee, 1978; and  others). To be
 appropriate  response indicators for  TVA  reservoirs,
 the principal source of DO demand must be from settl-
 ing and decomposition of autochthonous organic mat-
 ter, the DO demand must  be exerted in the hypolim-
 nion,  and the  hypolimnion must not receive ogygen
 from  mixing or underflows. These requirements are
 not  met  by the tributary  reservoirs  as a group.
 Chlorophyll  concentrations  (as  a  measure  of
 autochthonous productivity) are not  correlated with
 hypolimnetic DO depletion  rates in these reservoirs (r
 =  0.26). The  summer  inflow of a  number of  the
 tributary reservoirs  enters as an underflow to  the
 hypolimnion.   In some   cases  the  BOD  and
 allochthonous detritus  of the underflows  exert  a
 stronger DO demand than settling  autochthonous
 matter, and in  other reservoirs a continual underflow
 of cold, oxygenated water obscures the autochthon-
 ous DO demand. In  some  of the tributary  reservoirs,
 settling autochthonous matter from the epilimnion is
 trapped at the thermocline, giving rise  to a negative
 heterograde DO profile with only a portion of the DO
 demand being  exerted in the hypolimnion. DO deple-
tion rates may also be biased by sediment oxygen de-
 mand, which reflects the enrichment of  years past as
well as present conditions. Consequently, the use of
 hypolimnetic  DO depletion rates as  indicators  of
trophic response in tributary reservoirs  was rejected.
Obviously, hypolimnetic  DO depletion is not a mean-
 ingful concept in the unstratified mainstem reservoirs.
Chlorophyll

Chlorophyll has received a great deal of attention as a
measure of  phytoplankton biomass and therefore a
trophic  response indicator (Sakamoto, 1966; Vallen-
tyne et al. 1969;  Natl. Acad. Sci., 1972; Dobson et al.
1974; U.S. Environ. Prot. Agency, 1974; Wetzel, 1975;
Brezonik, 1976; Carlson, 1977; and others). Chlorophyll
data  must  be  interpreted  with caution  because
chlorophyll  concentrations   are  related  to  the
physiological state and  light  history as well as the
mass  or volume  of   the   algal  standing  crop.
Nonetheless, chlorophyll remains the best of the easi-
ly determined measures of algal biomass because of
                                                 13

-------
LAKE AND RESERVOIR MANAGEMENT
its  specificity for algal material and  its ability  1o
distinguish between viable biomass and detritus.
  Chlorophyll  is valuable  as a trophic response  in-
dicator  in  open, water  areas  dominated  by phyto-
plankton, but  is inappropriate for reservoirs or por-
tions  of reservoirs dominated by macrophytes and
floating algal  mats. These macroscopic growths  re-
quire an independent estimate of standing crop.
Phytoplankton Community Structure Indices

Phytoplankton indices  of lake trophic type were in-
troduced by Thunmark (1945) and Nygaard (1949) and
have  been  further  developed  by  Palmer  (1969),
Stockner (1971), Taylor et al. (1979) and others. These
indices give various measures of the diversity of the
community or look for occurrence of certain indicator
taxa.  Lake evaluations  by phytoplankton communily
indices like those of Nygaard have  generally shown
poor correlation with other  quantitative trophic stale
indicators  such as phosphorus concentration or
chlorophyll. An EPA study by Taylor et al. (1979) based
on National Eutrophication Survey data demonstrated
that many of the so-called indicator taxa are actual y
tolerant of a wide variety of conditions.
  Even as phytoplankton community indices are refin-
ed,  they will be of limited utility for the routine evalua-
tion of trophic status. Computation of index values re-
quires that samples be studied by a highly skilled ta<-
onomist, and the value of the index is strongly depen-
dent on the number of samples studied and the tirre
of collection. Index values are not directly related to
water quality as perceived by  the water user,  and the
indices give little information on the degree or type of
enrichment or  the type of  management  strategy
necessary for control.

Macrophytes

Macrophyte colonization in  TVA mainstem reservoirs
is controlled by available substrate, light penetration,
turbulence, and pool elevation fluctuations. Nitrogen
and phosphorus are present in excess and do not limit
growth. Nonetheless,  macrophyte  infestation is a
manifestation of excessive  productivity and is an ap-
propriate trophic response indicator for the mainstem
reservoirs.  Macrophyte  colonization in the tributary
reservoirs is insignificant.
APPLICATION TO TVA RESERVOIRS

All trophic state indices are  biased by the choice of
component indicator variables. The bias can only be
reduced by choosing nonredundant variables that are
directly relevant to the eutrophication process in that
particular system. As previously emphasized, the ap-
propriate component variables vary from one system
to another. Because of the fundamental differences
between  TVA's tributary and mainstem reservoirs, a
separate index composed of the appropriate variables
was developed for each.

Tributary Reservoir Trophic State  Index

Chlorophyll  concentration is the most appropriate
trophic  response  variable  for  characterizing  the
tributary reservoirs. However, because of the seasonal
variability in chlorophyll concentrations, especial y
when there are sporadic cyanophyte surface blooms,
it is desirable to add trophic potential variables to the
index. Based on TVA's experience, carbon,  nitrogen,
and  phosphorus are the trophic potential variables
that should be used in assessing trophic status. In the
index, nutrient concentrations are weighted by a func-
tion of the  N:P ratio  so that  the  N concentration
receives greater emphasis under low N:P conditions
while P receives greater emphasis under high N:P con-
ditions (Table 1). The tributary reservoir index value is
defined to be  the sum of the trophic response rank
and the mean trophic potential  rank (i.e., chlorophyll
rank plus the mean of the C, N, and P ranks).
  The inclusion of  the chlorophyll trophic  response
variable ensures that the occasional reservoir with
relatively high nutrient concentrations but low produc-
tivity  is not drastically misranked. Conversely, the in-
clusion of nutrient  data stabilizes  the  chlorophyll
ranking, which is strongly dependent on when sampl-
ing is conducted relative to seasonal phytoplankton
dynamics. Data requirements for the computation of
the tributary reservoir index values are minimal, and
the index is relatively  straight  forward and easy to
compute. Index values may also be computed on a
site-by-site basis when longitudinal variations in reser-
voir water quality are of interest.

Mainstem Reservoir Trophic State Index

Chlorophyll concentrations in the main channel and
the percentage of the reservoir surface with rooted or
floating macrophytes and algal mats are the most ap-
propriate trophic response indicators in the mainstem
reservoirs.
  Phytoplankton in  the main channel are limited  by
shallow light penetration relative to the mixed depth
and by hydraulic washout. Factors controlling the oc-
currence of floating  algal mats have not been studied
by TVA, but algal  mats are generally restricted  to
areas  with  submerged  macrophytes.  Macrophyte
establishment  is limited by available substrate, light
penetration,  pool elevation fluctuations, turbulence,
and herbicide application. Therefore, hydraulic reten-
tion time, Secchi depth,  percentage of the  reservoir
with depth less than 5 feet, and the magnitude of pool
elevation  fluctuations are appropriate trophic poten-
tial indicators for mainstem reservoirs. Note that Sec-
chi depth is used here as an indicator of inorganic tur-
bidity rather than as an indicator of chlorophyll as in
traditional indices  (that  is, it is used  as a trophic
potential variable rather than as a trophic response
variable).
  As with the tributary reservoir index, it is desirable
to include trophic  potential variables along with
trophic  response variables  because  estimation  of
trophic response may vary from week to week during
the growing  season. The trophic potential variables
supplement and stabilize the index values, making
them  less dependent on the date that measurements
are made. Like the tributary  reservoir  index, the
mainstem reservoir trophic state index  is defined  as
the sum of the mean trophic response rank and the
mean trophic potential rank [mean trophic response
rank = (chlorophyll rank + macrophyte cover rank)/2;
mean trophic potential rank = (Secchi depth rank  +
retention time  rank  + percentage reservoir area less
than 5-feet deep rank + drawdown  rank)/4] (Table 1).

Caveats for Application of the TVA
Reservoir Trophic State Indices

The calculated index values do not have an  absolute
physical meaning but rather serve to rank the  reser-
voirs by trophic status. This facilitates assignment of
water quality  management  priorities  and suggests
which variables might be managed to improve water
quality.
                                                 14

-------
                                                                          WATER QUALITY ASSESSMENT METHODS
                              Table 1.—Calculation of TV A reservoir trophic state index.
                                           TRIBUTARY RESERVOIRS

  TSI VALUE = chlorophyll  + (carbon rank + nitrogens  rank + phosphorus rank)
                  rank                            3

  Chlorophyll = n(mean chlorophyll in reservoir X - minimum chlorophyll for n reservoirs)	
      rank                                   range in chlorophyll values for n reservoirs

                                 where n = number of reservoirs included in analysis

  Carbon rank computed as  above using mean epilimnetic free CO2+ H2CO3 values based on mean alkalinity and mean pH

  Phosphorus rank = n(P*x  - P*min)
                        P*
                        r range

  where n = number of reservoirs included in analysis
        Px = mean epilimnetic total phosphorus in reservoir X
        Px = Px(log (mean epilimnetic N:P)X), the weighted phosphorus value for reservoir X
        P^,in =  minimum of mean P*x values of n reservoirs
        prange = range of mean P'x values in n reservoirs

  Nitrogen rank = n (N,,' -  N^in)

                     "range

  where n = number of reservoirs included in analysis
        TSINX = mean total soluble inorganic nitrogen in the epilimnion of reservoir X
        NX* = TSINX/ log (mean epilimnetic N:P)X, the weighted nitrogen value for reservoir X
        N^ln =  minimum of mean Nx values of n reservoirs
        Hange = range of mean N,,' in n reservoirs
                                           MAINSTEM RESERVOIRS

  TSI VALUE = (chlorophyll + macrophyte) + (retention + Secchi  + percentage + drawdown)
                  rank          rank         time      depth       area         rank
                          2                 rank       rank     •< 5 ft.
                                                                 rank
  Chlorophyll, macrophyte infestation, retention time, Secchi depth and percentage of surface area less than 5-feet deep ranks
  computed as follows:

  Rank =  n(value for reservoir X - minimum value for n reservoirs)
                         range for n reservoirs

          where n = number of reservoirs included in analysis

  Since drawdown is inversely related to trophic status, drawdown rank is computed as follows:

  Rank =  n(1 - [value for rexervoir X - minimum value for n reservoirs])
                           range for n reservoirs
         where n = number of reservoirs included in analysis

   Because the proposed indices are relative, inclu-
sion of extreme cases in the data set will obscure the
differences in the more  similar reservoirs  by com-
pressing the relative scale. Also, inclusion of new data
outside  the   range  of  the  original  data  base
necessitates recalculation of all index values. The in-
dices are proposed for use in  evaluating TVA reser-
voirs; the indices are not intended for natural lakes in
the Tennessee Valley region, nor are they intended for
reservoirs In general. They can be properly applied on-
ly to reservoirs where the mechanics and manifesta-
tions of eutrophication are comparable to the TVA
reservoirs.
   The most eutrophic reservoirs as indicated by the
trophic  state  indices  need   not  be considered
"eutrophic" by the standards of other researchers.
The judgment of what is eutrophic  is subjective and
will vary greatly by geographic region, the water quali-
ty that users are accustomed to,  and the intended
uses of the water. The trophic state indices proposed
here are not water quality indices and reservoirs with
high trophic state rankings should not automatically
be considered to have decreased value as a resource.
REFERENCES

Bazin, M., and G. Saunders. 1971. Mich. Acad. 3:91-106.
Brezonik, P. 1976. Trophic Classifications and Trophic State
  Indices: Rationale, Progress, Prospects. Tech. Ser. Vol. 2
  No. 4. State of Florida Dep. Environ. Reg.
Carlson, R. 1977. Limnol. Oceanogr. 22:361-69.
Chapra, S. 1975. Water Resour. Res. 2:1033-34.

	1977. J. Environ. Eng. Div. Am. Soc. Civil Eng. 103:
  147-61.
                                                     15

-------
LAKE AND RESERVOIR MANAGEMENT
Dillon, P.,  and  W.  Kirchner. 1975. Water Resour. Res.
  2:1035-36.
Dillon, P., and F. Rigler. 1974. Limnol. Oceanogr. 19:767-73.
Dobson, A., M. Gilbertson, and P. Sly. 1974. J. Fish. Res.
  Board Can. 31:731-38.
Edmondson, W. 1972. Nutrients and phytoplankton in Lake
  Washington. Pages 172-193 in G. Likens, ed. Nutrients and
  Eutrophication: The Limiting Nutrient Controversy. Spec.
  Symp. Am. Soc. Limnol. Oceanogr. I. Lawrence, Kans.
Edmondson, W., G. Anderson, and D. Peterson. 1956. Limnol.
  Oceanogr. 1:47-53.
Fruh, E., K. Stewart, G. Lee and G. Rohlich. 1966. J. Water
  Pollut. Control Fed. 38:1237-58.
Higgins, J., and B-R. Kim. 1981. Water Resour.  Res. 17:371-
  576.
Hooper, F. 1969. Eutrophication indices and their relation to
  other indices  of  ecosystem change.  Pages  225-35 in
  Eutrophication: Causes, Consequences, Correctives. Natl.
  Acad. Sci., Washington, D.C.
Hutchinson, G. 1957. A Treatise on Limnology. Vol I. John
  Wiley and Sons, New York.
Jones, J.,  and R. Bachman. 1976. J. Water Pollut. Control
  Fed. 48:2176-82.
King, D. 1970. J. Water Pollut. Control Fed. 42:2035-49.
Kirchner, W., and P.  Dillon. 1975. Water Resour.  Res. 2:182-3
Larsen, D., and H. Mercier. 1976. J. Fish. Res.  Board Can.
  33:1742-50.
Martin, J., B. Bradford, and  H. Kennedy.  1969.  Factors  Af-
  fecting the Growth of A/a/'as in  Pickwick Reservoir. Natl.
  Pert. Dev. Center,  Tenn. Valley Author.
National Academy  of Science and National Academy of
  Engineering. 1972. Water Quality Criteria. A report of 1he
  Committee on Water Quality Criteria. Washington, D.C.
Nygaard, G. 1949. Kgl. Danske Vendenskab. Selskab, Biol.
  Shr. 7:1-295.
Paerl, H., and J. Ustach.  1982. Limnol. Oceanogr. 27:212-17.
Palmer,C. 1969. J. Phycol. 5:78-82.
Peltier, W., and E. Welch. 1968. Factors Affecting Growth of
  Rooted Aquatic Plants. Div. Health Safety, Tenn. Valley
  Author.
Placke, J. 1983. Trophic Status Evaluation of TVA Reservoirs.
  TVA/ONR/WR-83/7. Tenn. Valley Author.

Rast, W., and G. Lee.  1978. Summary Analysis of the North
  American (U.S  Portion) OECD Eutrophication  Project:
  Nutrient   Loading—Lake  Response  Relationships and
  Trophic State Indices. EPA-600/3-78-008. U.S. Environ. Prot.
  Agency, Washington,  D.C.
Reckhow, K. 1979. Empirical lake models for phosphorus:
  development, applications, limitations and uncertainty.
  Pages 193-221  in  D. Scavia  and A.  Robertson, eds.
  Perspectives on Lake Ecosystem Modeling. Ann Arbor Sci.
  Publ., Inc., Ann Arbor, Mich.
Sakamoto,  M. 1966. Arch. Hydrobiol. 62:1-28.
Schindler, D. 1978. Limnol. Oceanogr. 23:478-86.
Stockner, J. 1971. J. Fish. Res. Board Can. 28:265-76.
Taylor, W.,  L.  Williams, S. Hern, and  V.  Lambou. 1979.
  Phytoplankton Water Quality Relationships in U.S. Lakes.
  VII.  Comparison  of Some New  and Old  Indices and
  Measurements of Trophic State. EPA-600/3-79-079. U.S. En-
  viron. Prot. Agency, Las Vegas, Nev.

Thunmark, S. 1945. Folia Limnol. Scand.  3:1-66.

U.S. Environmental Protection Agency. 1974. An approach to
  a relative trophic index  system for classifying lakes and
  reservoirs. Work. Pap. No. 24.  Corvallis, Ore.

Vallentyne, J., J. Shapiro, and A. Beeton. 1969. The process
  of eutrophication and criteria  for trophic state determina-
  tion. Pages 58-67 in  Modeling the Eutrophication Process.
  Proc. Workshop, St. Petersburg, Fla.
Vollenweider, R. 1975. Schweiz. Zeitschrift. Hydrobiol. 37:53-
  84.

	1976.  Mem. Inst.  Ital. Idrobiol. 33:53-83.
Wetzel, R. 1975. Limnology. W. B. Saunders Co., Philadelphia.
Williams, L, V. Lambou, S. Hern, and R. Thomas. 1978. Rela-
  tionships  of Productivity and  Problem Conditions to Am-
  bient Nutrients: National Eutrophication Survey Findings
  for 418 Eastern Lakes.  EPA-600/3-78-002.  U.S.  Environ.
  Prot. Agency, Las Vegas, Nev.
                                                       16

-------
WISCONSIN'S  LAKES: A TROPHIC ASSESSMENT
RONALD H. MARTIN
Center for Energy and Environment Research
Wisconsin Department of Natural Resources
Madison, Wisconsin
            ABSTRACT

            A cooperative program between the Wisconsin Department of Natural Resources and the University
            of Wisconsin-Madison (UW-MSN) has resulted in the assessment of the trophic condition of approx-
            imately 3,000 significant inland lakes in Wisconsin. A sophisticated  set of computer programs has
            been developed by the UW-MSN that locates the lakes that are to be  classified, extracts the spectral
            reflectance values from Landsat digital data for those lakes, and then adjusts the lake data for at-
            mospheric effects. The corrected lake statistics from Landsat data are correlated to  measured field
            parameters collected on a limited number of lakes. Finally, the relationships developed are used to
            predict the trophic condition of the significant lakes in the State.
INTRODUCTION

Wisconsin is rich in lakes. There are about 15,000 in-
land lakes in the State, totalling  almost 1,000,000
acres. As a result, lakes are a very  important natural
resource to  Wisconsin with their recreational use a
significant part of the State's economy. As the trend
continues towards increasing recreational use of Wis-
consin's inland lakes, so does the rate of eutrophica-
tion. At the same time, though, Wisconsin  has in-
creased its awareness of the importance of protecting
lakes as a resource. Efforts have been directed toward
lake trophic classification as well  as lake manage-
ment, protection, and rehabilitation.  Both Federal
legislation (P.L 92-500 of the Federal Water Pollution
Control  Act Amendments) and State legislation (Wis-
consin's Lake Protection and Rehabilitation Law) re-
quire the Department of Natural Resources (DNR) to
monitor and classify all the significant lakes in the
State.
BACKGROUND

Initial efforts by the DNR to respond to P.L. 92-500
were to classify by conventional field methods about
1,100 lakes over 100 acres in size (Uttormark and Wall,
1975). Because such methods proved to be very costly
in terms  of collecting the lake  data, the Wisconsin
DNR in 1974 initiated  a project to investigate the
feasibility of using Landsat (land satellite) imagery to
monitor inland lake water quality. The DNR contracted
with the  Institute for Environmental Studies at the
University of Wisconsin-Madison  to conduct  the
research  project.  The goal of  the  research was to
develop a nearly  automated  system  which,  with
minimal human interaction, would locate, extract, and
correct the satellite data and then would classify each
lake in the State at minimal cost.
RESULTS OF INITIAL INVESTIGATIONS

The lake classification program was designed around
an interactive graphics terminal and the Madison Aca-
demic Computing Center's  UNIVAC 1110 computer.
Secchi disk  depths from  37 lakes  and turbidity
measurements from  27  lakes  were correlated with
band 5 digital data from Landsat (Landsat is a space-
craft traveling around the earth every 103 minutes at
an altitude of 570 miles).  The relationship between
band 5 scene brightness and Secchi disk depth show-
ed some scatter around a regression line (Holmquist,
1977; Fig. 1). Much of the scatter could be explained
by the interval between the field sampling date and
the Landsat overpass date. The correlation between
measured turbidity (field data) and predicted turbidity
from the satellite was  quite good (Holmquist,  1977;
Fig. 2).
  In still other investigations (Scarpace et al. 1979), it
was noted that brightness values in band 5 increased
in the summer months and decreased in the fall. The
increase in band 5 reflectance was correlated to an in-
crease in algal  turbidity levels in the summer.  From
these investigations it was apparent that just one date
of Landsat multispectral scanner (MSS) data would be
inadequate to  monitor something as dynamic as a
water body. Consequently, it was  decided  that  a
minimum of three dates of Landsat data from spring
through fall would be necessary to monitor the trophic
condition of Wisconsin's lakes.
  In addition, after carefully examining satellite im-
agery, it was found that light atmospheric haze signifi-
                 10 0    150     200

                    SECCHI DEPTH(FEET)
Figure 1.—Secchi disk versus band 5 exposure. Exponential
regression represented by the solid line.
                                                17

-------
 LAKE AND RESERVOIR MANAGEMENT
 cantly increased reflectance values (Scarpace et  al.
 1979). Since day to day differences in lake reflectance
 resulted  partly from atmospheric effects, obviously,
 data needed atmospheric correction.
   As a result of initial investigations it was concluded
 that (1) the Landsat multispectral scanner was cap-
 able of monitoring lake trophic conditions, (2) multi-
 temporal data were necessary,  and (3) corrections for
 atmospheric effects on data needed to  be made.
   Subsequent  to  these early investigations, a pilot
 study was conducted jointly by the DNR and the Uni-
 versity of Wisconsin-Madison on about  1,300 lakes in
 1976 with funding from the U.S.  Environmental Protec-
 tion Agency and  National Air and Space Administra-
 tion. A computer software package was developed to
 process multiple dates of Landsat multispectral scan-
 ner data for lake trophic assessment. Analyses of the
 results generated  from this demonstration project  in-
 dicated that remote sensing could be used to assess
 lakes' trophic condition (Martin and Holmquist, 1979).
   Because this pilot study generated interest both at
 State  and Federal  levels, continued funding was ob-
 tained (an EPA  grant) to conduct a full scale investiga-
 tion of Wisconsin lakes. Details of the various proce-
 dures and the results of the classification  follow.
cated in  the scene and the lake pixels (picture ele-
ments) within the polygons were extracted (Fig.  4).
From these extracted pixels, the lake's spectral values
for bands 4, 5, and 6 and the means and variances of
those band values were stored for classification. Usu-
ally the only interpretive assistance necessary in this
process was in the satellite navigation procedure and
in the inspection of the extracted output to confirm
that the navigation was accurate.
   After the data were  extracted and just  prior  to
classification,  an  atmospheric  correction  of the
satellite  data was  performed. The recorded signal
from the satellite was corrected for the effects of scat-
tering and absorption in the atmosphere. All dates of
data were then normalized to the clearest day.
   After correction for atmospheric  effects, the spec-
tral information was stored in  files, so that the Land-
sat radiance values could  be correlated  with field  in-
formation collected at the time of the Landsat over-
pass.
   The program package described  was  used to pro-
cess  34 different Landsat  tapes from  11  scenes  (a
scene is  approximately  110 square miles) to provide
multiple date  monitoring of the significant  lakes  in
 METHODOLOGY FOR EXTRACTION OF
 LANDSAT DATA

 A total of 2,925 significant  Wisconsin lakes—those
 larger than  20 acres and deeper  than 8 feet—were
 classified using information  from Landsat  digital
 tapes for 1979-81 (Martin et  al. 1983). Only a short de-
 scription of  the data extraction procedures is related
 here.
   Each lake was located on a U.S. Geological Sur/ey
 topographic map and the coordinates of its bounding
 polygon were  digitized and stored on a computer file
 (Fig. 3).  In addition, control points corresponding to
 easily recognized points on  satellite imagery were di-
 gitized, and  their latitude/longitude coordinates were
 placed on computer files.
   Each Landsat computer compatible tape (CCT) v/as
 navigated by an affine transformation  program using
 the digitized control  points. Each  lake was then  lo-
 to 5
 -n-OOOOOOOOOOOOOOOOOOOOOX++-t**-
                                                                 ++»*»*XOOOOOOOOOOOOOOOOOOOOOO»»»+»t
                                                                 •n-n-tXOOOOOOOOOOOOOOOOOOOOOOOO+*****
                                                                **+-t**OOOOOOOOOOOOOOOOOOOOOOOOOXn-»»*
                                                               »»+***ooooooooooooooooooooooooooaf+-f.i..i.
                                                               •n.**+XOOOOOOOOOOOOOOOOOOOOOOXXXXOX»ttt+
                                                               »tt*XOOOOOOOOOOOOOOOOXXOOOOX**+++*****
                                                              -f*»»XOOOOOOOOOOOOOOOOX*+»»XOX»»*-f»-ft**
                                                              *»xooooooooooooooooxx*»****ox*******
                                                               +XOOOOOOOX»OOOOOX»tt*-
                                                                                                         1130
                                                                                                         14140
                                                                                                         1150
                                                 1460
                                                                                                          1480
                                                             1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800
                                                 1490
Figure 2.—Plot of observed (measured) versus predicted tur-
bidity.
Figure  4.—Digital  output  from  program  EXTRACT  for
Pewaukee Lake.
                                                      18

-------
                                                                      WATER QUALITY ASSESSMENT METHODS
 Wisconsin.  Unique correction factors had to be ap-
 plied for each scene to correct for atmospheric haze.
 PROCEDURE FOR TROPHIC
 CLASSIFICATION

 To generate more accurate lake classification results,
 Wisconsin was divided into four regions based on sim-
 ilarity of soils, geology, vegetation, and water quality.
 Water quality samples were collected on lakes within
 each region (Fig. 5) concurrent with satellite overpass
 dates during the period of study (1979-81). A total of
 289 samples was collected statewide for ground truth
 (about 10 percent). The information collected on each
 lake included. Secchi disk transparency, chlorophyll a,
 turbidity, and color. The sampled data were used to
 establish relationships between the  MSS data (or
 spectral reflectance information)  and water quality
 data  for lakes within each region (Table 1). Stepwise
 regression procedures contained within the MINITAB
 statistical program package (Ryan et al.  1976) were
 used to accomplish this.  Only chlorophyll a content
 and Secchi disk transparency were shown to correlate
 well with MSS data.
                                         :GION 4
                                         90 samples
                                         on 58 lakes
Figure 5.—Ground truth sampling by region.
                                                  Thus two models—one based on chlorophyll a con-
                                               tent and one based on  Secchi disk transparency-
                                               were used to classify  lakes in each region. Landsat
                                               MSS data were correlated with the field data collected
                                               for each region. The lakes were classified by predic-
                                               ting chlorophyll a content  and Secchi  disk measure-
                                               ments from the Landsat  data.
                                                  In the final phase of  the classification process, pre-
                                               dicted Secchi disk measurements and chlorophyll a
                                               content were related to certain interrelationships be-
                                               tween water quality parameters measured in the field.
                                               The approach was patterned after work done by Carl-
                                               son (1977) in Minnesota. Carlson  developed a single
                                               criterion index based  on interrelationships between
                                               Secchi disk transparency, chlorophyll, and total phos-
                                               phorus.
                                                  A numerical scale from 0-100 was used in Carlson's
                                               trophic  state index (TSI).  By determining empirical re-
                                               lationships  such  as between biomass  (as measured
                                               by chlorophyll a) and transparency, Carlson developed
                                               different computational forms of TSI equations such
                                               as:

                                                    TSI(Chl) = 10 (6 - 2.04 - 0.68 In Chl-a)
                                                                          In 2

                                                    where Chl-a is chlorophyll a in ug/l

                                                 Wisconsin's  lake trophic classification  methodol-
                                               ogy was based on the ability of  Landsat to predict
                                               sampled parameters (i.e., Secchi  disk transparency
                                               and chlorophyll a). The next step  involved incorpora-
                                               ting these values into Carlson's TSI index. TSI equa-
                                               tions similar to Carlson's were developed from data
                                               collected on 705 Wisconsin lakes from 1976 through
                                               1981 (Martin et a). 1983). Table 2 lists the equations re-
                                               lating Secchi depth to chlorophyll  a content for each
                                               region.
                                               RESULTS OF THE LAKE CLASSIFICATION

                                               The lake classification (TSIs) resulting from the two
                                               linear models—TSI-Secchi disk and TSI-Chlorophyll
                                               a—were grouped  into  traditional  trophic  classes.
                                               Varying water quality characteristics were observed in
                                               the different regions of the State based on the percen-
                                               tage of lakes in each trophic category. For example,
                                               northern Wisconsin (Region 2) had a high percentage
                                               of oligotrophic lakes while southern and western parts
                                               of the State (Regions 1 and 3) had a higher percentage
                                               of eutrophic lakes. Table 3 lists the percentage of Wis-
                                               consin lakes in each trophic category.
                        Table 1.—Equations relating water condition data to MSS data.
Region

LnSD
LnChl

Region

LnSD
LnChl

Region

Ln SD
Ln Chi

Region

LnSD
Ln Chi
 I

= 0.914 - 0.013(652) - 0.810 (B6/B5) + 0.024(V6)  + 0.004(842) + 11.70(1/84)
 =  -11.034  + 3.77(B5)1'2 - 0.004(842)  + 23.00(1/85) - 0.013(V5) + 0.230 (V4)


= 4.183 - 15.90 (B5/(B4 + B5 + B6)) + 0.039 (B4/V4) + 5.50(1/66) - 0.038 (B5/V5)
 =  14.118 +  86.00 (B5/(B4 + 65 + 66)) - 25.80(65/64) - 43.00 (B4/(B4 + 65 +  66))
= 2.591 + 0.167 (V4) - 0.0047(662) + 1.72(86/86) + 3.00 (B4/(B4  + 65 + 66))
=  -0.406 + 0.041 (B5/V5) + 1.07(B51'2) _ 0.143 (V5)

IV

= 0.882 - 0.002 (652) - 0.054 (65/V5) + 0.025 (V6) - 0.990 (85/84)  + 7.20 (1/85)
=  -2.116 + 1.190(B6/V6) + 3.480(65/64) + 0.002(652) + 0.012 (V4) + 0.016 (84/V4)
 R2

54.19
58.40


41.72
51.76


59.82
35.30


60.55
64.46
                                                  19

-------
LAKE AND RESERVOIR MANAGEMENT


                        Table 2.—Equations relating Secchi depth to chlorophyll a content.
Region
1
2
3
4
Equation
Ln SD = 1.528 -
Ln SD = 1.98 -
Ln SD = 1.665 -
LnSD = 1.621 -

.4514 (Li CHL)
.529 (Ln CHL)
.4962 (Li CHL)
.4596 (Li CHL)
R2
53.8
53.6
64.8
56.0
Number of
Samples
135
326
110
134
                         Table 3.—Percentage of V/isconsin lakes in each trophic category.
   TSI
Trophic Classification
TSI-SD
TSI-CHL
 ^46
  47-49
  50-53
 ^•54
Oligotrophic
Oligo-mesotrophic
Mesotrophic
Eutrophic
  Average trophic conditions for a given lake v/ere
based on predictions of the chorophyll a concentra-
tion and Secchi disk transparency during the satellite
observation period (1979-81).
  Predicted TSI-Chlorophyll  a and TSI-Secchi  disk
values from  Landsat showed a TSI range over the
sampling season. Oligotrophic lakes had a range  of
about 9 TSI units while eutrophic lakes  had a range of
about 16 TSI units. TSI ranges similar to those pre-
dicted from Landsat were obtained from lakes sam-
pled quarterly by the DNR from 1966-78.
  Results from  the  TSI-Chlorophyll a  models v/ere
preferred because lakes with high color or turbidity
levels are more likely to be biased (that is, they were
classified as more eutrophic) by the TSI-Secchi disk
models.
LITTORAL ZONE MACROPHYTES

In addition to the investigations described earlier, de-
velopment of  a lake classification methodology for
Wisconsin  also  included examining measures  to
incorporate macrophyte abundance as a trophic:  in-
dicator. The macrophyte portion of the lake classifica-
tion project (Martin et al. 1983) focused on whether it
was feasible to use remote sensing to measure lake
macrophyte  abundance,  especially  submergent
growth. The conclusion of that portion of the report
was that it  is  not now possible, nor is it likely in the
near future,   to  detect  or  accurately  measure
macrophytes  using  satellite  systems.  Low altitude
photography can  be used  to estimate  macrophyte
abundance, although expense will limit the use of this
approach to only lakes of high interest.
  Since a technique to assess  inlake macrophyte
problems via remote sensing could not be developed,
weed growth abundance was not used as a trophic in-
dicator in this  study. However, it  is recognized  lhat
macrophyte growth is an important indication of nutri-
ent availability and thus lake trophic condition.

SUMMARY AND CONCLUSIONS

Trophic conditions using Landsat digital data were
used to classify  a total of 2,925 significant lakes in
Wisconsin. A sophisticated set of computer programs
were developed by the University to accomplish mis
task. Field data were collected on about 10 percent of
the lakes to correlate with Landsat parameter values.
Landsat  tapes were obtained during  the  growing
season  (May-October) from 1979-81 for  use in the
classification process.
20.7%
26.4%
25.5%
27.4%
 13.9%
 16.5%
 30.4%
 39.2%
                                Some general conclusions  regarding the  results
                              and Landsat as a technique for lake classification in-
                              clude the following:
                                1. Landsat provides an overall synoptic view with
                              technical measurements being made over all the lakes
                              in the same way. This can be an advantage over using
                              subjective judgments for classification.
                                2. The R2 values (Table 1) relating water condition
                              data to MSS data are reasonable for this type of ana-
                              lysis.
                                3. TSI ranges similar to those predicted from Land-
                              sat were obtained from conventional field sampling.
                                4. Multiple date analyses of Landsat data provide a
                              good first-cut of the water quality of Wisconsin lakes.
                              To collect an equivalent  amount of conventional field
                              data would  be time consuming and expensive.
                                5. The presence of cloud cover as well  as faulty
                              data from Landsat  frequently created problems with
                              analyses of the reflectance data.
                                6. Techniques for correction of  atmospheric  haze
                              have to be improved.
                                7. Predicted values from  Landsat for chlorophyll a
                              and Secchi  disk transparency are reasonably good in-
                              dicators of the trophic condition for any given Wiscon-
                              sin lake. However, lakes  with high levels of color and/
                              or turbidity  are more likely to be biased towards the
                              eutrophic end of the scale.
                                8. Additional research is necessary to determine if
                              satisfactory correlations can be established between
                              satellite observations and lake color and turbidity.
                              REFERENCES

                              Carlson, R.E. 1977. A trophic  state index for lakes. Limnol.
                                Oceanog. 22(2): 361-69.

                              Holmquist, K.W. 1977. The Landsat lake eutrophication study.
                                Unpubl. Master's Degree Rep. Univ. Wisconsin, Madison.
                              Martin, R.H., and K. Holmquist. 1979 Remote sensing as a mech-
                                anism for classification of Wisconsin lakes by trophic condi-
                                tion. Dep. Nat. Resour. Madison, Wis.
                              Martin, R.H., et al.  1983. Wisconsin's Lakes-A Trophic Assess-
                                ment Using Landsat Digital  Data. Dep. Nat. Resour., Univ.
                                Wisconsin-Madison.
                              Ryan, T.A., B.L Joiner, and B.F. Ryan.  1976. MINITAB Student
                                Handbook. Duxbury Press, North Scituate, Mass.
                              Scarpace, F.L, K.  Holmquist, and L.T. Fisher. 1979. Landsat
                                analysis of lake quality. Pages 623-33 in  Photogrammetric
                                Engineering and Remote Sensing. Vol. 45.
                              Uttormark, P.O., and P.J. Wall. 1975.  Lake Classification—A
                                Trophic Characterization of Wisconsin Lakes. EPA 660/3-75-
                                033. U.S. Environ. Prot. Agency, Corvallis, Ore.
                                                   20

-------
NUTRIENTS  IN  CANAL FLOWS  TO LAKE HEFNER,
OKLAHOMA
DALE  W.  TOETZ
Oklahoma State  University
Stillwater, Oklahoma


             ABSTRACT

             Lake Hefner is a terminal offset reservoir located in Oklahoma City and used as a water supply lake.
             The lake is eutrophic and customers frequently complain of tastes and odors in finished water. Dur-
             ing 1980 to 1981, water and nutrient budgets were constructed to learn the causes of eutrophication
             of the lake. Most water and nutrients entered the lake via a 11.6 km canal from the North Canadian
             River. The loadings were regulated so that flow events in the canal coincided with releases of water
             from Canton Lake, which is located upstream on the North Canadian River, or when rain fell on its
             watershed. Discharge and nutrient content were monitored during these flow events. Linear regres-
             sion of concentration of nutrients on discharge showed soluble reactive phosphorus decreased with
             increasing discharge rate  in both North Canadian and Canton flow events. Ammonia, nitrate, and
             Kjeldahl nitrogen increased as discharge rate increased  only in North Canadian flow events  Canal
             discharge did not have typical features of stream hydrographs. Therefore, canal discharge may not
             be usable as an infallible predictor of nutrient concentrations. Estimation of nutrient loading of lakes
             via canals may of necessity remain highly empirical.
INTRODUCTION

Algal growth in impoundments causes water quality
problems because algal respiration and/or decay exert
a  demand for oxygen. Organic production  in  the
epilimnion  may  result in  oxygen deficiency  in  the
hypolimnion and its  attendant increases in reduced
compounds such as ammonia and  sulfide. Further,
algae may impart tastes and odors to the water. Algae
may also promote the growth of actinomycetes, which
also cause taste and odor problems.
  Water from  lakes in Texas and Oklahoma typically
has tastes and odors in the latter part of the summer
as a result  of a bloom of blue-green  algae, which
develops in early July, and a population maxima of ac-
tinomycetes, which follow the blue-green algae bloom
(Silvey et at. 1959). Since the actinomycetes are  not
photosynthetic and  since  organic  substrates  for
microorganisms have a short half-life in water, it can
be assumed that upon decay  or through excretion,
blue-green algae are  a prime energy source for other
microorganisms. Therefore, the ultimate solution of
the taste and odor problem may rest in controlling
blue-green algae.
  This paper reports  part of a research effort to con-
trol algae in a southwestern lake by determining its
nutrient budget. Lake Hefner, Okla., has a long history
of poor potable water caused by midsummer blooms
of algae and actinomycetes (Silvey et al. 1959).
  The lake is fed mostly by a regulated canal, a rather
unusual situation for most reservoirs. The objective of
this study was to determine whether a relationship ex-
isted between nutrient concentration and discharge
rate  in the  canal, which would allow a prediction of
nutrient concentration   based  upon  discharge
measurements alone.
DESCRIPTION OF LAKE

Lake Hefner is a terminal, offset water supply reser-
voir  owned  by Oklahoma  City  and  located  in
Oklahoma County (Fig. 1). It receives most of its water
from the North Canadian River (Fig. 2).
  The lake was constructed on a small stream, Bluff
Creek, by building an earth-and-fill dam of about 5.63
km (3.5 miles) with a clay core. The lake rests on flat-
lying red shale and silt sandstones of the Hennessey
shale (U.S. Geolog. Surv., 1952). Soil thickness around
the margin of the lake is only 30 to 90 cm deep; the
lake thus rests essentially on bedrock (U.S. Geolog.
Surv., 1952). Vegetation  is native grassland or  park
lawn. Most of the watershed is city park.
  Lake  Hefner  has an area of 1,044 ha at 358.5  m
(1,195 ft.) elevation, a mean  depth of 8.86 m  and a
shoreline development of 3.2 (Okla. Water Resour.
Board,  1976). The residence time  for water is  1.96
years, an unusually high value for an artificial lake.
The immediate watershed of the lake is only 1,317 ha.
Much of the flow from the drainage basin is retained in
farm ponds, so runoff into the lake is generally small.
Figure 1.—Lake Hefner, Oklahoma.
                                                 21

-------
 LAKE AND RESERVOIR MANAGEMENT
   The canal dominates all lake loadings. For example,
 from July 1,1980, to June 30,1981, canal discharge in-
 to the lake and precipitation accounted for about 91.6
 and 8.2  percent  of  incoming  water, respectively,
 (Toetz, 1981).
 METHODS

 The observations  were made  during 1980-1981. The
 U.S. Geological Survey provided data on the daily flow
 of canal water into the lake, their measurement sta-
 tion being at point Y (Fig. 1). It was assumed that  all
 water passing  point Y would  enter the lake.  During
 1980, analyses were completed for total phosphorus
 (TP) and total nitrogen (TN) only. TN was the sum of
 nitrate-N  and  Kjeldahl-N. During  1981,  additional
 measurements  were  made  for  soluble reactive
 phosphorus (SRP) and ammonia-N. Samples for SRP,
 nitrate-N and ammonia-N were filtered through rinsed
 Reeve Angel 984 H glass fiber  filters before analyses.
   Analysis for TP, SRP, and Kjeldahl-N followed the
 U.S. Environmental Protection Agency (1974). Analysis
 of nitrate-N and ammonia-N followed Strickland ard
 Parsons (1968) and Solorzano (1969), respectively.
 RESULTS

 Figure 3 shows that a major flow event in the canal
 from Canton Lake occurred between April 7 and 19,
 1980. It was followed by releases from the North Cana-
 dian River during the rest of the  spring and summer.
 Low flow rates occurred from June through August
 1980, preventing fishkills in the canal. Another major
 release of Canton Lake  water occurred Nov. 16-27,
 1980,  but  no  River  releases  occurred afterwards.
 Figure 3 shows that during  1981 a major release from
 Canton Lake occurred between  April 22 and  May (3,
 1981. This was followed by large releases of water
 from the River during the remainder of May and Juno.
                                     t
                                     N
              CANTON LAKE
 NORTH CANADIAN
        RIVER
LAKE HEFNER
    40 Miles
                       Low flows also occurred in the summer and  fall of
                       1981, but no Canton release occurred in late fall.
                         Typically, the water released from Canton Lake is
                       allowed to bypass the canal for several days  before
                       the lake managers take the water. Also, flows from the
                       watershed may be treated in the same way, so there is
                       no hydrograph in either case. In  fact,  flows  in the
                       canal are either very large (100-600 cfs) or very small.
                         I performed correlations of nutrient concentration
                       on daily mean discharge rate. I could not establish a
                       relationship between either total N or total P and the
                       discharge rate for the 1980 data (Table 1). In the data
                       for spring and  fall, Canton  and the  North Canadian
                       River watershed  releases were analyzed  separately
                       from each other.
                         Using 1981 data the analyses were more successful
                       as significant relationships were established between
                       concentrations and  SRP and discharge rate in both
                       the Canton and River releases (Table 2). Strong rela-
                       tionships were also established between ammonia-N,
                       nitrate-N,  and  Kjeldahl-N, and the  discharge rate
                       (Table 2). The relationship was especially good for
                       nitrate.
                         Regression of nutrient concentration on discharge
                       rate was also done in those cases where the relation-
                       ship  was statistically  high.  Table 3  shows  the
                       resulting regression equations. In both Canton and
                       North  Canadian  River releases, SRP decreased  as

                       Table 1.—Correlation between nutrient concentrations and
                        Hefner Canal discharge rate during water diversions into
                          Lake Hefner during 1980, giving values of R2, F and
                                       significance levels.
Canton Lake Releases
April (n = 12)
Total phosphorus
Total nitrogen
November (n = 11)
Total phosphorus
Total nitrogen

R2
0.0179
0.0667

0.3294
0.3247

F
0.13
0.50

3.93
3.37

P
-------
                                                                       WATER QUALITY ASSESSMENT METHODS
discharge rate increased, but the slopes were marked-
ly different. The  slope for River releases  was twice
that of the slope for Canton  releases (- 0.077 versus
 -0.037). The intercepts were  of the same order of
magnitude (292.8 versus 492.0)  and the standard error
or estimates was likewise similar.
   All nitrogen compounds increased in concentration
with discharge rate (Table 3).  However, since all of the
relationships are specific to  the River releases, they
may be indicative of the relationship existing in the
River as well as the canal.
   Table 4 shows that nutrient concentrations in the
canal at point Y (Fig. 1) were the same as those in the
River at the entrance of the canal. Table 4 also shows
measurements made during  a Canton  Lake release.
The data indicate a substantial increase in ammonia,
Kjeldahl-N, SRP, and TP between the River at Watonga
and the canal.
 DISCUSSION
Reckhow (1979) states that phosphorus concentration
decreases  with stream flow in  streams affected by
sizeable  point  sources and increases with flow in
streams lacking major point sources. The concentra-
tion of SRP decreased with  flow in the canal, leading
one  to suppose that the  River was  perturbed by
sizeable  releases  of  P from point sources. Several
small communities upstream between Lake  Hefner
and Canton Lake may be releasing effluents that con-
tribute significant quantities of SRP to the lake at low
flow.
  During Canton releases or high watershed runoff in
the River, dilution  apparently reduces the concentra-
tion of SRP. This scenario seems likely for a semiarid
region such  as Oklahoma, where  sewage plant ef-
fluents are often the major source of stream water dur-
ing drought. Thus, the strategy of admitting water into
the lake only at high  flow seems very appropriate.
  However, only 50  percent of  Canton Lake water
released into the River reaches Lake  Hefner because
of losses to  the alluvium in the river bed (E. Hearn,
pers. comm.). Thus, there may be two good reasons to
transfer  water  in  pipelines  and  not  rivers when
transferring water  from lake to  lake. Water  loss is
avoided and nutrient  loading does not occur. Further,
if the  pipeline is strategically  placed in a receiving
reservoir, destratification could be achieved as well by
inducing turbulent mixing by incoming water.
  Further research is  needed to clarify the quantity of
SRP being contributed  by communities along  the
River  above   Lake  Hefner.  If  sewage  effluent
represents a large fraction of lake loadings and if
release rates  and  natural  runoff events are highly
                       variable,  the  prediction  of  lake  loadings  using
                       algorithms in Table 4 is not apt to  be very useful.
                        CONCLUSION

                        The concentration  of  SRP,  ammonia, nitrate,  and
                        Kjeldahl-N was estimated from discharge rate when
                        the River flowed in the Hefner canal. The concentra-
                        tion of SRP was predicted from discharge rate when
                        Canton  Lake water flowed  in the canal. Therefore,
                        nutrient concentration  could not be predicted from
                        discharge rate in the Hefner canal in  all cases and
                        estimation of nutrient loadings via the Hefner canal
                        may remain highly empirical.
                        Table 2.—Correlation between nutrient concentrations and
                         Hefner Canal discharge rate during water diversions into
                           Lake Hefner during 1981, giving values of R2, F and
                                         significance levels.

                                                    R2         F       P < F
                        Canton Lake Releases
                        (April and May) (n = 20)
                        Phosphorus
                          Total                    0.0139      2.11    0.170
                          Soluble reactive           0.6570     22.98    0.0004
                        Nitrogen
                          Ammonia                0.0260      0.24    0.630
                          Nitrate                   0.0650      0.83    0.380
                          Kjeldahl                  0.0060      0.08    0.780
                        North Canadian River Watershed (n = 20)

                        Phosphorus
                          Total                    0.0278      0.14    0.721
                          Soluble reactive           0.7389     14.15    0.013
                        Nitrogen
                          Ammonia                0.7061     12.01    0.018
                          Nitrate                   0.9216     47.05    0.002
                          Kjeldahl                  0.8521     28.81    0.003


                        Table 3.—Regression equations relating concentration of
                           nutrients and micrograms per liter to Lake Hefner
                           canal discharge (D) as cfs during 1981. (Standard
                                  error of estimate is in parenthesis).

                        Canton Releases
                        Soluble reactive P = -0.037(0) + 292.8(37.3)
                        North Canadian River Watershed
                        Soluble reactive P = -0.077(0) + 492.0(39.6)
                             Ammonia-N  = 0.37(D) + 46.0(20.2)
                               Nitrate-N  = 0.82(0) + 13.8(24.5)
                               Kjeldahl-N  = 8.99(D) + 666.6(316.7)
          Table 4.—Concentration of nutrients in waters entering Lake Hefner as micrograms N or P per liter.
 Site
  Nitrate-N
Ammonia-N
                                                                  Kjeldahl-N
               Soluble
              Reactive P
             Total
               P
 Canal at Lake Hefner
 North Canadian River
   at Hefner Canal


 Canal at Lake Hefner
 North Canadian River,
   Watonga
 Canton  Dam
April 26, 1981

      3
 May 3, 1981
      6

     38
      4
    54

    49


    165

    46
    95
1903

1712


1808

 761
 866
 65

 81


119

 31
 52
334

267


304


 81
136
                                                   23

-------
LAKE AND RESERVOIR MANAGEMENT

ACKNOWLEDGEMENTS: This research was supported by
the Office of Water  Research and Technology, U.S. Depart-
ment  of  the  Interior,  Washington,   D.C.,  Project
A-091-Oklahoma. I thank Mel McFarland, Nyena Vijjeswarapu,
and  Henry Chau for  technical  help, Daryl Walters,  U.S.
Geological  Service,  for providing data on canal  flows, and
the following persons with the Oklahoma City Department of
Water Resources: Betty Fox, Earl Hearn, Jane Webster, and
Bill Criswell.
REFERENCES

Oklahoma Water Resources  Board.  1976.  Okalahoma's
  Water Atlas. Publ. 76. Water Resour. Board.
Reckhow, K. 1979. Quantitative Techniques for the Assess-
  ment of Lake Quality. EPA-440/5-79-015. Office Water Plan.
  Stand. U.S. Environ.  Prot. Agency.
Silvey, J.K.G, J.C. Russell, D.R. Redden, and W.C. McCormick.
  1959.  Actinomycetes  and common tastes and odors. J.
  Am. Water Works Ass. 42:1018-26.

Solorzano, L. 1969. Determination of ammonia in  natural
  waters  by the  phenolhypochlorite  method.  Limnol.
  Oceanogr. 14:799-801.

Strickland, J., and T. Parsons 1968. A practical handbook of
  seawater analysis. Fish. Res. Board Can. Bull. 167.

Toetz, D. 1982.  Nutrient control of blue-green algae in  a
  southwestern  reservoir. Tech. Comple. Rep. Office Water
  Res. Technol.  Project  A-091-Okla. U.S. Dep. Interior.

U.S.  Environmental Protection Agency. 1974. Methods for
  chemical analysis of  water and  wastes. Office Technol.
  Transfer.

U.S.  Geological   Survey. 1952.  Water loss  investigations,
  Lake Hefner studies. Vol  1, Tech. Rep. U.S. Geolog. Surv.
  Circ. 220.
                                                      24

-------
FLORIDA LAKES ASSESSMENT: COMBINING
MACROPHYTE, CHLOROPHYLL, NUTRIENT,
AND  PUBLIC  BENEFIT PARAMETERS  INTO
A MEANINGFUL LAKE MANAGEMENT SCHEME
H.  L. EDMISTON
V.  B. MYERS
Bureau of Water Management
Florida Department of Environmental  Regulation

Tallahassee,  Florida


            ABSTRACT

            Numerous indices developed during the last 10 years to quantify the concept of trophic state turned
            out to be too ambiguous to be useful to lake managers. Based on empirical relationships between
            various water quality indicators, the most widely used indices use Secchi disk transparency, chlorophyll
            a concentration, and total phosphorus as a measure of lake trophic state. Indices based on data from
            north temperate lakes, however, are not directly applicable to Florida lakes. Many Florida lakes are
            known to be nitrogen rather than phosphorus limited. Florida lakes also commonly have macrophyte
            problems which are not accounted for by such indices. A series of indices based on Secchi disk
            transparency, chlorophyll a concentration, macrophyte abundance, total phosphorus concentration,
            and total nitrogen concentration were developed for Florida lakes. The Secchi disk index serves as
            a physical measure of trophic state. The  public perceives water clarity as an important attribute of
            lakes, and Secchi disk transparency is a good measure of water clarity  The biological measures of
            trophic state are based on chlorophyll a concentration and macrophyte coverage. Chlorophyll a con-
            centration is a good indicator of algal populations and macrophyte coverage is related to potential
            aquatic weed problems. Since phytoplankton and macrophytes compete for similar habitats, Florida's
            shallow lakes usually do not experience nuisance conditions of both these autotrophs simultaneous-
            ly. The chemical measures of the trophic state are total phosphorus and total nitrogen concentra-
            tions. These elements are the principal nutrients limiting primary productivity in aquatic systems and
            therefore provide useful information on the nutritional status of lakes.  When the lake is  primarily
            phosphorus limited, based on the ratio  of total nitrogen to total phosphorus concentration,  the
            phosphorus trophic index is used. If the lake is nitrogen limited, based on the concentration ratio,
            the nitrogen index is used. Many Florida lakes are not limited by a single nutrient and are relatively
            well balanced. In this case an average of the nitrogen and phosphorus indices is used to determine
            the trophic status. The overall trophic state index for a Florida lake is determined by combining  the
            appropriate values obtained from the physical, chemical, and biological indices. Over 570 lakes in
            the State were ranked by this method. The relative simplicity of the trophic index,  combined with its
            accuracy and  reliability, emphasizes its utility in statewide lake management.
INTRODUCTION

The Federal Water Pollution Control Act Amendments
of 1972 established a program to help States restore
polluted and  degraded public lakes. Section 314 of
this act, the Clean Lakes Program, required all States
to classify their lakes according to trophic state as
part of the overall  strategy for development of lake
restoration programs. The Florida Department of En-
vironmental Regulation received a Lake Classification
and Prioritization grant from the U.S. Environmental
Protection Agency  in  1981. This grant was  used to
compile lake  water quality data, group  lakes accor-
ding to trophic condition, and develop a management
scheme to prioritize lakes in need of restoration or
preservation.
FLORIDA LAKES DATA BASE

The first step in this  project  was to determine the
most  effective prioritization  scheme  for  Florida's
numerous aquatic systems. A compilation of  major
lake studies in Florida and the data contained therein
was used to rank the lakes. The Florida Department of
Environmental  Regulation  and  the University  of
Florida worked jointly to compile the State's lake in-
formation. The major tasks involved identifying major
water quality data sources for Florida lakes; determin-
ing the  data  compatibility  of the  different  data
sources; organizing the data by studies and by lakes;
computerizing  and  modifying the data  sets into
similar structures with similar variables; putting the
Florida Lakes Gazetter (Fla. Board Conserv., 1969) on
computer tape; and developing  an extensive biblio-
graphy of Florida lakes. When completed, the large
computerized data set contained physical, chemical,
biological, and geographical data on over 800 different
lakes from over 20 different sources.
TROPHIC STATE INDEX DEVELOPMENT

Although numerous methods are available to classify
lakes (Hutchinson, 1957), trophic state is generally ac-
cepted as one of the best and most accurate tech-
niques to describe  the  nutritional  status of lakes.
Many other trophic state indices have  been used in
the past (Carlson, 1977; Shannon and Brezonik, 1972;
Reckhow, 1981) but most are either based  on northern
                                                 25

-------
LAKE AND RESERVOIR MANAGEMENT
lake  data or  required  parameters  not  routine!/
measured. The Carlson (1977) index, probably most
widely used, has a good theoretical basis and relies
on  three  water quality  indicators that are  easil/
understood and quantified. As with other trophic in-
dices, it was derived from data on temperate lakes and
is not directly applicable  to Florida's warm, sub-
tropical lakes. To make it applicable to Florida lakes
the University of  Florida developed a  water quality
trophic  state index based  on Carlson's. The major
modifications include (1) regression relationships for
Florida  lakes,  (2)  inclusion  of  relationships  for
nitrogen-limited and nutrient balanced lakes; (3) bas-
ing the indices on chlorophyll a, and (4) adding  a
macrophyte index.  To assess trophic  state as ac-
curately as possible separate indices  for algal  bio-
mass, macrophyte coverage, Secchi disk transparen-
cy, and nutrient concentrations were considered.
  Algal biomass was selected as the  basis  for the
trophic indices since it is directly related to nutrient
enrichment. Biomass  estimates  are also readily ob-
tained by  measuring  chlorophyll a concentrations.
Other indices for Secchi depth, total phosphorus, and
total  nitrogen  were  derived  after relating  these
variables to chlorophyll a concentration. Log-log data
transformations were used for all regressions in order
to normalize the data and linearize the model.
  The trophic state (Chla) index was developed using
two guidelines: (1) a doubling of chlorophyll a concen-
tration would represent a 10-unit increase in the index,
and (2) an  index value of 60 would be equal to a 20 ^g/l
chlorophyll a  concentration. The equation developed
was:
    TSI(Chla) = 10(1.68 + 1.44 In Chla)
(1)
where Chla is in jug/I and factor 1.68 is simply a scaling
term, and TSI is the trophic state index.
  The TSI for Secchi depth (SD) was developed from
the regression between chlorophyll a concentration
and Secchi disk transparency. Substituting into Equa-
tion 1 yields the following relationship:
    TSI(SD) = 10(6.0 - 3.0lnSD)
(2)
where SD is in meters. From this equation a Secchi
depth of 1.0 m corresponds to a TSI (SD) of 60.
  The development of nutrient TSI's was  somewhat
more complicated  because of the variety of nutrient
limited lakes in Florida. In Florida, diverse  geological
and physiographic  features permit nitrogen as well as
phosphorus to be a limiting nutrient. Smith (1982) had
found that lakes with total nitrogen  to total  phos-
phorus (TN/TP) ratios >30 were primarily phosphorus-
limited and lakes  with TN/TP < 10 were primarily
nitrogen-limited. Lakes with TN/TP ratios between 10
and  30 were  assumed to have a balanced  nutrient
status.  From  these  assumptions, three  separate
nutrient indices were calculated using subsets of the
data determined by TN/TP ratios.
  The  TSI  for   phosphorus-limited  lakes   was
developed from the regression between chlorophyll a
concentration  and total  phosphorus concentration
using only lakes with a TN/TP ratio > 30. Substituting
into  Equation 1:
    TSI(TP) = 10(2.36 In TP - 2.38)
(3)
where TP is in pig/l. From this equation a total phos-
phorus concentration of 34 ^g/l corresponds to a TSI
(TP) of 60. Since lakes with a TN/TP ratio of > 30 are
      considered to  be primarily phosphorus-limited, TSI
      (TP)  is  the  best  nutrient-related  predictor  of
      chlorophyll a concentration for these lakes.
         For nitrogen-limited  lakes (TN/TP < 10), the algal
      biomass relates more closely to total nitrogen concen-
      tration than total phosphorus concentration. The TSI
      for nitrogen-limited  lakes was developed from the
      regression  between chlorophyll a concentration and
      total nitrogen concentration  using data exclusively
      from lakes  having a TN/TP ratio of < 10. Substituting
      into Equation 1:
          TSIfTN) =  10(5.96 + 2.15 In TN)
                                               (4)
      where  TN  is  in  mg/l. From this equation a  total
      nitrogen concentration of 1.02 mg/l corresponds to a
      TSI  (TN)  of 60.  Since  the relationship  between
      chlorophyll  a  and  total  phosphorus concentration
      generally becomes less exact as the ratio  of TN/TP
      decreases, the TSI (TN)  index  should  be  used for
      nitrogen-limited lakes (TN/TP <10) as the best nutrient
      estimator of trophic state.
        For those lakes that have a TN/TP ratio between 10
      and 30 it is not possible to assign a single limiting
      nutrient.  These lakes,  considered  relatively  well-
      balanced with regard to nutrients, respond to changes
      in  loadings and concentrations of nitrogen or phos-
      phorus. Therefore,  both  nutrients  should  be  con-
      sidered and related to chlorophyll a concentration
      fusing subsets of data  with TN/TP  ratios^10 but^SO.
      'Substituting into Equation 1 yields:
    TSI(TNB) = 10(5.6 +  1.98lnTN)

    TSI(TPB) = 10(1.86lnTP - 1.84)
(5)

(6)
where TN is in mg/l and TP is in ^g/l. The proper TSI to
use when TN/TP ratios are between 10 and 30 is:

    TSI (N-Bal) = 0.5 [TSI (TNB) + TSI (TPB)]       (7)

  Although  Carlson  (1977)  strongly  recommended
that his indices be used separately and not combined,
it was felt that for classification and ranking purposes
a combined  index  (single value) was  more useful.
Therefore, the algal  based TSI (WQ) used in this study
was determined by the indices from the three com-
ponents of trophic state.
  The three component indices used were:
          1)  Biological
             TSI (Biol) = TSI (Chla)

          2)  Physical
             TSI (Phys) = TSI(SD)

          3)  Chemical
             a) If TNrTP > 30
               TSI (Chem) = TSI (TP)

             b) If TN/TP < 10
               TSI (Chem) = TSIfTN)
             c)lf 10^-
               TSI (Chem) = TSI(N-Bal)
                                               (8)


                                               (9)



                                              (10)


                                              (11)


                                              (12)
The  algal based trophic state index used to rank
Florida lakes was:

    TSI (WQ) =  1/3 [TSI(Biol) + TSI(Phys) +
    TSI(Chem)]                                (13)
                                                 26

-------
                                                                    WATER QUALITY ASSESSMENT METHODS
  Using this equation and the data in the data base,
580 lakes were initially ranked using only water quality
data. The resulting TSI (WQ) values generally ranged
from 4 to 99 with several extreme outlier values pre-
sent. A more detailed description of the development
of the algal based trophic state index can be found in
Huber, et al. (1982).
  A common problem in Florida lakes is infestation by
aquatic macrophytes encouraged by climate, nutrient
levels, and  the introduction of many noxious exotic
plants into the State. Because of this it was desirable
to quantify  the macrophyte problem and include it  in
the trophic state  index. Several macrophyte indices
have been  proposed (Porcella et al. 1980), although
none have been routinely accepted. Some of the main
problems seem to be quantifying the  macrophytes
and relating a standing crop estimate to water quality
parameters.
  The type of  macrophyte data available for Florida
lakes  was  areal coverage by specific  macrophytes
(Tarver et al. 1979). It should be noted that overlapping
coverages were not identified and the calculated total
coverage in some lakes was greater than 100 percent.
Areal  coverage  is  a  difficult  method to  use  in
estimating biomass, but it seems to be one of the few
presently available on a large scale for Florida lakes.
  The TSI (MAK)  index for macrophytes consists of
three separate subindices which were averaged and
appropriately scaled. The three subindices  were bas-
ed on percent  macrophyte coverage (12 species), per-
cent coverage  of pest species (five species), and per-
cent coverage related  to littoral  zone  (lake cir-
cumference).
  The first subindex used was percent coverage:
         The TSI (MAK) index was then based on the average
      of the three indices:
    MC = 100[(MAC/AREA)/MC(MAX)]
(14)
where MAC is the sum of the acreage of the 12 most
prevalent macrophyte species (Table 1), and AREA is
the lake acreage. Coverage values were scaled by
dividing by the  value of the lake with  the maximum
coverage (MC(MAX)) and multiplying by 100.
  The second subindex was pest coverage:
    MP = 100[(MACP/AREA)/MP(MAX)]
(15)
where MACP is the sum of the acreage of the five
most prevalent pest species in Florida (Table 1). The
same type of scaling calculation used above was ap-
plied.
  Although the total lake coverage of macrophytes
and whether or not they are pests are important, the
location of macrophytes can determine the severity of
the problem. For this reason a third subindex has been
incorporated to relate percent coverage to lake shore-
line:
    M L = 100 [(M AC/V4AREA 7r)/M L(M AX)]      (16)

  It was assumed that each lake was circular. The sur-
face area of each lake was converted to square miles
and a circumference  (V 4 AREA n) in miles was
calculated. The total coverage of macrophytes (sq.
mi.) was then divided by the circumference of the lake
in miles.  The value obtained is roughly half the
distance the macrophytes extend from the shoreline
and is also related to the size of the littoral zone.
These numbers were also scaled from 0 to 100.
  All  three of  these  subindices  were  averaged
together:
    MA = (MC + MP +  ML)/3                  (17)
           TSI (MAK) = 10[1.68 + 1.44 In (MA)]
                                             (18)
  The  algal  water-quality-based  and  macrophyte-
based  trophic indices  were combined into  a final
trophic index (TSI(F)) for Florida lakes. The single TSI
(F) value was thought the  easiest to accept  and
understand by the public and also the best approxima-
tion of trophic state for Florida lakes. The TSI (MAK)
value was given equal  weight because of the large
number of lakes in the  State with aquatic weed pro-
blems. A trophic state cutoff value of 60 was used to
prioritize lakes with problems or potential problems.
This  value corresponds  to a Secchi depth of ^1 m, a
chlorophyll a concentration of ^ 20 ^g/l,  a nitrogen
concentration of^1 mg/l, a phosphorus concentration
of ^34 i*gl\, a macrophyte  coverage of ^100 percent
and a pest coverage of ^75 percent.
  The TSI(F) was calculated as follows:
  1.  If TSI (MAK) or TSI (WQ) was missing, the value
present was used as the final TSI (F).
  2.  If both TSI's were  present:
     a. if either TSI (MAK)  or TSI (WQ) was below 60
and the other was above 60, the greater value was us-
ed for TSI (F).
     b.  otherwise the two values were averaged to
determine the TSI (F).
  Regression analysis was performed for all possible
relationships utilizing the independent variables used
to calculate the trophic state index. Total nitrogen
concentration  was responsible for 50 percent of the
variation in TSI (F). Total nitrogen concentration and
macrophyte cover variables together explained 68 per-
cent of the variation while the addition of the other
three variables explained only an additional 4 percent.
This was probably because total phosphorus concen-
tration, chlorophyll a concentration, and Secchi depth
were all interrelated to total nitrogen concentration
while macrophyte cover was not.

PRiORITIZATION CONSIDERATIONS FOR
FLORIDA LAKES

The scientific endeavor up to this point has been to
separate lakes into two classes:  those with  "poor"
water quality and those  with  "fair to good" water
quality. The separation was accomplished  by utilizing
water quality values that have been used by other in-
vestigators (Carlson, 1977; Shannon and Brezonik,
1972; Reckhow, 1981)  as  a  dividing  line between
eutrophic and mesotrophic lakes. We have been hesi-
tant  to use these terms since  so  many   people
automatically  relate the  term eutrophic to man-
induced problems. Florida  has many eutrophic lakes
that  are not related to man's activities,  but occur
naturally. We have therefore separated Florida lakes
such that only those lakes with  "poor" water  quality
will be prioritized. Prioritization simply implies  a rank-
ing of lakes based upon available data.
  Using the trophic state index developed  earlier, the
list of Florida lakes was reduced to 202  lakes with
"poor" water quality (TSI (F)> 60). A method for further
reducing this list had to be developed so that impor-
tant  nonquantitative factors such  as public interest,
recreational usage, and impaired use could be taken
into account.
  A list of lakes deemed important water resources to
the State was compiled. Lakes on this list  came from
the following sources:
                                                 27

-------
LAKE AND RESERVOIR MANAGEMENT
   1. Top  25 fishing  locations  in  Florida (Fernald,
1981).
   2. Lakes  located in a State Park (Fla. Dep. Nat,
Resour., 1981).
   3. Lakes  located in Fish Management  Areas and
Wildlife Management  Areas (Fernald, 1981).
   4. Lakes  located in National  Forests, Wilderness.
Areas, Wildlife Refuges,  Recreational Trails System,
and Rare II study areas (Fla. Dep. Nat. Resour., 1981)
   5. Lakes  located in State Forests,  Preserves, anc
Wilderness Areas (Fla. Dep. Nat. Resour., 1981).
   6. Lakes used as potable water supplies (Fla. Dep
Environ. Reg., 1983).
   These lakes were felt to  represent lakes that were
well known  to the public and were also used  more'
often than lesser known lakes. There were 102 lakes in
this category.
   Recreational  usage factor  is difficult to quantify,
Because of the long coastline in Florida more people
use the ocean, gulf, and coastal areas for recreation
than  in landlocked  States.  For this  reason,  boat
registration  data cannot be used.  Fishing licenses.
however, are required only for freshwater fishing in
Florida. Although this type  of data includes river and
stream fishing and excludes swimming, boating, and
other recreational uses it gives an indication of lake
recreational usage.
   Because of the large number of tourists who visit
the State every year, nonresident data were also used.
Information  was collected on a county basis, for the
year 1981, and  came  from  the files of the following
agencies:
   1. Resident and  nonresident  fishing licenses  —
Florida Game and Fresh Water Fish Commission
   2. County  populations  — University of  Florida,
Bureau of Economic and  Business Research.
   3. Florida tourist statistics —  Florida Division of
Tourism, Office of Marketing Research.
   From this data, a  recreational usage factor was
calculated for each  county  using the following equa-
tion:
    RECUSEA = FR/CP + FNR/T
(19)
where FR is the resident fishing licenses sold in the
county,  FNR is the nonresident licenses, CP is the
county population, and T is the number of tourists
estimated to have visited the county.
  The interest the public displays in a lake is inter-
related  with recreational usage,  population living
nearby, and condition of the lake. Two methods have
been derived to assess public  interest in order to in-
clude this important factor in ranking Florida  lakes.
  The  first  method  used direct public  response to
questionnaires sent to lake associations, environmen-
tal groups, concerned citizens,  and State and local of-
ficials. Results of a public participation process were
also used. This included lakes requested to  be con-
sidered  for  restoration after review of a preliminary
restoration list.
  The second method devised  to determine public in-
terest was more indirect. In  many lakes, public in-
terest can be traced to a large population that  lives on
or near the lake.
  An equation was developed:
    PI = Urban/fTotland - Water)
(20)
        tion, the percent of the watershed that is developed
        can be approximated.  Data used  in this  equation
        came from the Florida Department of State Planninq
        (1979).
          Impaired use factors are very important in any lake
        restoration ranking scheme since the public can easi-
        ly see fish  kills  or lakes  closed to  swimming and
        fishing. Relating impaired use to actual causes, how-
        ever, is difficult. Two methods were devised to include
        impaired use of lakes in the ranking scheme. A list of
        lakes in which fish kills were reported to EPA was ob-
        tained from STORET files for 1960 through 1980. This
        list was used later in ranking Florida lakes.
          Another method devised to indicate impaired use in-
        volved relating "poor" water area to total water area.
        Lakes with TSI (F) values greater than 60 were grouped
        by county and their acreages summed by county. This
        "poor" water acreage was divided by the total acreage
        of all named lakes in that particular county.
          While trophic state ranking is an essential aspect of
        any lake management program, nutrient loadings can
        also be used to pinpoint potential problem lakes. The
        concentration  of  nitrogen  and  phosphorus in  lakes
        depends on various factors including watershed land
        use characteristics, basin geology, hydrology, physio-
        graphy, and climate of an area. Inputs from land uses
        provide the  majority  of  nutrient loadings  to  most
        Florida lakes.
          To determine nutrient loadings, it was necessary to
        have accurate and comprehensive land use data. Data
        compiled by the Division of State Planning in conjunc-
        tion with  the Florida Department of Environmental
        Regulation (Dep. State Plan., 1979) was used for this
        assessment. Land use data for 309 Florida lake water-

               SECCHI  DISK VS. CHLOROPHYLL A
                    ... .  x
       Figure 1.—Secchi disk depth (m) versus chlorophyll a con-
       centration (pig/I).

         Table 1.—List of species used in macrophyte index for
                  coverage and pests calculations.
where Totland is the lake's watershed in acres, Water
is the area of the lake, and Urban is the area of the
watershed classified as urban land. From this equa-
Common Name
Floating water-hyacinth
Hydrilla
Fragrant water lily
Maidencane
Spatterdock
Bladderwort
Cattail
Knotgrass
Torpedo grass
Southern naiad
Eurasian watermilfoil
Pickerelweed
Species Name
Eichhornia crassipes
Hydrilla verticillata
Nymphaea odorate
Panicum hemitomon
Nuphar luteum
Utricularia sps.
Typha sps.
Panicum geminatum
Panicum repens
Najas quadalupensis
Myriophyllum spicatum
Pontederia lanceolata
(Pest)
*
*

*





*
*

       'One of five species considered a pest and used in macrophyte pest coverage
       sub-index
                                                 28

-------
                                                                        WATER QUALITY ASSESSMENT METHODS
sheds  was  available in five categories:  urban,  agri-
cultural, wetlands, forest, and water area.
  After accurate land use data were obtained, nutrient
export coefficients were needed for each  type of land
use. Nutrient export coefficients were used to relate
land use and nutrient fluxes to water quality problems.
Export  coefficients  are  difficult   to   accurately
calculate  because of variability  caused  by regional,
climatic, geological, hydrographic, and seasonal fac-
tors.  However, two authors (Baker et  al.  1981;
Shahane,  1982) have reviewed Florida-based loading
data and the median export values were  used in this
study. Table 2 lists the export coefficients used for the
various land   uses  and  also  coefficients for at-
mospheric input. Data on point source discharges to
lakes were also compiled.
   Using the above data, mass loadings to lakes were
calculated using the formula:
M  = (AcxA) + (UcxU)
     (AtcxS) + PSI
                             (WcxW) + (FcxF) +
                                                (21)
where:

     M      =  Mass loading (g/yr)
     Ac     =  Agricultural export coefficient (g/m2/yr)
     A      =  Agricultural land area (m2)
     Uc     =  Urban export coefficient (g/m2/yr)
     U      =  Urban land area (m2)
     We    =  Wetland export coefficient (g/m2/yr)
     W      =  Wetland land area (rn.2)
     Fc     =  Forest export coefficient (g/m2/yr)
     F      =  Forest land area (m2)
     Ate    =  Atmospheric input coefficient (g/m2/yr)
     S      =  Surface area of lakes (m2)
     PSI    =  Point source inputs (g/yr)

   The mass  loading values for  nitrogen and phos-
phorus  were then divided by the lake surface area to
normalize the loadings  (g/m2/yr). These  values were
then used to rank the lakes based on highest loadings.
Some of the loadings were  unrealistically high, there-
fore phosphorus loadings greater than 10 and nitrogen
loadings  greater than 200 were eliminated from con-

    Table 2.— Nutrient loading coefficients for Florida.
                   Loading, g/m^/yr.
Activity
                       TNL
TPL
Agriculture
Urban
Wetlands
Forest
Atmospheric
                       2.06
                        .57
                        .55
                        .22
                        .75
0.67
 .082
 .025
 .032
 .051
       Table 3.—Point system for lake prioritization.
 Parameter
                                        Points
Trophic state index
Recreation index
Public interest
  a) Questionnaire return
  b) Public interest index
  c) Requested lake
Impaired  use
  a) Fish kills
  b) Percent degraded water
Nutrient loading index
Important public waters
          Total
                                        0-40
                                        0-10

                                        0,5
                                        0-5
                                        0,5

                                        0,10
                                        0-10
                                        0-10
                                        0,10
                                        0-105
                                                      CHLOROPHYLL A VS. TOTAL PHOSPHORUS
                                                    LNCHLA
                                                  Figure 2.—Chlorophyll a concentration O^g/l) versus total
                                                  phosphorus  concentration  (mg/l)  for phosphorus  limited
                                                  lakes (TN/TP ratio > 30).

                                                        CHLOROPHYLL A VS. TOTAL NITROGEN
                                                                             a s>

                                                                             LNTN
                                                  Figure 3.—Chlorophyll a  concentration (iug/l) versus total
                                                  nitrogen concentration (mg/l) for nitrogen lakes (TN/TP ratio <
                                                  10).

                                                       CHLOROPHYLL A VS. TOTAL  PHOSPHORUS
      -54   -48   -42   -36   -30   -24   -18   -12
                          LNTOTP

Figure  4.—Chlorophyll  a  concentration (wg/1)  versus total
phosphorous concentration (mg/l) for nutrient balanced lakes

dO^JB ^30).


        CHLOROPHYLL A VS. TOTAL NITROGEN
                ^*— -;,,-—v
                                    LNTN
         Figure 5.—Chlorophyll a  concentration (/jg/l) versus total
         nitrogen concentration (mg/l) for nutrient balanced lakes
                                                    29

-------
  LAKE AND RESERVOIR MANAGEMENT
  sideration  since few lake loading predictions occur
  above these values (Baker et al. 1981).
    The lakes were  ranked  separately by nutrient and
  the values  scaled from 0 to 1 by dividing each lake by
  the maximum loading rate for each nutrient. If a lake
  was nitrogen-limited, nitrogen  loading  ranking was
  used. Phosphorus  loading  ranking was used for phos-
  phorus-limited  lakes.  For well-balanced  lakes  an
  average of the two rankings was used.
  FINAL RANKING OF FLORIDA LAKES

  After the individual indices were developed and values
  for each determined, a system was devised for awar-
  ding points to individual lakes. The maximum score
  possible was 105 points. All lakes with TSI(F) values
  greater than 60 were included in the trophic state rank-
  ing. The values were scaled from 0 to 40 so that the
  maximum point award in this category was 40 points.
  Trophic  state,  being  the most  important,  was
  weighted much heavier than the other indices. Most of
  the other indices were scaled from 0 to 10 so that a
  maximum point award of 10 was possible. The lakes
  on  the four lists that were compiled (questionnaire
  returns,  fish kills,  requested lake,  and  important
  public waters) were given a set number of points for in-
  clusion on  each list. The indices were treated  in-
  dependently and composite rankings were determined
  by  summing the points awarded in each category.
                             Table 3 lists the parameters used  and the range of
                             possible points awarded in each category.
                               After point awards were calculated, the lakes were
                             ranked for restoration consideration. The list was ar-
                             bitrarily reduced to the 50 top ranked lakes. The loca-
                             tion of these lakes can be found in  Figure 1.
                               It was gratifying to note that most of the lakes listed
                             are  also those that  most knowledgeable  people in
                             Florida would  choose for restoration purposes. The
                             majority of  these  lakes are  affected  by cultural
                             eutrophication, either point  sources or urban-related
                             nonpoint source discharges. Most of these lakes are
                             also located in the central Florida region, which has
                             experienced an explosive population growth over the
                             last 20 years.
                               This paper has  attempted to devise a method for
                             prioritizing Florida  lakes in  a scientific manner. The
                             method developed could  be used by other States with
                             modifications to fit  information  available to them.
                             Perhaps the most important aspect of any method is
                             to eliminate guesswork and  bias in the selection pro-
                             cess, which we feel has been accomplished.
                            REFERENCES

                            Baker, LA., P.L. Brezonik, and C.R. Kratzer. 1981. Nutrient
                              Loading-Trophic State Relationships in  Florida  Lakes.
                              Publ.  56, Water  Resour.  Res.  Center,  Univ.  Florida,
                              Gainesville.
              87°
   31
   30°
   29°
  28°
  2T
86"
85°
84°
                                                                                            80e
                     10
                     {}
 STITH
 TS»l» IP
 CUT
 FPANris
 TM}wnTrS
 ^FiLflA 1
 orm
 EU'TTS
 SPIF F IN
  'III U«105
  HIUSIFHUT.
  LIKE
  LVKE
  LUE
  LIRE
l
7
1

0
1

i
<.
3
*>
7
FP..KIS
TXEE^HIHCC
APTPf A
r A9 L T ON
FAIPVIFU
LA JNF
fi[TLA*'n
H«I:«INF M
"APJAN
FPHTPFML K,A
SF-ilNntE
C01NIF
f FFIF
-"A1C1C"
MPLt I NC-S^IP TM
MOW ACp
'(IINI F5
LULU



1CFTY
3r>1IT
S H I P P


S Tc L 1 A
'JHEECHJ1EE
n« » H 5 E
OR A*JGE
1R AN3E
W « t G E
3» »1-,E
15CE1H
1SCEOH
o>;enL A

•Nf Ll«S
POM
POL*
POL<

POL<
POL<
»OL<

PUL<
POK
POL<
POL1
POL<

POL<
PJ H A1
PUI^AI
Figure 6.—Location of top 50 lakes in need of restoration.
                                                 30

-------
                                                                               WATER QUALITY ASSESSMENT METHODS
Carlson, R.E. 1977. A trophic state  index for lakes. Limnol.
  Oceanogr. 22: 361-69.
Fernald, E.A., ed. 1981. Atlas of Florida. Florida State Univ.
  Foundation, Inc., Tallahassee.
Florida Board  of Conservation. 1969. Florida Lakes, Part III
  Gazetteer. Div. Water Resour., Tallahassee.
Florida Department  of  Environmental  Regulation.  1983.
  Water quality standards. Chap. 17-3, Fla. Admin. Code.
Florida Department of  Natural Resources. 1981. Outdoor
  Recreation in  Florida -1981. Div. Recr. Parks, Tallahassee.
Florida Division of State Planning.  1979. Construction of a
  Land Characteristic Data Base  for Comprehensive Water
  Management  Planning. Info. Systems Section, Tallahas-
  see.
Huber, W.C., P.L Brezonik, J.P. Heaney, and R.E. Dickinson.
  1982. A classification of Florida lakes. Rep. ENV-05-82-1.
  Dep.  Environ.  Eng. Sci., Univ. Florida, Tallahassee.
Hutchison,  G.E.  1957.  Eutrophication.  Am.  Sci.   61(3):
  269-79.
Porcella, D.B., S.A. Peterson, and D.P. Larsen. 1980. In Index
  to Evaluate Lake Restoration. Proc. Am. Soc. Civil Eng., J.
  Environ. Eng. Div. 106: 1151-69.
Reckhow, K.H. 1981. Lake Data Analysis and Nutrient Budget
  Modeling.  EPA-600/3-51-011. U.S. Environ.  Prot. Agency,
  Corvallis, Ore.
Shahane, A.M. 1982.  Estimation of pre- and post-develop-
  ment nonpoint water quality loadings. Water Resour. Bull.
  18(2): 231-37.
Shannon, E.E.,  and  P.L  Brezonik. 1972.  Eutrophication
  analysis: a multivariate approach. Proc. Am. Soc. Civ. Eng.
  J. San. Eng. Div. 98(SA1): 37-57.
Smith, V.H., Jr. 1982. The  nutrient  and  light dependence of
  phytoplankton productivity. Ph.D. Thesis. Univ. Minnesota,
  Minneapolis.
Tarver, D.P., J.A. Rodgers and M.J. Mahler. 1979.1979 Florida
  Aquatic Flora Surv. Rep. Dep. Nat. Resour.,  Bur. Aquat.
  Plant Res.  Control, State of Florida, Tallahassee.
                                                         31

-------
 TOOLS  FOR ASSESSING  LAKE EUTROPHICATION
 IN THE  PUGET SOUND  REGION,  WASHINGTON
 ROBERT J.  GILLIOM
 U.S. Geological Survey
 Reston, Virginia


             ABSTRACT

             Assessment of eutrophication of lakes in watersheds undergoing development is facilitated by estimates
             of (1) background phosphorus (P) loading and concentration, (2) present-day P concentrations and
             amounts and sources of P loadings in excess of background levels, (3) the sensitivity of the lakes
             to future increases in P loading, and (4) relationships between P concentration and other factors that
             determine lake water quality. Methods have been developed for making such estimates for lakes in
             the Puget Sound Region based on data already available for most lakes in the region. Background
             P loadings were computed from P concentration data for 24 undeveloped lakes in the region using
             a mass balance model, and predictive relationships were developed from these loadings to estimate
             background loading for other lakes. The standard error of estimate for background loadings and con-
             centrations averages 25 percent for most lakes in the region. Present-day P loadings were then estimated
             from measured P  concentrations for 28  lake:>  in basins  containing  residential land. Differences
             between present-day and background loading were attributed to land use changes. P loadings from
             septic systems, computed as the  difference between total present-day  loading and the sum  of
             background and residential runoff loading, were f Dund correlated with the presence of old homes around
             the lakes (r2 = 0.36). The regression relationship expressing this relation can be used to estimate
             septic-system loading for other lakes. If necessary, P loading from agricultural land can then be estimated
             on a lake-by-lake basis as the difference between total present-day loading and the sum of background
             and both types of residential area loading. Methods are presented for estimating the reliability of all
             estimates, which varies. These empirical relations allow approximations of the cumulative impact that
             development has had on P loading and the amounts of loading from generalized land use categories.
             The mass balance P  model also estimates lake sensitivity  to future increases in P loading Finally,
             predictive relationships were developed between P concentrations and Secchi disk transparency and
             chlorophyll a concentrations, two key indicators of lake water quality related to eutrophication, based
             mainly on data for 17 well-studied lakes in the region.
 INTRODUCTION

 Management of lake eutrophication in regions con-
 taining many lakes presents some unique difficulties
 that are not often encountered in the same severity as
 when  attention  is focused  on one or  a few lakes,
 Foremost,  detailed study of  each lake with water
 quality problems is not economically feasible.  The
 Puget Sound region contains more than 500 lakes that
 have a wide range of nutrient levels and land use set-
 tings.  For  this  region  and  other lake-rich  regions,
 methods are needed that use existing data or minimal
 new data to assess the quality of lakes, the unique
 cause  and  effect relationship between water quality
 and local land uses for  individual lakes, and the sen-
 sitivity of a lake's water quality to future changes in
 nutrient loadings. These assessments enable lakes to
 be prioritized so that limited resources for investiga-
 tion and management may  be best used to design
 detailed studies of individual  lakes.
  Following is a brief summary and discussion of the
 derivation  and  application  of  methods  recently
 developed  for assessing lake eutrophication  in .the
 Puget  Sound region.  The main  components of the
 eutrophication assessment procedure are (1) predic-
tive relationships between P (phosphorus) concentra-
tions and Secchi disk transparency and chlorophyll a
 concentrations   (Gilliom and  Bortleson,  1983);  (2)
 evaluation  of lake sensitivity  to  P loading (Gilliom,
 1982a), and (3) methods for estimating the amounts of
 P loading under background (pre-development) condi-
tions (Gilliom, 1981) and from major land use sources
 under present-day conditions  (Gilliom,  1982a). The
 reader is referred to the references cited for details
 not covered in this short space.


 PHOSPHORUS CONCENTRATION AND
 LAKE WATER QUALITY

 The  first questions  that  arise when considering a
 regional  scale evaluation of  lake eutrophication are
 often:
   1.  Is Pthe limiting nutrient?
   2.  If so, what type of P measurement is most impor-
 tant to evaluate (dissolved or total, summer or winter)?
   3. Are there consistent relationships between  P
 concentrations and measures of water quality effects,
 such  as  Secchi  disk transparency and  algae levels,
 and what are those relationships?
   Both N:P ratios and  correlations  between chloro-
 phyll  a  concentrations and  various seasonal  and
 chemical measures of both N (nitrogen) and P indicate
 that  P is the limiting  nutrient in almost all Puget
 Sound region lakes. For 73 lakes, the median  N:P ratio
 by mass for total N and total P in the epilimnion during
the summer was 24:1, with only one lake having a ratio
 less than 5:1 and only four with ratios between 5:1 and
 10:1. The theoretical critical value for the break be-
tween N  and P limitation is 7:1. Furthermore, mean
summer Secchi disk transparency and mean summer
chlorophyll a levels, both indicators  of water quality
                                                  32

-------
                                                                        WATER QUALITY ASSESSMENT METHODS
during the season of greatest  recreational use, are
more highly correlated with mean summer total P in
the epilimnion waters of region lakes than are other
measures of P or N.
  The quantitative  relationships  between summer
total  P concentrations  (STP) and both Secchi disk
transparency and chlorophyll a concentrations during
the summer are shown in Figures 1 and 2. These rela-
tionships were derived mainly from data for 17 inten-
sively studied  lakes  in the region—the data are ex-
plained  in  detail in  Gilliom and Bortleson  (1983).
Figure 1 shows the regression  relationship between
chlorophyll a and total P. Total  P explains 76 percent
of the variance in chlorophyll a; the standard error of
prediction is about 5 ^g/l. Figure 2 shows the regres-
sion  relationship  between Secchi disk transparency
and summer total P for P concentrations greater than
15/43/1. The regression equation  explains 49 percent of
the  variance in Secchi disk. The standard  error of
prediction is about 1 m.
   The  relationships  in  Figures  1  and  2  enable
estimates of Secchi disk transparency and chlorophyll
a levels for a particular measured or  predicted mean
summer total P concentration  for any  lake  in the
X cc
  t
o<
rr cc
O O
T°
X DC
00
tt 5
til *
5Z
  ~
      40
      30
      20
      10
          	1          i          r~

                      EXPLANATION
           	 SCHLA = 0 42 • STP-2.0 (percent variance   /
                      explained: r1 = 0.76)         /
           	Standard errors of predictions
            SUMMER TOTAL PHOSPHORUS (STP),
                IN MICROGRAMS PER LITER

 Figure 1.—Relationship between summer chlorophyll a and
 summer total phosphorus concentrations.
                                                     region falling within the range  of  P concentrations
                                                     represented.  For more general  lake water  quality
                                                     ratings, Table 1  was developed from Dillon and Rlgler
                                                     (1975), the regression relationships just explained, and
                                                     the authors' experience in the region.
                                                      RELATIONSHIP BETWEEN PHOSPHORUS
                                                      CONCENTRATION AND LOADING

                                                      A predictive relationship between P concentration and
                                                      P loading links the effects and causes of eutrophica-
                                                      tion and indicates the sensitivity of lake P concentra-
                                                      tion to future changes in loading. An  effective and
                                                      commonly applied means of relating P concentration
                                                      to P loading is a simple mass balance model of P on
                                                      an  annual  time  scale.  Several versions  of  these
                                                      0    8
                                                      (/)
                                                      O
                                                       o
                                                       z
(fl
Z V)
 z
9-
z
o
O
IU

CC
Ul
5
5
               	1	1	1	

                    EXPLANATION
               SSD = 4. -0.05 STP (percent variance
                                                                          explained: r2 = 0.49)
                                                                	Standard errors of predictions
                                                             15   20
                                                                        30
                                                                                 40
                                                                                        SO
                                                                                                 60
                                                                                                         70
                                                             SUMMER TOTAL PHOSPHORUS (STP),
                                                                  IN MICROGRAMS PER LITER

                                                      Figure 2.—Relationship between summer Secchi-disk trans-
                                                      parency and summer total phosphorus concentration  for
                                                      lakes with phosphorus concentrations of 15 ng/l or greater.
    Table 1.—Characteristic relationships between lake-water phosphorus concentrations, and general lake-water quality.
     Water-quality
        group
                                                     Group characteristics
  STP = 0 to 10
          B
  STP =  10 to 20
  STP = 20 to 30 /ig/l
   STP greater than
        30 jig/l
                    Low algal productivity; high suitability for all recreational uses. Algal blooms are rare and the
                      water is extremely clear, with a Secchi disk transparency that is usually 5 m or greater.
                      Summer chlorophyll a concentrations generally average less than 3 ngl\.
                    Moderate algal productivity; generally compatible with all recreational uses. Algal blooms are
                      occasional, but generally of low to moderate intensity. Oxygen depletion is common in the
                      bottom waters and coldwater fisheries may be endangered in some shallow lakes. In
                      many lakes, however, the fishery may be enhanced by the increased productivity. Secchi disk
                      transparency is usually 3 to 5 m, and chlorophyll a averages 2 to 6 ^g/l in most lakes.
                    Moderately high algal productivity; still compatible with most recreational uses, but algal blooms
                      are more frequent and intense, and oxygen depletion is more serious. This can increase fisheries
                      problems, though productivity may still be enhanced. Water is often somewhat murky and
                      Secchi disk transparency is usually 2-4 m. Chlorophyll a usually averages 4-12 ^g/L
                    High algal productivity; lake suitability for most recreational uses is often impaired by frequent and
                      intense algal blooms which may form floating scums. The water often takes on a "pea soup"
                      color resulting in extremely murky water. Fish kills may be common because of depleted oxygen
                      especially in shallow lakes. Secchi disk transparency is generally less than 3.0 m, and
                      chlorophyll a concentration is usually greater than 10 /jg/l.
                                                     33

-------
 LAKE AND RESERVOIR MANAGEMENT
 "Vollenweider-type"  models and  their validity  and
 ranges of applicability were reviewed by Reckhow
 (1979). The model chosen for this investigation  was
 described by  Larsen and  Mercier (1976). The  model
 may be written as
     (Pjco
                - R)
 where
(P)oo

L
R

Q
               Q
           is the long-term mean concentration of
           total  P in the lake, /^g/l,
           is the total-P loading to the lake, kg/yr,
           is the lake-retention coefficient, dimen-
           sionless,
           is the annual average flow through the
           lake,  (km2.m)/yr,
 and R is approximated by
               1
     R =
                                              (2)
 Flushing rate, p, is the annual flow through the lake
 divided by lake volume and can be estimated from
     P =
           WSA-RO
               V
 where WSA is watershed area, including lake surface,
 RO is average annual runoff, and V is lake volume.
 LAKE SENSITIVITY
The factor -
          1 -  R
            Q
           in equation 1 is a constant for a par-
ticular lake because it is comprised of the unchanging
average values for the estimated P retention coeffi-
cient and the annual flow through the lake. Equation 1
can be expressed in abbreviated form as
    (P)°° = S.L

where
    S =
     d - R)

        Q
                                         (4)
                                             (5)
and S is termed the "lake sensitivity coefficient." The;
value of S is the predicted change in (P)°° that would
be caused by a I  kg change in annual average P
loading. For most region lakes the standard error in S
SES,  appears to be about ± 20 percent. For regiona
scale analyses, the sensitivity coefficient provides a
simple and inexpensive means of comparing  many
lakes. Lakes can be ranked based on their relative sen
sitivity to increased loading, and these ranks can be
used to help set priorities for research and manage-
ment.

ESTIMATION OF PHOSPHORUS LOADING
FROM DIFFERENT LAND USES

The most difficult aspect of eutrophication to assess,
and yet the most important to understand for devising
 control strategies, is P loading from different nonpoint
 sources, usually basin areas with different land uses.
 Commonly, one of three general approaches is used
 to estimate land use contributions of P to a particular
 lake:  (1) direct  measurement, (2) estimation  from
 measurements  made for  similar  land  uses in the
 region of the lake, or (3)  estimation from data col-
 lected for various land uses over a wide geographic
 area,  such as the entire United States or a major part
 of it.  The reliability of estimated P loading generally
 decreases from (1) to (3). Since direct measurements
 of loading are impractical if many  lakes are involved,
 most  regional  evaluations are best accomplished
 using  empirical  relations  between  land  uses and
 loading for the region where the lakes are  located.
  Ideally, to develop regional land use/P loading rela-
tionships, one needs data on loading from similar land
uses in a variety of hydrologic and geologic settings in
the region. But, reliable and consistent measurements
of P loadings from various land uses are not available
for many regions, including the Puget Sound. An alter-
nate approach was used in this study to estimate land
use loadings indirectly from lake water P concentra-
tions  using the  mass  balance  P model  described
earlier.


ESTIMATION  OF PHOSPHORUS LOADING
FROM MEASURED CONCENTRATIONS

Phosphorus  loading  estimates  were  made  from
measured lake  water P concentrations, which  are
much  easier to  obtain than loading measurements.
Concentration data are already available for  many
lakes. Loading can be computed from concentration
using  a simple rearrangement  of equation  4. Though
equation  4 is expressed in terms of the whole lake
mean  concentration  of total  P,  that term can be
replaced  with any other type of phosphorus measure
as long as (1) there is a consistent proportional rela-
tionship between (P)°° and the alternate measure, and
(2) the loading term is redefined.  As demonstrated
earlier, mean summer epilimnion total P, (P)ss, is the
most effective focus for eutrophication assessment in
the Puget Sound region. It is also correlated with both
annual and winter-spring mean total P and appears to
average about 83 percent of (P)°°. Equation 4 was thus
redefined:
                                                       (P)ss = L*-S
                                                                                            (6)
                                                   or
                                                  L* =
                                                              (P)s
                (7)
                                              where  L*  is an  empirically derived P loading rate
                                              which  is  generally  lower  than the actual  total P
                                              loading. The standard error of L* can be calculated by
                                              use of equation 8
           /SE(P;
    SEL. = (	
           \   S2
SE(P)SS2 + (P)SS2.SES2 1/2

   S2          S4
                                                                                           (8)
                                                By estimating P loading to a lake using equation 7,
                                              the lake is,  in  effect,  used  as a  time-integrated
                                              sampler of  P loadings to the lake from all sources.
                                              Such loading esimtates are in one sense better esti-
                                              mates  of long-term average P loading than actual
                                                34

-------
                                                                    WATER QUALITY ASSESSMENT METHODS
measurements  of  loading  over  only  a short time
period, such as 1 year, because lake P concentrations
reflect to some extent the variable history of loadings
to the lake. Furthermore, only loadings of P that affect
summer epilimnion  concentrations are incorporated
into  L*. In this  sense, L*  can be called  "active P
loading" and it probably is most heavily influenced by
loads of soluble, biologically available P that occur in
late  winter. For example, large  loads  of  sediment-
related total P resulting from storm flows may make
up a large portion of the total P load but have a small
effect on  summer  epilimnion P  concentrations and
thus not be an important part of L*.
  Sources of P loading to a lake were considered in
two categories: background (predevelopment) sources
and  cultural (human-related)  sources. All cultural
sources were evaluated according to  the  increases
above background loading levels that they produce. In
mathematical terms,  phosphorus loading  to a  lake
can be described as follows:

     L* = (PREL.A)  + (FORY-WSAbg) + ARR +
          AWW + AAG,                         (9)
where
    PREL

    A
    FORY

    WSAbg

    ARR, AWW,
    and AAG
is the areal rate of loading by
precipitation, (kg/km2)/yr,
is lake area, km2,
is the yield of P from forested
areas, (kg/km2)/yr,
is the area of land in the lake's
drainage basin

are the incremental increases in
loading about background levels
attributable respectively to:
residential area runoff, nearshore
septic tank systems,  and  agricul-
tural land, kg/yr.
  Each  loading  and yield term  in equation 9  was
evaluated by steps, proceeding from left to right. Addi-
tional terms assessed by Gilliom (1982a), but not in-
cluded here, described  loading  contributions from
upstream lakes.  The approach  was to (1) estimate P
loadings  from measured concentrations  for undevel-
oped lakes in the region and use these estimates to
develop an empirical relationship for estimating back-
ground loading to other  lakes in the region; (2) com-
pare predicted background loadings  for developed
lakes, with the present-day loading estimated from
measured P concentrations to evaluate the amount of
loading caused by development; and (3) evaluate the
residual  loadings  attributed  to   development for
selected  groups  of lakes with only  certain land uses
present in order to develop regional relationships be-
tween loading rates and  specific P sources.
Background Sources of Phosphorus

For lakes with no significant development in their
drainage basins, equation 8 reduces to:
    L*bg = (PREL.A) + (FORY-WSAbg)
                             (10)
where L*bg is the loading from background sources as
calculated from  the measured P concentration  in a
lake using equation 7. There are two unknowns, PREL
and FORY, in equation 10. To solve the equation for
                                   FORY  for individual undeveloped  lakes, a regional
                                   average value of PREL was first  determined from
                                   published data for the region and was assumed to be
                                   a constant for the study area:
                                        PREL = 20(kg/km2)/yr
                                             (11)
                                   The uncertainty in PREL was not directly assessed,
                                   but is included as a source of uncertainty in estimates
                                   of loading from forest land.
                                      The P  yield  from forest land, FORY,  can be
                                   calculated by difference for any lake with a forested,
                                   undeveloped drainage basin by use of equation 12.
                                        FORY =
                                                  L*bg - (PREL.A)
                                             (12)
                                                  WSA
                                                      bg
                                    FORY was calculated for 24 undeveloped lakes and
                                    found to be highly correlated with annual runoff in the
                                    vicinities of the lakes (Fig. 3). The following regression
                                    equation, which explained 73 percent of the sample
                                    variance, can be used to estimate FORY for any lake
                                    that is in a locality of the Puget Sound region that has
                                    an annual runoff approximately in the range of 0.1 to
                                    1.5m.
    FORY = 7.1-lnRO + 16.6.
(13)
Standard errors are shown in Figure 3.
  With estimates of PREL and FORY, the total back-
ground  loading of P  for a  developed  lake  can be
calculated using equation 10. The standard  error of
this estimated background loading can be calculated
using equation 14.
                                                        SEL*bg = SEFORY -WSAbg,
                                             (14)
whre  SEL*bg is the standard error  of the  loading
estimate. The uncertainty in values  of  FORY from
equation 13 is the result of the combined variability in
all model terms (including (P)ss and PREL) and model
error. The uncertainty in  all  model terms and in the
                                     25 -
                                   CL
                                   O
                                   <  15 -
                                                    IE
                                                    O
                                                    Li.

                                                    5
                                                    O
                                                    tL
                                      10 -
                                                                Standard error of estimate
                                                                for one particular lake. SE FORY
                                                  0.5          1.0          1.5

                                                    ANNUAL RUNOFF (RO).INm
                                                                                                     20
                                    Figure 3.—Relationship between phosphorus yield from
                                    forest land and annual runoff.
                                                 35

-------
 LAKE AND RESERVOIR MANAGEMENT
 lake model is incorporated  in estimates of standard
 errors in FORY from the regression equation.
 Cultural Sources of Phosphorus

 Cultural sources of phosphorus can now be assessed
 by evaluating  the difference  between background
 loading  and present day loading, L*, estimated from
 the P concentration in a lake. Equation 9 can be rewrit-
 ten as:
     L* - L*bg = ARR + AWW + AAG.
 (18)
 Evaluations of ARR, AWW, and AAG are described in
 the next section.
 Residential Area Runoff, ARR

 The increased  loading of P caused by single famliy
 residential  areas  for  lakes with only  forest and
 residential land use and no wastewater disposal in
 their basins, can be calculated by difference.
     ARR = L*  - L*
                   bg
 (16)
   The standard error of loading from residential area
runoff, SEARR calculated by this method is a function
of the standard errors in the present day loading, and
the background part of the present day loading:
     SEARR = (SEL.2 + SEL.bg2)i/2.
(17)
   Equation 16 is useful for only a few lakes in the
Puget Sound  region, because at most lakes there are
other potential P loadings (such as from septic tank
systems or agricultural land). For estimating ARR for
other lakes,  a  simple  empirical relationship  was
developed  based  on  computed yields of P from
sewered residential areas for four lakes in the region
that meet the conditions required for use of equation
16.
   Increased P yields caused by residential develop-
ment, ARRY, were calculated for each of the four lakes
by using equation 18:
    ARRY =
                ARR
              WSAr
(18)
where WSAres is  the area, in square  kilometers, of
residential land use. The mean increase in P yield from
developed areas over undeveloped area in the basins
of the four lakes was 7.0 (kg/km2)/yr with a standard
deviation of 2.9 (kg/km2)/yr. The  values  of ARR and
standard errors, SEARR_ for other lakes in the region
can be estimated from these data by using equations;
19 and 20.

    ARR = 7.0.WSAres»                       (191

    SEARR  = 2.9.WSAres«                      (201


Wears/lore Septic Tank Systems, hWW

Phosphorus loading to  a  lake from nearshore septic:
tank systems can  now be  calculated by difference for
lakes that are unaffected  by agricultural land use.
        The standard error of AWW, SEAWW, is calculated by
        equation 22, which sums the effects of uncertainties
        in total present day P loading and all other previously
        evaluated sources of loading.
      SEAWW = (SEL.2
                              SEL.bg2
SEARR2)l/2     (22)
   For estimating loading from septic tanks for lakes
 that have agricultural land use in their basins, an em-
 pirical  regional relationship between AWW and the
 numbers of nearshore dwellings around lakes was
 developed.  Values of AWW calculated from equation
 21  for  a  sample  of 24  lakes not influenced by
 agricultural land use were used to make this analysis.
   Analysis  of the available data for effects of both re-
 cent and past development revealed a correlation be-
 tween calculated septic system P loadings, AWW, and
 numbers of nearshore dwellings present in 1940 at the
 24 lakes. The fitted regression relationship, which ex-
 plains 36 percent of the sample variance, and stan-
 dard errors of  estimates are shown in Figure 4. For
 lakes with  no  1940 dwellings, the average value of
 AWW was about zero. The apparent effects of  other
 variables on loading  from  septic  tanks,  such  as
 geology and soil type, were inconsistent and could not
 be distinguished with the available data. The results
 suggest that the age of septic tank systems is in some
                               Explanation

                                AWW 0.68(1940
                                buildings) - 0.2

                                Standard error of
                                estimation

                                Lake in glacial outwash

                                Lake in glacial ti
    AWW = L* - L*bg - ARR«
(21)
    -40
               10      20      30      40
            NUMBER OF BUILDINGS IN 1940
Figure 4.—Relationship between estimated  phosphorus
loading from  near-shore septic systems and  numbers  of
near-shore buildings present in 1984.
                                                 36

-------
 way related to the amount of P they contribute to an
 adjacent lake. This issue is explored in more detail by
 Gilliom and Patmont (1982).
 Agricultural Land, A4G
 Increased P loading caused by agricultural land use,
 when present, can be calculated by difference, with
 other loadings estimated independently.
     AAG =  L*  - L*bg - APR - AWW

 The standard error of estimated loadings from  agri-
 cultural land can be calculated using  equation 23,
 which  combines the effects of uncertainties in all
 terms in the right hand side of equation  23.
     SEMG =  (SEL.2 + SEL.bg2 + SEARR2 +
              SEAWW2)1'2                         (24)
 Because of the extreme variability  in P  yields from
 agricultural land, depending on the type of agriculture,
 intensity, and other factors, no attempt was made to
 develop a  regional  empirical relationship  for esti-
 mating increased loadings from agricultural  areas.
 APPLICATION OF ASSESSMENT TOOLS

 Methods described in this report allow estimation for
 a lake of present day water quality, background P con-

            Watershed Land-Uses
         (shown on map areas around lakes)
         Forest and unproductive

         Agriculture

         Residential

         Lake surface

         Basin boundary
         Phosphorus Loading Sources
               By Proportions
          (shown on right-side bar graph)

         Forest and unproductive-land runoff
         combined with precipitation on lake surface

         Agricultural-land runoff

         Residential-land runoff

         Nearshore septic-tank systems

         Upstream lakes
                         WATER QUALITY ASSESSMENT METHODS

       centration  and water quality, amounts of nonback-
       ground P loading attributable to different land uses,
       and the sensitivity of the lake to future changes in P
       loading. Detailed consideration of uses and reliability
       of these estimates for lake management is provided in
       Gilliom (1982a), and an example of regional applica-
       tion of the procedure is described in Gilliom (1982b).
       An  overview of selected  aspects of method applica-
       tion follows, beginning with how to display the esti-
       mates for regional assessment.
         After considerable experimentation in cooperation
(23)    wjtn |oca|  ancj state agency representatives, the for-
       mat of the example  in Figure 5 was chosen to display
       results on  a geographic  basis.  For  each  lake
       evaluated, the drainage basin and  land use areas are
       delineated, and  a bar graph shows the estimated
       background  P level by  the  left  side  bar and  the
       measured  present day P level by  the right side bar.
       Above each bar is the standard error of that concen-
       tration estimate. The  right side bar is subdivided to
       show the estimated  proportions of above-background
       P loading attributable to  different  land uses. The bar
       graphs are shown with scales for both P concentra-
       tion and water quality groups from Table 1. Beneath
       the name of each lake is  a sensitivity rating.
         The reliability  of  loading estimates for land use
       sources is not shown on the graphs but can be readily
       computed. Table 2, for example, lists standard errors
       for land use related increases  in P loading for 14 lakes
       included in the regional  analysis by Gilliom (1982b).
       Each of the 14  lakes  had apparent present  day P

         Water-Quality Sensitivity  to Increased
                Phosphorus Loading

      Extreme —  10 kilograms increase in annual phosphorus
                loading rate results  in 20 micrograms per
                liter greater increase in concentration
         High —  10 kilograms increase in annual phosphorus
                loading rate results in a 10 to 20 microgram
                per liter increase in concentration
     Moderate —  10 kilogram increase in annual phosphorus
                loading rate results in a 3 to 10 microgram
                per liter increase in concentration
         Low —  10 kilogram increase in annual phosphorus
                loading rate results in less than  a 3 micro-
                gram per liter change in concentration
   Model-estimated background phosphorus levels


             Measured present-day phosphorus levels
                                                                        • Water-quality sensitivity
                                                                         to increased phosphorus
                                                                              loading
v Figure 5.—Example of eutrophication assessment.
                                                     37

-------
 LAKE AND RESERVOIR MANAGEMENT


                               Table 2.—Summary of phosphorus loading evaluation.

                              L*  = (PREL'A)  + (FORY«WSAbg) + ARR + AWW +  AAG

 1. estimate PREL for all lakes from published data:

     PR EL  = 20 (kg/km2)/yr

 2. compute FORY for any undeveloped lake:

             L* -  (PREL»A)
     FORY =
                 WSA
                     •bo
3. evaluates values of FORY calculated for 24 undeveloped lakes for correlation with regional variables so that it can
be estimated for developed lakes (fig. 3):

     FORY = 7.1 •  In RO + 16.6 (RO is annual runoff)

4. compute ARR for any lake with only undeveloped land and sewered residential areas in its basin:

     ARR = L*  - L*bg

     where L*bg =  (PREL»A) + (FORY»WSAbg)

5. average values of ARR calculated for 4 lakes for application to other lakes in the region that do not meet the criteria for
step 4:

     ARR = 7.0 • WSAres

     where WSAres is the area of basin land in residential land use

6. compute AWW for any lake with only undeveloped and and unsewered residential areas in its basin:

     AWW  =  L*  - L*bg - ARR

7. develop regional relationship from values of AWW calculated for 24 lakes for estimating AWW for lakes that do not meet
the criteria for step 6 (fig. 4):

     AWW  =  0.68»(number of nearshore dwellings in 1940)  - 0.20

8. estimate loading from agricultural land by difference:

     AAG = L*  - L*bg - ARR - AWW
    Table 3.—Percent standard errors in phosphorus loading estimates for 14 selected lakes evaluated by Gilliom (1982).
Lake Back-
ground
Anderson
Armstrong
Beaver
Big
Cassidy
Cranberry
(Skagit County)
Howard
Loma
Lone
McMurray
Pass
Shoecraft
Stevens
Weallup
60
30
30
20
30
100
40
30
70
30
50
40
20
30
Present
day1
50 (20)
40 (20)
40 (20)
40(20
40 (20)
40 (20)
40 (20)
70 (20)
50(20)
40 (20)
40 (20)
30 (20)
40 (20)
60 (20)
Cumulative
increase in
loading above
background
level
70 (30)
100 (60)
100 (70)
70 (50)
60 (40)
300 (200)
70 (50)
200 (60)
50 (30)
100 (60)
60 (30)
90 (70)
100 (60)
100 (60)
IIIWI V IMMM* UWUI W<9 Wl IW«UII 1^ || I\»I OC19C
Residen-
Upstream tial Septic
lakes runoff systems
_ _ _
— — 200 (200)
_ _ _
100 (60) - -
_ _ _
— — —
— — 70 (50)
— — 200 (60)
_ _ _
_ _ _
_ _ _
— — 90 (80)
— 40 (40) 200 (100)
100 (60) - -

Agriculture
70 (30)
200 (200)
100 (70)
80 (60)
60 (40)
300 (200)
—
—
60 (30)
100(70)
60 (30)
—
—
—
'Values in parentheses are precent standard errors that would result if the standard error in the mean lake-water concentration of phosphorus was ± 10 percent for
lakes, reduced from the present level of ±30 to ±50 percent



                                                       38

-------
loading that exceeded estimated background loading
by more than 50 percent. Errors for cultural sources
that accounted  for  less  than  25 percent  of the
cumulative  increase  in loading above background
levels are not shown. Using estimates of present day P
concentrations from  available data, error values for
present day loadings range from ± 40 to ± 50 percent.
The loadings calculated from more precise estimates
of mean lake concentration, hypothetically set to ± 10
percent, have a standard error of about ± 20 percent.
Standard errors of estimated cumulative loading in-
creases in  lakes in Table 3 are mainly in  the ±50 to
±100  percent range.  Improved concentration esti-
mates  would  reduce  these standard errors substan-
tially, as shown. In general,  loading increases from
small  P  sources  (about 25  percent of background
loading or  less), and  all loading increases for lakes
where  more than two major  nonpoint P sources are
present, are estimated  with standard errors of ±100
percent or greater.
  Two key applications of the information contained
in Figure 5, when repeated and mapped for all lakes of
interest in  a region, are (1) prioritizing management
and research efforts,  and (2) designing more detailed
studies. Priorities for more intensive lake-by-lake in-
vestigations can be assigned based on the severity of
water  quality degradation  in relation to background
levels,  as shown by the bar graphs,  and based on lake
sensitivity ratings which indicate whether or not a lake
is likely to  respond  dramatically to  a  small  or
moderate increase in loading. The design of studies of
lakes chosen for further investigation is enhanced by
both the land use/P loading assessment shown in the
bar graphs, which  indicate which  P sources require
the most careful  study, and the  sensitivity rating.
Assessments of highly sensitive lakes tend to depend
                 WATER QUALITY ASSESSMENT METHODS

greatly on an accurate water balance, and studies of
such  lakes  should  include  a  careful   hydrologic
analysis. These and other applications of the assess-
ment  results are enhanced by the mapped, graphical
display  of  data which  encourage  recognition  of
geographic  patterns  in land  uses and  lake  quality
within a region.

REFERENCES

Dillion, P.J., and F.H. Rigler.  1975. A simple method for pre-
  dicting the capacity of a lake for development based on
  lake trophic status. J. Fish. Res. Board Can. 32:1519-31.
Gilliom, R.J.  1981. Estimation of background loading and
  concentrations  of  phosphorus for lakes  in the  Puget
  Sound region, Wash. Water Resour. Res. 17(2):410-20.
	. 1982a. Estimation  of nonpoint  sources of phos-
  phorus to  lakes in the  Puget Sound region, Wash. U.S.
  Geol. Surv. Open-file rep. 82-161.
	1982b. Lake-water quality and  land-use relation-
  ship? for selected lakes in  the Port Townsend quadrangle,
  Puget Sound region, Wash. U.S. Geol. Surv. Open-file rep.
  82-684.
Gilliom, R.J., and  C.R. Patmont. 1982.  Lake  phosphorus
  loading  from septic systems  by seasonally perched
  ground water, Puget Sound region, Wash. U.S. Geol. Surv.
  Open-file rep. 82-907.
Gilliom, R.J., and G.C.  Bortleson. 1983.  Relationships  be-
  tween water  quality and phosphorus concentrations for
  lakes of  the Puget Sound region, Wash. U.S. Geol. Surv.
  Open-file Rep. 83-255.
Larsen, D.P.,  and  H.T.  Mercier. 1976. Phosphorus retention
  capacity of lakes: J. Fish.  Res.  Board Can. 33:1742-50.
Reckhow, K.H. 1979. Quantitative Techniques for the Assess-
  ment of Lake Quality. EPA-440/5-79-015.  U.S. Environ. Prot.
  Agency.
                                                   39

-------
 SURFACE RUNOFF WATER QUALITY FROM DEVELOPED AREAS
 SURROUNDING A  RECREATIONAL LAKE
 JAY A. BLOOMFIELD
 JAMES W.  SUTHERLAND
 JAMES SWART
 Bureau of Water Research
 New York State Department of Environmental  Conservation
 Albany, New York


 CLIFFORD  SIEGFRIED
 State Museum
 New York State Education Department
 Albany, New York


            ABSTRACT

            During 1980, as part of its Nationwide Urban Runoff Program (NURP), the U.S. Environmental
            Protection Agency entered into a cooperative agreement (P002229-01-1) with the New York State
            Department of Environmental Conservation to study urban runoff at Lake George, N.Y., located
            in the southeastern Adirondack Mountains. The purpose of the study was to determine the ef-
            fect of runoff from a developed watershed on the water quality of the lake and its tributaries.
            More than 40 storm events were sampled during a 2-year period at six tributary sampling sta-
            tions to assess the loading of plant nuttients and other contaminants from developed and
            underdeveloped areas to the open waters of the lake. Additionally, the nearshore and open
            waters of Lake George were sampled during storm and nonstorm periods, to assess the impact
            of stormwater runoff on the trophic conditions of the lake. Runoff from developed areas ac-
            counts for 13.6 percent of the annual phosphorus loading to Lake George, which is 15.1 percent
            of the load to the South  Lake and 6.5 percent of the load to the North Lake. In  addition,
            developed areas contribute 28.9 percent of the annual phosphorus load to the study area at the
            extreme south end of the lake.
INTRODUCTION

Lake  George and the surrounding watershed have
become a major tourist and recreation area in New
York State during the past decade, with resulting in-
creases in the permanent and seasonal population of
communities situated along the lake, land use rezon-
ing toward the  tourist-commercial  and  residential
categories, and development throughout the water-
shed, especially along the south portion of the lake.
The economy of communities in the watershed, being
almost totally tourist and recreation related, depends
upon a high level of water quality in the lake.
  Widespread public  concern  for water quality has
been partially responsible for a large number of Mm-
nological investigations on Lake George during the
past 15 years. Distinct differences in water quality in-
dicators have been reported, with the south, more-
developed, portion of the Lake exhibiting lower trans-
parencies (Ferris and Clesceri, 1977; Wood and Fuh.-j,
1979; Wood, 1982; Pope, 1981,  1982; Siegfried, 1982;
Siegfried et al. 1983), lower hypolimnetic dissolved ox-
ygen concentrations (Wood and Fuhs, 1979; Siegfried,
1982; Siegfried et al. 1983), higher phosphorus (Aulen-
bach and  Clesceri,  1971; Siegfried et al. 1983)  and
chlorophyll a concentrations (Wood and Fuhs, 1979;
Wood, 1982; Siegfried  et al. 1983), and a trend toward
seasonal blooms of blue-green algae (Monheimer and
Baker,  1982; Siegfried,  1982; Siegfried et al. 1983).
These differences in water quality indicators are
associated with, or could result from, higher levels of
cultural activity  (i.e.,  increased sources  of  phos-
phorus) in the southern portion of the watershed, and
continued development will tend to accentuate these
differences (Dillon, 1983; Shapiro, 1983).
  Several  investigators  have  constructed  nutrient
budgets for Lake George based on relatively little or
nonexistent data (Aulenbach,  1979; Aulenbach and
Clesceri, 1971,1972,1973,1977; Aulenbach et al. 1979;
Gibble, 1974; Hetling, 1974; Wood and Fuhs, 1979). Al-
though the estimates vary, all of the budgets indicate
that  atmospheric deposition and surface runoff are
the major sources of nitrogen and phosphorus input to
the Lake. In his evaluation of these nutrient budgets,
Dillon (1983) estimates that, on an  annual basis, the
mean contribution of total phosphorus  in runoff from
developed areas  is approxmately 20 percent of the
total loading to the south portion of the lake. Unless
certain controls are implemented, phosphorus loading
will increase as development continues in this portion
of the watershed. It would appear that any water quali-
ty management program for Lake George should ad-
dress the issue  of  runoff control  from developed
areas.
  During June of 1980, a study was initiated with the
following  objectives:

  1.  To identify and quantify (in terms of concentra-
tion and load) the major  runoff contaminants trans-
ported to  Lake George by streams and  storm sewers
                                               40

-------
                                                                     WATER QUALITY ASSESSMENT METHODS
located in the developed, south portion of the water-
shed, and
  2. To determine the water quality response in south
Lake George to the total loadings of contaminants dis-
charge from urbanized areas under present levels of
development.
  The data presented here are the results of the water-
shed sampling program. The results of the lake sampl-
ing program are presented in Siegfried (1982), Wood
(1982) and Siegfried et al. (1983).
STUDY AREA DESCRIPTION

Lake  George is located  in the eastern Adirondack
Mountain Region of New York State near the Vermont
border  and within the  Lake Champlain Drainage
Basin. The  lake is 51 km  long, averages 2.3 km wide,
and is aligned  in a nearly north-south direction. At
mean lake level (97 m AMSL) the surface  encom-
passes  114 km2. The lake consists  of two  distinct
basins  of  nearly equal surface area (57 km2) and
volume  (1.05 kmS) which are referred to as North and
South Lake George. A notable difference between the
basins is the watershed area:  179 km2 for North Lake
George  and 313 km2 for South Lake George. The lake
flows from  south to north and the water retention time
has been calculated to  be eight years (Ferris and
Clesceri, 1977).
  The most recent land use data for the Lake George
watershed  was presented by  Hetling (1974) and was
based upon 1968  aerial  photography. According to
this report,  97 percent of the total watershed  was un-
developed and about 75 percent of the developed area
was concentrated along the shoreline of South Lake
George. Most of the development  is along the lake
shoreline because  of the steep topography of the
watershed.
  The study area was located at the south end of Lake
George  and included stream, storm sewer, and direct
drainage in the portion of the watershed south of Tea
Island (Fig. 1). The total  area of this  section of the
Lake George watershed is 59.95 km2 and the land use
characteristics include  urban, agriculture, forest, and
water—with the forest (87 percent) and urban (12 per-
cent)  categories constituting  the major  land  usage.
Direct drainage along the shoreline has a relatively
small surface area (0.57 km2)  when compared to the
other  drainages, but has the highest proportion of
developed area with approximately 97 percent urban
land use and 36 percent impervious area.
  Three tributaries and three storm sewer catchments
were  monitored during the study. Figure 1 identifies
the drainages and Table 1 contains a summary of mor-
phometric,  land use, and population  data for each
drainage. A more detailed discussion of the drainage
characteristics is presented in Sutherland et al. (1983).
METHODOLOGY

Each study drainage had a primary station where sur-
face runoff was  monitored for flow and sampled for
physical and chemical parameters. The primary sta-
tion for most drainages was located along the stream
or storm sewer conveyance near its outflow to Lake
George (coded with four-digit numbers on Fig. 1). Pro-
spect Mountain  Brook is a section of the Sheriff's
Dock storm sewer that drains a forested  region and
was included as a control in  the  study. The station
was  located west of the developed  area of  Lake
George Village. Atmospheric deposition samples were
collected at two sites shown on Figure 1.
  Equipment used at the primary stations to monitor
flow and collect samples is summarized in Sutherland
et al.  (1983).  Continuous discharge  records were
developed at each site with temporal resolution rang-
ing from 5 to 60 minutes depending on the rapidity of
catchment response to precipitation.
  Runoff events of more than 1.0 cm of precipitation
or equivalent estimated snowmelt were  sampled for
chemical  quality. Discrete samples  generally were
selected for analysis. In some cases, discrete water
samples were composited on a discharge-weighted
basis, with each  subsample volume proportional to
the amount of discharge (in  m.3) represented by that
discrete sample. Compositing generally was limited to
the three tributary stations during periods of the year
without ice effects. Samples also were composited at
Cedar Lane Storm Sewer (3702), where the discharge
was controlled by a Palmer-Bowlus flume.
  Meteorological  data collected included precipita-
tion, air temperature, and snowpack. These results are
summarized in Sutherland et al. (1983).
  Samples were  analyzed for pH, specific conduc-
tance, total alkalinity, plant nutrients, major  ions, and
lead.  Selected samples  were analyzed  for fecal
bacteria, trace organics, and trace metals (Sutherland
et al. 1983). A detailed discussion  of the  chemical
analyses and  analytical  techniques is given  in the
Quality  Assurance/Quality Control Plan  for  this pro-
ject (N.Y. State Dept. Environ. Conserv., 1981).
                                         Tea Island
                                            east
                                       East Brook
                            4I  English Brook
                            40 Marine Village
                            39 Sheriff's Dock
                            38 West Brook
                            37 Cedar Lane
                            I  I Monitored drainage
                            II Unmonitored drainage
                            I  I Direct drainage
                            A  Primary sampling station
                            £>  Atmospheric deposition
                                station
Figure 1 .—Drainages and sampling stations at the south end
of Lake George.
20 km
                                                 41

-------
LAKE AND RESERVOIR MANAGEMENT

RESULTS AND DISCUSSION

Hydrology

Stream discharge was measured continuously at all
sites, with 5-minute resolution at Stations 3901  and
4001; 15-minute resolution at Stations 3702, 3801, aid
4101; and 60-minute resolution at Station 3950. Table 2
summarizes the total runoff (mm) per sampled period
by catchment. In general, the runoff rates are greatest
during spring snowmelt conditions. Approximately,
two thirds to three quarters  of the annual runoff oc-
curs between February and June (Sutherland  et al.
1983).
  Table 2 also presents direct runoff coefficients lor
each sampled period. These coefficients are the ratio
of direct runoff to water input for the period. Water n-
put is defined as rainfall plus snowmelt. Again, the
spring periods represent the time  when the largest
fraction of available  water  becomes  direct runoff,
usually from 20 to 30 percent. During certain periods,
these coefficients and baseflow runoff rates were esti-
mated because of  incomplete continuous discharge
records. These estimates were based on values  at s.d-
jacent similar sites, or similar periods at the  sane
site, with continuous  records. In turn, the estimated
direct runoff coefficients were used to fill in values for
direct discharge during certain periods. This estima-
tion procedure allows development of discharge and
loading values at all sites for the entire study period.
Thus, the loadings and discharge summaries for the
Marine Village storm  sewer (4001) should be viewed
more critically,  because it  lacks  a complete con-
tinuous discharge record during the early parts  of the
study.
  Since runoff per unit area varied among the  sites,
primarily because of the large variation in drainage
area and to  a lesser extent the permeable nature of
the soils, estimates were made of the amount of  water
leaving the smaller drainages via subsurface seepage.
It was assumed that groundwater losses to the la4001 >3901
                            >3801 >4101 >3950. Prospect Mountain Brook(3950)
                            can be considered as a natural tributary.  English
                            Brook (4101) exhibits  slightly higher levels of  total
                            phosphorus, chloride, total suspended sediment,  lead,
                            and nitrate nitrogen than 3950,  but, in  terms  of this
                            data  set,  has  been affected only slightly by  land
                            development. The remaining four drainages  show a
                   Table 1.—Characteristics of drainage areas at the south end of Lake George.
         Drainages
   Physical
      Channel
Area Length Slope
(km2)(km)    (%)
               Land Use

Urban  Agric.  Forest  Water  Impervious
Population Density
   (persons/ha)
Permanent  Total3
MONITORED
Cedar Lane Culvert
West Brook
'Sheriff's Dock Culvert
Prospect Mountain Brook
Marine Village Culvert
English Brook
OTHER
Direct
2East Brook
Tea Island East
Tea Island West
Drainage Total (weighted)


0.31 0.5
21.60 8.0
2.24 1.6
0.99 0.9
0.66 0.6
21.2411.2

0.57 —
9.08 1.4
2.58 —
1.98 —

59.95km2

1.6
1.2
9.1
18.5
3.8
2.9

—
1.5
—
—



42.02
7.47
27.90
4.89
76.87
4.77

96.55
21.97
18.15
12.96
11.73%
7.03km2

—
0.05
—
—
—
0.45

—
0.46
—
6.94
0.48%
0.29km2

57.98
91.69
72.10
95.11
23.13
94.46

3.44
77.00
81.85
80.10
87.31 %
52.32km2

—
0.79
—
—
—
0.32

—
0.57
—
—
0.48%
0.29km!

3.60
1.57
9.02
3.64
17.97
1.90

35.76
2.72
2.32
3.62
2.74%
1.64km2

2.0
0.1
3.1
0.0
5.1
0.2

0.4
0.2
0.8
2.5
0.5


13.2
0.8
9.2
0.0
16.1
0.7

25.8
1.9
4.0
6.0
2.1

'includes Prospect Mountain Brook
includes Cedar Lane Culvert
'total includes permanent plus average seasonal population
                                                 42

-------
 gradual  progression from natural to impacted water
 quality,  with the storm  sewers at Cedar  Lane and
 Marine Village exhibiting EMC's for each constituent
 typical of many urban areas.
   Figure 2 presents histograms of EMC's for total
 phosphorus  for each  primary sampling  site.  The
 gradual  progression from natural  to developed con-
 ditions is quite noticeable. The same pattern occurs
 for  chloride,  total  suspended sediment,  and lead
 (Sutherland et al. 1983). Again, Prospect Mountain and
 English  Brook  represent  the most natural conditions,
 while  West Brook  and the  storm sewers  seem the
 most impacted by land development.

 Total Phosphorus Loading Calculations

 One of the main objectives  of this study was to im-
 prove  estimates of the  phosphorus contribution to
 Lake George from the surrounding watershed. To this
 end, a detailed analysis of the phosphorus data was
 conducted.
   Approximately 30 runoff events  were sampled for
 from each of the six drainages between Oct. 1, 1980
 and Sept. 30,  1982.  Approximately 20  chemistry
 samples also were  collected at each site during non-
 event  periods.  For  total phosphorus, an  arithmetic
 average  of these values was used for each site to
 calculate the baseflow phosphorus component. There
 was no  evidence of seasonal or discharge relation-
 ships  with nonevent  phosphorus concentrations at
 any site. The baseflow total phosphorus concentra-
 tions (in ^g/l) were as follows: 22 at 3702, 8 at 3801,10
 at 3901, 3 at 3950, 14 at 4001, and 4 at 4101.
   The  duration of  the  study was  divided  into six
 periods,  commencing the  first  days of  October,
 February, and July in each of the 2 water-years. These
 periods were  chosen to represent fall and winter con-
 ditions, spring snowmelt, and hot weather conditions,
 respectively.  During  each  period,  at  each  site,
 hydrograph separation  (Chow, 1964) was used  to
 separate direct  runoff from  baseflow for  both indi-
vidual sampled events and the total period. The total
 phosphorus load (in grams)  for the sampled events
then was calculated by summing the individual loads
for each  event during a  period. Then the  baseflow
phosphorus load was calculated for  both sampled
events  and the total period (Sutherland et al. 1983).
  When the direct runoff phosphorus loads are com-
bined with  the baseflow phosphorus loads, the total
                  WATER QUALITY ASSESSMENT METHODS

 phosphorus loads (in kg) can be calculated for each
 site. These results are shown in Table 3. When the
 period loadings are recalculated as percentages of an-
 nual load,  the importance of the spring snowmelt  is
 quite evident, with this period accounting for 46.04
 percent to 81.15 percent of the annual phosphorus
 load at the six sites. When all guaged drainages are
 combined,  76.73 percent of the annual phosphorus
 load occurs during the spring.
   When the direct runoff and total runoff  loads are
 standardized by drainage area, the areal loads (g/ha/-
 day) in the direct runoff and total runoff (Table 3) can
 be determined,  as can the discharge-weighted con-
 centrations in the direct runoff (Table 3). With the ex-
 ception of  West Brook (3801), all drainages exhibit a
 dilutional relationship between phosphorus and direct
 runoff rates.  This indicates a finite source of phos-
 phorus in each drainage which, during periods of high
 runoff such as the spring snowmelt,  can  be  diluted
 (Fig. 3).
      I0  50 100 500      10 50  IOO 500

                   Total Phosphorus (pg/t)
                                     10  50 IOO 500
Figure 2.—Histograms of Event Mean Concentrations for
total phosphorus at  primary sampling stations during the
study, (n = number of events sampled).
Table 2.—Total runoff1 (in mm) and direct runoff coefficients2 (unitless) at primary sampling stations during sampled periods.
           Total runoff also is present for water years 1980-81 and 1981-82, and for the study period (in m3).
Stations
Period
10/80 - 1/81

2/81 - 6/81

7/81 - 9/81

10/81 - 1/82

2/82 - 6/82

7/82 - 9/82

WY 80-81
WY81-82
Total Study (m3)
3702
'(8.1 18E)
2(0.007E)
114.300
0.119
21.620
0.034
34.686
0.037
152.700
0.101
(9.01 6E)
(0.01 OE)
1 144.038
1 196.402
1 105,536
3801
82.533
0.022
247.500
0.191
79.672
0.044
153.750
0.098
445.650
0.293
82.892
0.015
409.705
682.292
23,587,135
3910
11.685
0.037
94.500
0.089
21.344
0.062
70.848
0.159
260.100
0.232
(2.852E)
0.005E
127.529
33.800
1,033,376
3950
(20.664E)
(0.060E)
(172.800E)
(0.200E)
51.980
0.142
161.745
0.247
335.400
0.358
4.140
0.006
245.444
501.285
739,262
4001
(114.390E)
(0.100E)
(264.750E)
(0.0250E)
(62.284E)
(0.090E)
(191.757E)
(0.200E)
426.900
0.341
47.104
0.029
(441.424E)
665.761
730,742
4101
89.667
0.080
283.800
0.307
42.872
0.072
214.881
0.175
468.900
0.400
28.244
0.019
416.339
712.025
23,966,451
E = estimated
                                                  43

-------
LAKE AND RESERVOIR MANAGEMENT

Annual Phosphorus Budget, South End of
Lake George

The portion of Lake George south of Tea Island (:>ee
Fig. 1) was chosen for development of an annual phos-
phorus budget because of the reduced water quality in
this region of the Lake (Siegfried, 1982; Siegfried et al.
1983). Information on areal phosphorus loadings from
guaged drainages was compared to the level of devel-
opment in each drainage (see Table 1) and this rola-
    500
    IOO
 I
 Q_
            02
                 04    06    08    10

                   Direct Runoff  (mm/doy)
Figure 3.—Relationship between total phosphorus concen-
tration and direct runoff rates at the primary sampling sta-
tions during the study.
tionship was used to develop phosphorus loads from
unguaged areas.
  Areal  phosphorus  loads  were  standardized  in
drainages 3702, 3901 and 3950 for groundwater seep-
age, using the assumption that these areas have aver-
age  daily runoff equal to the average of drainages
3801, 4001 and  4101  (1.520  mm/day), although the
channel runoff of  the former drainages is  less. The
phosphorus concentrations in the groundwater flow
was assumed to be equal to the baseflow concentra-
tions. The results are presented in Table 4, and it can
be seen that  there is direct relationship between cor-
rected areal loading and percent development (Fig. 4).
The regression line that describes this relationship is:

     LAp = -0.149 + 0.241 Ln (%D) (ft = 0.952, n = 6)

where LAP = corrected total phosphorus areal loading
      rate (g P/ha/day)
      %D = percent  catchment developed (unitless)

When this relationship is used to estimate  the phos-
phorus loadings from unguaged drainages there is an
annual loading to the south  end of Lake George of
908.0 kg P/yr. This corresponds to a watershed areal
loading  rate  of 0.415  g  P/ha/day for the study area.
Figure 5 summarizes the total for the study area.
  Table 5  compares  the  present loading  of phos-
phorus with various development scenarios, from un-
developed conditons (LAP  = 0.100 g P/ha/day) to the
watershed 100 percent developed. The small amount
of atmospheric loading to the  lake surface,  37.7 kg
P/yr, also is  considered and lake areal phosphorus
loading rates are presented. If the south embayment
of the lake is assumed to behave according  to Vollen-
wieder and  Dillon (1974),  the various areal  loading
rates can  be  plotted as to trophic status (Fig. 6). The
letters correspond to  each scenario in Table 5. The
present loading (scenario d) is in the transition region,
while all scenarios with > 25 percent development are
 Table 3.—Total phosphorus load' (kg), areal loading of total phosphorus2 (g/ha/day) in total runoff and discharge-weighted
 average total phosphorus concentrations3, (ug/l) in direct runoff at primary sampling stations during the sampled periods.
 Total phosphorus load also is given for Water Years 1980-81 and 1981-82, and areal loading of total phosphorus in total runoff
                                      is presented for the total study.
Period
10/80 - 1/81
2/81 - 6/81
7/81 - 9/81
10/81 - 1/82
2/82 - 6/82
7/82 - 9/82
WY 80-81
WY 81-82
Total Study
3702
'(0.332E)
2(0.087E)
3(475E)
5.144
1.106
372
1.663
0.573
438
1.489
0.391
345
3.271
0.704
163
(0.294E)
(0.103 E)
(475E)
'7.109
'5.054
20.537
3801
17.714
0.067
34
158.727
0.490
91
30.304
0.152
60
65.769
0.248
59
399.780
1.234
109
15.973
0.080
34
206.745
481.522
0.436
3901
6.291
0.228
272
42.564
1.267
163
8.948
0.434
192
7.953
0.289
62
22.642
0.674
74
(0.90 2 E)
(0.044E)
(380E)
57.803
31.497
0.546
Stations
3950
(0.532E)
(0.044 E)
(31 E)
(1.99E)
(0.134 E)
(25E)
1.094
0.120
23
1.994
0.164
21
2.839
0.191
13
0.044
0.005
35
(3.61 6E)
4.877
0.118
4001
(5.651 E)
(0.696E)
(260E)
(10.892E)
(1.1 OOE)
(165E)
(4.865 E)
(0.801 E)
(230E)
(8.625E)
(1.062E)
(165E)
10.875
1.098
75
1.481
0.244
343
(21.408E)
20.981
0.880
4101
17.742
0.068
25
114.689
0.360
45
31.342
0.160
58
61.818
0.237
38
207.425
0.651
43
(4.721 E)
(0.024E)
(40E)
163.773
273.964
0.282
390-3950
5.759
0.375
40.574
2.164
7.854
0.683
5.959
0.388
19.803
1.056
0.858
0.075
54.187
26.620
0.886
 E = estimated
                                                  44

-------
                                                                        WATER QUALITY ASSESSMENT METHODS
 in the eutrophic range. Undeveloped scenarios (a and
 b) are in the oligotrophic range.

 Phosphorus-Trophic State Relationships

 Study Area

 The study area and the south end of Lake George have
 the characteristics shown  in  Table  6. In a  fashion
 similar to Wood and Fuhs (1979), the relationship be-
 tween  spring total  phosphorus  and   phosphorus
 loading rates  (Vollenweider, 1976):
where: TPSPR = spring total phosphorus (mg/m3)
       Ip     = areal phosphorus loading (mg/m2-yr)
       qs     = areal water loading (m/yr)
       z      = lake mean depth (m)
and between summer chlorophyll a and spring total
phosphorus (Dillon and Rigler, 1974):
Figure  4.—Relationship  between  total phosphorus areal
loading corrected for groundwater seepage  and  percent
development  in the study  drainages (see text  for
explanation).
Figure  5.—Annual total  phosphorus  loads  (kg/yr) from
drainages at the south end of Lake George.
 Table 4.—Areal loads of total phosphorus (g/ha/day) at the primary sampling stations corrected for groundwater flow (see
                                           text for explanation).
runoff
(mm/day)
Station
3702
3801
3901
(3901-3950)
3950
4001
4101
Direct Flow
0.161
0.369
0.321
0.136
0.554
0.533
0.570
Total Flow
0.466
1.496
0.632
0.323
1.023
1.517
1.546
Baseflow
0.305
1.127
0.311
0.187
0.469
0.984
0.976
Groundwater
Flow*
1.054*
0.000
0.888*
1.197*
0.497*
0.000
0.000
phosphorus areal load
(g/ha/day)
Direct Flow
0.471
0.345
0.515
0.841
0.103
0.763
0.243
Total Flow
0.537
0.436
0.546
0.886
0.118
0.880
0.282
Corrected
Total
0.765
0.436
0.635
1.176
0.134
0.880
0.243
 Note"—estimated annual total runoff (average of 3801, 4001, 4101), Qt = 1 520 mm/day.

               Table 5.—Projected annual phosphorus loadings under various development scenarios.
Scenario
a) Undeveloped conditions
b) Entire watershed like 3950
c) Entire watershed like 4101
d) Present conditions
e) 25% developed
f) 50% developed
g) 75% developed
h)95% developed
i) 100% developed
Watershed Areal
Loading Rate
(g P/ha/day)
0.100
0.143
0.243
0.415
0.626'
0.7921
0.8901
0.9471
0.9591
Notes, 'from equation in text.
'watershed area = 59.95 km2, Lake area = 2 62 km*
'includes atmospheric deposition on Lake (100% wetfall + 25%
Annual Phosphorus
Load from Watershed2
(kg P/yr)
218.8
312.9
531.7
908.0
1,369.8
1,733.0
1,947.5
2,072.2
2,098.5
dryfall, 144mg P/m'-yr)
45
Total Phosphorus
Load3
(kg P/yr)
256.5
350.6
569.4
945.7
1,407.5
1,770.7
1,985.2
2,109.9
2,136.2

Lake Areal
Loading Rate2
g P/mz-yr
0.098
0.134
0.217
0.361
0.537
0.676
0.758
0.805
0.815


-------
LAKE AND RESERVOIR MANAGEMENT
 Log10(CHLA)sum = 1.449 Log10(TP)spR - 1-136
 where (CHLA)sum  = average  summer chlorophyll a
                    (mg/m3)
 were used to assess the trophic status of the study
 area of the lake. The use of the two equations yields
 values for (TP)SPR =  14.4 mg/m3 and (CHLA)sum =  3.5
 mg/m3. These values are slightly above those actually
 measured by Siegfried  et al. (1983) at Station 1 in "he
 study area during 1981 (Table 6).  This result makes
 sense, since much of the phosphorus load in the study
 area is particulate phosphorus,  and the loss  of "he
 material by sedimentation in the littoral zone may lead
 to the violation of Vollenweider's assumption of all in-
 flow reaching the pelagic zone.
   By using Siegfried's  measured summer chlorophyll
 a value and the Dillon and Rigler (1974) equation, one
 obtains a value of (TP)SPR = 11.8 mg/m3, indicating a
 slight  overestimation  of the  effective phosphorus
 loading (22.0 percent too high).  This  overestimat on
 could be caused by several  factors:
   1. Atmospheric phosphorus  loadings are  over-
 estimated;
   2. The south embayment cannot be considered to
 behave as a lake;
   3. Watershed phosphorus loads do  not completely
 reach  the pelagic  zone  due to nearshore  sedi-
 mentation. (See for example, Sutherland et al. (1981),
 concerning the  buildup of deltaic material off  of
 English Brook; or
   4. General uncertainties  occur in various measure-
 ments and calculations.
   Despite  this result, the  loading estimates corras-
 pond with observed phosphorus  and chlorophyll  a
 levels  quite well.  In general,  the  work of  Siegfried
(1982) and  Siegfried  et al. (1983) shows north-south
 gradients in phosphorus, chlorophyll a and  Secchi
                           disk depth  corresponding  to  land development pat-
                           terns and watershed area-lake volume ratios.

                           Lake George

                           Table 7 is an annual phosphorus budget for the North
                           and South Basins of Lake George. Estimates of mor-
                           phometry, phosphorus transfer, and outflow are iden-
                           tical to those of Wood and Fuhs (1979). The Wood and
                           Funs (1979) estimate for the Bolton Landing sewage
                           treatment  plant  also  was  used even though  this
                           number (570 kg  P/yr) is probably several times too
                           high. Only additional field studies on this facility will
                           allow refinement  of this estimate.
                                                                 ^ Dangerous


                                                                 , Permissible
                                                             "Ohgotrophic"
                          Figure 6.—Areal loading  rates of total phosphorus under
                          various development scenarios (a - i; see Table 5) versus
                          mean depth-lake hydraulic retention time relationship (Z/TH;
                          see Table 6) to give lake trophic status.
                      Table 6.—Limnological characteristics, south embayment, Lake George.
 Symbol
Meaning
Source
Value
Units
Al
Aws
VI
r
IQ
qs
z
TH
LP
IP
(TP)spr
(CHLA)sum
lake surface area
watershed surface area
lake volume
watershed runoff rate
annual runoff
lake inflow rate
lake mean depth
lake hydraulic retention time
annual phosphorus loading
(watershed + atmospheric)
areal phosphorus loading
spring total phosphorus
summer chlorophyll a
Hutchinson et
this study
Hutchinson et
this study
this study
this study
this study
this study
this study
this study
Siegfried et al.
Siegfried et al.
al. (1981)

al. (1981)







(1983)
(1983)
2.62
59.95
3.11 x107
0.555
3.34x10?
12.75
11.87
0.93
908.1
361
9.4
2.6
km2
km2
m3
m/yr
m3
m/yr
m
yrs
Kg/yr
mg/m2-yr
mg/m3
mg/m3
                           Table 7.—Lake George annual phosphorus budget (kg P/yr).
Source
Atmospheric deposition
Watershed, developed
Watershed, undeveloped
Bolton Landing sewage treatment plant
Transfer from South Lake
Total sources
Outflow
Phosphorus retention
South Lake
829
525
1,557
570
—
3,481
1,300
2,181
North Lake
812
210
902
—
1,300
3,224
1,660
1,564
Total
1,641
735
2,459
570
—
5,405
1,660
3,754
% Total
30.4
13.6
45.5
10.5
—
100.0
30.7
69.3

-------
                                                                          WATER QUALITY ASSESSMENT METHODS
  Atmospheric deposition and watershed  contribu-
tions are estimated from  the  present study. An  at-
mospheric loading of 14.4  mg P/m2-yr was applied to
the lake surface area estimates by basin presented in
Wood and Fuhs (1979). The annual loads arrived at are
very close to the Wood and Fuhs (1979) estimates, but
about one half of the estimates of Dillon (1983).
  Loading of phosphorus from developed areas was
calculated  by applying  the  100  percent developed
loading  rate  (0.959  g   P/ha/day) to  estimates  of
developed land in the North Basin (6 km2)  and the
South Basin (15 km2). These estimates of developed
area loading are the lowest reported for Lake George
to date, and the only values  based on data collected
using  event-oriented sampling at  the lake. These
estimates are  about one half those projected  by
previous  investigators,  including  Dillon  (1983). This
discrepancy probably is due to previous investigators
considering that the developed area in Lake George is
equivalent to typical urban and suburban land in North
America. The area around Lake George is a seasonal
recreational community, with intense  use for several
months and light  use during  most of the year.
  The loadings from undeveloped areas were calcu-
lated using the areal loading measured  at  Prospect
Mountain Brook (0.143 g P/ha/day) and applying this
value to the estimates of undeveloped  land in the
North (178.2 km2) and South Basins (298.2 km2). These
results are significantly higher than all  previous  in-
vestigators except Dillon (1983), probably because the
present study used event-oriented, instead of fixed in-
terval, sampling of the tributaries. The estimates  for
undeveloped runoff are  virtually identical to those of
Dillon (1983).
  In summary, runoff from developed areas accounts
for only 13.6 percent of the annual phosphorus loading
to Lake George, which is 15.1  percent of the load to
the South Lake and 6.5 percent of the load to the North
Lake. In contrast,  developed areas contribute 28.9 per-
cent of the annual phosphorus load to the study area
at the south end of the lake.
REFERENCES

Aulenbach, D.B. 1979. Nutrient budgets and the effects of
  development on trophic conditions in lakes. Rep. no. 79-2.
  Fresh Water Inst. Rensselaer Polytech. Inst. Troy, N.Y.
Aulenbach, D.B., and N.L Clesceri. 1971. Results of lead time
  studies of baseline chemical nutrients in Lake George and
  nitrogen and phosphorus cycles in the Lake George eco-
  system. Eastern Deciduous Forest Biome, IBP. EDFB- IBP
  Memo rep. no. 71-121. Oak Ridge, Tenn.
	1972. Sources  and sinks of nitrogen and phos-
  phorus: Water quality management of Lake George. Rep.
  no. 72-35. Fresh Water Inst.  Rensselaer Polytech. Inst.
  Troy, N.Y.
	1973. Sources of nitrogen and phosphorus in the
  Lake George drainage basin: a double lake. Rep. no. 73-1.
  Fresh Water Inst. Rensselaer Polytech. Inst. Troy, N.Y.
	1977.  Means  for protecting the drinking water
  quality of Lake  George, N.Y. Rep. no.  77-1. Fresh Water
  Inst. Rensselaer Polytech. Inst. Troy, N.Y.
Aulenbach, D.B., N.L Clesceri, and J.R. Mitchell. 1979. The
  impact of sewers on the nutrient budget of Lake George,
  N.Y. Rep. no. 79-8. Fresh Water Inst. Troy, N.Y.
Chow, V.T. 1964. Handbook of Applied Hydrology. McGraw-
  Hill Book Co., New York.
Dillon, P.J. 1983. Nutrient budgets for Lake George, N.Y. In
  C.D. Collins, ed. The Lake George Ecosystem 3. (In press.)
Dillon, P.J., and F.H. Rigler. 1974. The phosphorus-chloro-
  phyll relationship in lakes. Limnol. Oceanogr. 19:767-73.
Ferris, J.J., and N.L Clesceri. 1977.  A  description of the
  trophic status and nutrient loading for Lake George, N.Y.
  Pages 135-181 in North American Project—a Study of U.S.
  Water Bodies. EPA-600/3-77-086. U.S. Environ. Prot. Agen-
  cy, Corvallis, Ore.
Gibble, E.B.  1974. Phosphorus  and  nitrogen loading and
  nutrient budget on Lake George,  N.Y. Masters thesis,
  Rensselaer Polytech. Inst., N.Y.
Hetling,  LJ.  1974.  Observations on the rate of phosphorus
  input into  Lake George and its relationship to the lake's
  trophic state. Tech. Rep.  no. 36. N.Y. State Dep. Environ.
  Conserv., Albany.
Hutchinson, D.R., et al. 1981. The sedimentary framework of
  the southern basin  of  Lake George,  N.Y. Quat. Res.
  15:44-61.
Monheimer, R.H., and M. Baker. 1982. Phytoplankton com-
  munity changes  in Lake George (N.Y.), 1975-79.  Pages
  41-47 in M. Schadler, ed.,  The Lake George Ecosystem 2.
New York State Department of Environmental Conservation.
  1981. Quality assurance project  plan for the Lake George
  Urban Runoff Project.
Pope, D.H. 1981. Data from Lake George monitoring program
  for the  year April  1980-April 1981.  Rep. Lake George
  Assoc.

	. 1982. Report on second year of the Lake George
  monitoring program, April 1981-November 1981. Rep. to
  Lake George Assoc.
Shapiro, J. 1983. An analysis of Lake George, N.Y. In C.D.
  Collins, ed. The Lake George Ecosystem 3. (In press.)
Siegfried,  C.A. 1982. Water quality  and phytoplankton of
  Lake George, N.Y.: Urban storm runoff and water quality
  gradients. Tech.  Pap. no. 66. Bur.  Wat. Res., N.Y. State
  Dep. Environ. Conserv., Albany.
Siegfried, C.A., J.A. Bloomfield, and J.W. Sutherland. 1983.
  Final report to the U.S. Environ.  Prot. Agency for the Lake
  George Clean  Lakes  Diagnostic Feasibility Study. N.Y.
  State Museum  and N.Y. State Dep.  Environ.  Conserv.
  Albany. (In prep.)
Sutherland,  J.W.,  et al.  1981.  First  Annual  Report: Lake
  George Urban Runoff Study. N.Y. State Dep. Environ. Con-
  serv. Albany.

Sutherland,  J.W..  J.A.  Bloomfield, and J.M. Swart.  1983.
  Final Report for the Lake George Urban Runoff Study, Na-
  tionwide Urban Runoff  Program.  Bur. Water Res., N.Y.
  State Dep. Environ. Conserv. Albany.
Vollenweider, R.A. 1976. Advances in defining critical loading
  levels for  phosphorus in  lake eutrophication. Mem. Inst.
  Ital. Idrobiol. 33:53-83.
Vollenweider, R.A. and P.J. Dillon. 1974. The application of
  the  phosphorus  loading  concept  to  eutrophication
  research. Can. Centre Inland Waters, Burlington, Ontario.
Wood, L.W. 1982. Trophic gradients and nutrient loadings in
  Lake George, N.Y. 1979-80. Final report to the New York
  State Department of  Education for work under the U.S.
  EPA Nationwide Urban Runoff Program. Environ.  Health
  Inst., N.Y. State Dep. Health, Albany.
Wood,  L.W., and  G.W.  Fuhs.  1979. An  evaluation  of the
  eutrophication process in Lake George based on historical
  and 1978 limnological data. Environ. Health Rep. No. 5. En-
  viron. Health Center, Div. Lab. Res., N.Y. State Dep. Health,
  Albany.
                                                     47

-------
 COMPUTER ASSISTED WATER QUALITY DATA ANALYSIS
 MICHAEL W. MULLEN
 STEPHEN  R. SMITH
 Engineering Analysis, Inc.
 Huntsville,  Alabama

 RICHARD E. PRICE
 TERRY S.  SMITH
 Vicksburg District
 U.S.  Army  Corps of Engineers
 Vicksburg,  Mississippi
             ABSTRACT

             Comparing water quality data with standards, criteria, or management guidelines can be a major
             part of any water quality assessment. The effort required for manual comparisons of data for a
             large reservoir or watershed project can become a formidable task. One solution to the problem
             is to utilize a digital computer in conjunction with appropriate software designed to compare
             numerical water quality data with Federal and State criteria or standards as  well  as user-
             specified management guidelines. Aside from reducing the labor cost of data analysis,  more ac-
             curate, rapid, and flexible analyses of water quality data can be readily performed. Utilization of
             such software can also allow a more extensive comparison of water quality data and criteria
             than would otherwise be possible. A software package of this type has been developed for the
             U.S. Army Corps of Engineers, Vicksburg District. The program, WATERCHECK, is used primari-
             ly in comparing project water quality data with Federal and State standards; a comparison
             which was not always feasible without the program. The processing of a data set which pre-
             viously required 2-4 man-days now can normally be accomplished in 1 man-day. WATERCHECK
             is especially useful for State and regional water quality management organizations responsible
             for assessment of large amounts of data. User-specified options which allow the data to be com-
             pared against criteria or standards for dif'erent water use classifications are useful for water
             resource planning purposes as well as water quality management.
INTRODUCTION

Computers have been used for a considerable time 1o
aid researchers and others working in environmental
management. Various water quality data bases such
as STORE! (EPA) and WATSTORE (USGS) are main-
tained to aid in the assessment and management of
water resources. The software associated with these
data bases can perform statistical analyses and other
manipulations of the data useful to researchers. How-
ever, until recently no general program was readily
available to assist a water resource manager in com-
paring ambient water quality data with State, Federa.l,
and other criteria and standards. This paper describes
WATERCHECK, a program developed by Engineering
Analysis, Inc. (EAI), for the Vicksburg District of the
U.S. Army Corps of Engineers. The program has been
applied to a variety of water  quality data analyses by
the Corps' Vicksburg District.
PROGRAM  DESCRIPTION

The objective of the WATERCHECK development wa.s
a general program that could readily be applied to any
State or  region to provide  cost-effective, accurate,
timely, and flexible data analysis. The method used 1o
obtain a  generalized  program was to make it data-
base-driven, based  on the inherent logical organiza-
tion of the criteria and standards. The WATERCHECK
software package currently requires a minicomputer
or mainframe. If  a  suitable  computer is available,
measured data from new regions can be analysed
simply by introducing new data files.
  State criteria or standards for water quality depend
the stream or waterbody segment, the waterbody use
designations, or general statewide standards. EPA
standards are based on classification of the water-
body uses. Thus, consistent with the organization  of
criteria and standards, the WATERCHECK tests can
be grouped as indicated in Figure 1. To compare water
quality data, a variety of tiles  are required in an inter-
nal data base. As Figure 2 indicates, the data base in-
cludes four  types of files: Criteria and  Standards,
General Operation/Information, Measurements Input
Data, and Output Data.

Internal Data Files

The basic operation of WATERCHECK involves identi-
fying the waterbody, the parameter, and the  proper
criteria or standard, then comparing the criteria  or
standard with the  measured value.  Figure 3 gives the
general logic through which the internal data files are
used by the  program: Waterbody  ID  numbers in the
IDWB files identify the State, river basin, type of water-
body (lake or river),  and waterbody segment. Water-
body use codes are stored in the same file. The TVSNS
                                                48

-------
                                                                         WATER QUALITY ASSESSMENT METHODS
file contains  ID numbers  (based  on EPA  STORE!
numbers) to designate parameters  recognized by the
program and a code indicating the units. State stan-
dards are given in three files which contain segment-
dependent (SDSS), use-dependent (UDSS), and state-
wide standards (SWSS). The EPA criteria file contains
the Federal  criteria sorted  by parameter  and use
classification.
  Two files not shown in Figure 3 are USEF, the use
code  file, and  UNITF,  the  units conversion file. The
data files and procesing sequence  of WATERCHECK
are based essentially on the logical organization of
the State water quality standards.
Input Data File (MEAST) and Special Criteria
File (SPC)

The  user's  input  data is entered  into the  program
through the MEAST file (shown in Fig. 3). Normally,
this is the only data file that has to be created by the
user.  Figure 4 is an example of a portion of a MEAST
file.  The  first  two  records  in the MEAST  file are
80-character Title and Column Headers which may be
left blank. However, the title appears in the output file
EAIOUT2 as the run ID and is useful in  identifying out-
puts.  The  third record contains  the waterbody ID
number, a Tributary Flag which can be used to number
unnamed  tributaries  to a waterbody,  and user-spec-
ified use flags by  which use  categories not currently
assigned for a waterbody segment can be added. The
fourth record is repeated for each measurement and
contains  the  parameter  ID number  which  is the
STORET number for the measured parameter. A 99999
indicates the end of a set of data for a waterbody seg-
ment,  at which point  further data  begin  with  a new
waterbody ID number. The fourth record also contains
the date, time, and depth at which the measurement or
sample was collected (all  optional  data);  and the
measurement  type code  for  water, sediment, or
elutriate data. The record makes provision to indicate
                 State Standards for
                 Water Qua!ity Data
                                    Test Type 1 = SDSS
                                       Water Body
                                    Segment-Dependent
                                       Standards
                                    Test Type 2 = UDSS
                                     Use-Dependent
                                       Standards
                                    Test Type 3 = SWSS
                                   State-Wide Standards
                                    - Specific to Lakes
                                    - Specific to Streams
                                    - General
                                    Test Type 4 = EPA
                                    EPA Water Quality
                                        and
                                    Elutriate Criteria
                                      by Use
                                    Test Type 5 = SEDM
                                    Sediment Standards
                                         or
                                       Criteria
                                    Test Type 6 = SPC
                                     User-Specified
                                    Special Criteria
that the measured data were less than the detection
limit of the analytical method. Finally, the record in-
cludes  the  measured value and the measured units
code for the measurement.
       Not a true file name
Figure 2.—WATERCHECK database functional organization.
                                       Input and Validation
                                       of Measurement Data
   UDSS
   Use Code
   Parameter ID
   Standard
   Units Code

   SWSS
   Parameter ID
   Standard
   Units Code
   EPA
   Parameter ID
   USE Code
   Criteria
   Units Code


   "SPC
   Waterbody
   Parameter
   Criteria
   Units Code
Process against
Standards specific to lakes
Standards specific to streams
General statewide standards
                       Compllance with
                       State Standards
Process against EPA criteria    1  Compliance with
based on USES             1  Federal Criter
                                                                         Process against user*
                                                                                               Compllance with Any
                                                                                               Special Criteria
Figure 1.—WATERCHECK test categories.
Figure 3.—WATERCHECK high-level logic flow.
                                                    49

-------
LAKE AND RESERVOIR MANAGEMENT
  A user can also add his own management or other
criteria for any parameter in the TVSNS file by using a
Special Criteria file (SPG). The program will check data
against  both the internal criteria and standards and
the limits input through the SPC file.

Output

WATERCHECK produces  three  types  of  output.
Messages are displayed on  the terminal screen to in-
form the user of program progress at important points
in its  execution. Two output files are created by pro-
gram execution. The first file, EAIOUT1, contains a list
of error messages and warnings.  The messages in-
clude warnings  of  a NONCOMPARABLE STOREiT
NUMBER  in the  form of  an  identification  of the
STORET  NUMBER, water value, date, time, measure-
ment  value, and other information. Typically such a
condition results from a STORET number in the inputs
for  which there are no established criteria or stan-
dards.
  An  example of the primary WATERCHECK output
file, EAIOUT2, is shown in Figure 5. EAIOUT2 gives in-
formation on each comparison made. The output iden-
tifies the test type (EPA, State tests, etc.), the water-
body use designation, the criteria limits, the  criteria
units  code, the measured value, measurement units
code, and conversion errors and finally flags measure-
ments that  fail to satisfy criteria or standards.
PROGRAM APPLICATION

The Water Quality Section of the Vicksburg District is
responsible for assessing the water quality of surface
water for the various use categories relative to Corps
projects. This involves collecting  all  available data,
summarizing it into a usable form, and assessing it for
a particular use. If the data set for evaluation is large,
04040100
00020
00010
00060
00070
00060
00095
00300
00301
00310
00400
00410
00500
00515
00530
00630
00650
00665
00300
00910
00315
00935
00340
00945
01003
01037
01053
01034
01042
01045
01051
01055
01067
01032
31501
31616
31679
33330
33340
33360
33365
33370
33380
39388
33330
39400
33410
33430
334E0
33600
33782
33333
               MISSISSIPPI R NR HELENR ORK
                    TEST BVERBGES
             110101001010010
 30. 5
 16. 7
 422.
 203.
 137.
1.313
 0. 64
0. 703
 146.
 19
1702
  £4


14725
 19B3
 1&8M?
 21.01?
 i?. 00
 0. 01
 0. as
 
-------
WATER QUALITY ASSESSMENT METHODS
JR COE VICKS&URG DISTRICT OFFICE
NG ANALYSIS. INC.
5SIPPI R NR HELENA ARK
CRITERIA LIMITS MEASUREMENT CONV EXCP
D tt PARAMETER NAME USE WIN MAX UNIT VALUE UNIT ERR TION
WATERCHECK ANALYSIS FC
ENGINEER
MISSIS
TEST
DOTE TIME WB ID tt WATER BODY NAME TRIB DEPTH TYPE P
S S Q S
Q Q Q &
r*) -T co r-n
OJ CO
Q S S S
& & S S
G> & G> Q
in a & is
oj ui in  & & s
a & & o
iiii
cr cr tr o:
cr cr cr cr
a a cr cr

cr cr cr tr
oi oi ai ui
01 01 tn tn
£ Z Z Z
|g II
Q Q S. S

Q Si S Q
S &  •-"
E 1- t- Q
. a •- .
MO: Q z
o a in CD
U E cr >
. UJ D X
u. i- H a
-I -* tv S
uj n
rr»
tn ui tn 01
a o o a
Q Si S &
S S G S
& & SG>
IIII
cr cr a cr
cr cr cr cr
era a a

cr cr tr cr
tn" oi ui tn
tn 01 in tn
z z z z
Q Q Q iii n
£G>KZHZEEEE
DUJOOOOOOOO
Z ZZZ Z ZZ ZZ
cr _i
i- _i cr
o cr t-
K 0) _l H _l O
i~ cr o a H-
01 «_F K H K »
O .O -H H t- *-» - UJ
XOHZOOOtrZI
tntnao: ^ - *_io:cr
O »CE Q Q UJ t- UJ O X
U) •* Q (*1 (1 fl ri n fl -*
oitntnoioitnoioioicn
3333333333
U101O1U1U10101O10101
-SS&S&Si&iS&S
ss&sssssse
1 t 1 1 t 1 1 1 1 1
crtrcccrcrcEcrcrcrcr
crcccrtrfrcrcrcrcrcr
crcrcracrcrcracra

croccccrtrtrcrtrcra:
oioitncnaioitntfiuiui
oioioitnoioitnoiuitn
EEZEEEEEEE
S>G>&&G>G&SrG<8

::::::::::
s s, s Q s Q s s & Q
* * * * + * *
* * * * * * *
* * * * * * *
*
S i£ -« M OJ OJ OJ OJ OJ OJ OJ OJ OJ OJ OJ OJ CU t^ ** (7* *D ^ CU OJ OJ OJ OJ OJ OJ OJ OJ OJ OJ OJ OJ OJ OJ OJ
f\J •-« OJ -*
ro co ro -« UT oj -t -« in r- uj co ri 03 uj cu -*• en *-< IIT * ® UT
OJ OJ „* **
Q U) ~> « OJ OJ OJ -i — OJ OJ ^ OJ CU AJ OJ OJ r- -H -i U3 -« OJ OJ OJ — OJ OJ OJ OJ CU OJ OJ OJ OJ OJ OJ OJ
OJ 1* OJ »*
r^ tn >-" ui -» tn in ui s M o *-• o & •* & r- m s>
ESZZZEEEEEEZEEEEEGSZGSZEZEZEZZZZZZZEEE
abTaoaoaoaaooaaoaa<Oifl&oooaoaaaaaaaaooa
Z ZZZZZZZZZ ZZZ ZZZ -> Z MZ Z.ZZ ZZ ZZ ZZZ ZZZZZ
• UJ « UJ •
t- Z H > Z J K
-i _ia_jo t- crcrj.o
n . J cr o at- *•* _i OK a t- 1-
u. t~ _j _i a H z i- »_j . z_jcr zo t- o -jj
— oi.acicri-j o M_JOCCCI .t-tn«--cii-_j _i ji-iKjOh-crcrcr
izt-zt-i- ^ocraujciuaiuxi-iDt-« z^t- .oaaocrouaaujoxt-i-
a. . LU LU u z D roaujoiii » in > ui d - .- .CTZD *-oa »o • A _J -ujx>ujtj
^QDh- — ^-«cri-i-zt- i zxxzerzza DI-«Q:)-K_JKZ i Dziuxz1--
_! .OlHUlQCCQ.OcrZZZQxt-ZE>> ^t-QCEQ.QcrtJZOSOQxn.HZUl
(*i f*i n ro n M ro fo n f*) n r*i ro
aacrcrcracraaciacrcrcrciaacccrcrcraacEacracr era aaaacraaa
0-acxaQ.Q.aQ-Q.aciQ.CLaQ.acLQ-Q.aQ.aQ.Q.aciaciaaciQ.aQ.ciaQ.a
UJLUUJLULdWLlJLUUJUJLUllJUJUJLUUJUUJLUUJUJWLyilJUJUJUJWW
Ssssssssss8ss8ssessssses8sssssssssssss
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 t 1 1 1 1 1 1 1 1 1 t 1 1 1
atrtrtracctEatratEtrtttrtrtEtratrtrtrtrortrtrtrtrtr tree trtrcctctrirtro:
crcrcrcrtrcrcECccrtrircKtrcrcrcrcrQicrcrcrcrarcEcrtrcrcz cca; cccrcticccra^crcc
aaaaaacraaaaaaaaaaciaaaaaaaaaa aa acrcr aaaaa

aorcrccarcrcrccKacrcrcrtrcrtrcrtrcrcrtrtrcrcrcrircrQ: cr cr cr or cr cr cr cr cr cr
oicnoitntnoioitJioioJtnwtnoioioioiuioitJitntnuioioitntnoioioioiwoiuiw
tnoioitnuioitnoitnmuioitnoioioitnuitnuiuitnoioioioitnoioioioiuioitnuitnw
EEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEZEZZEEZ
&&&&iS&&&SiS.'SiS'S&&&SQ&QQ&&iSS&&&&'Sr&'S'S&SliSS'S

sssssssssssess8S8Sessssesss8SSSSesssss
SS8SSS,8S888S8S88SSS8S88888888SS88S888S
Figure 5.— EAIOUT2 output file example.
*
*
en
03
UT
z
a
z
tn
01
Q
IU
13
X
O
U
a
en
&
i
tr
cr
a
tr
oi
tn
Z
s
&
-T
-t
&

-------
LAKE AND RESERVOIR MANAGEMENT

criteria or standards. Application  of the program to
other  locations by creation of the  database for addi-
tional regions should require little  or no modification
of the program. Therefore, any entity responsible lor
reviewing  large amounts of  water quality data can
benefit from applying WATERCHECK, principally in-
creasing the capability to review data without n-
creasing manpower requirements.
  A limitation of the program is that  the user must
have access  to a minicomputer or mainframe. That
limitation could be overcome by creating a version of
the program to run on microcmputers. Other limita-
tions involving the manual creation of MEAST files
can be eliminated by developing a "user friendly"  in-
put routine and a utility program to read and reformat
existing water quality data files.
                                                 52

-------
 KENTUCKY RESERVOIR ASSESSMENT OF WATER QUALITY
 AND BIOLOGICAL CONDITIONS
 NEIL E. CARRIKER
 MAHLON P. TAYLOR
 Office of Natural Resources
 Tennessee  Valley Authority
 Chattanooga, Tennessee, and Muscle Shoals, Alabama


            ABSTRACT

            Spatial and temporal variations in water quality and biological conditions were investigated in the lower
            77 miles of Kentucky Reservoir in a series of monthly surveys conducted from February to September
            1982. Factors investigated included dissolved oxygen and nutrient dynamics, algal community struc-
            ture and standing crop, primary productivity, limiting nutrients, and water chemistry. Significant
            longitudinal and seasonal variations were observed for most parameters. Only minimal variations were
            observed with depth. Although thermal stratification was not observed, weak dissolved oxygen stratifica-
            tion did occasionally develop. Reservoir hydrodynamics appears to have pronounced effects on water
            chemistry and algal productivity.
BACKGROUND

Kentucky  Reservoir, the largest  in the Tennessee
Valley system, is the last in a series of nine reservoirs
on the main stem Tennessee River. Impounded in
1944, it extends over 180 river miles from  Kentucky
Dam at Tennessee River mile 22.4 to Pickwick Dam at
river mile 206.7. Normal reservoir operation includes a
gradual drawdown from July to December to provide
flood storage capacity followed by a return to full pool
during April (Fig. 1). Principal features of  Kentucky
Reservoir are summarized in Table 1.
  Although the impoundment extends to  Pickwick
Dam, the upper 100 miles of Kentucky Reservoir is
more riverine than lacustrine, except for embayment
areas. An information search  revealed that although
Figure 1.—Operating curve for Kentucky Reservoir.
several investigations had been conducted for limited
reaches of the reservoir, primarily in the more riverine
section, only two comprehensive investigations were
available that include stations throughout the lower
end of the reservoir. These were studies conducted by
the Tennessee Valley Authority from 1966 through
1968 (Tenn. Valley Author., 1974) and the U.S. Environ-
mental Protection Agency's National Eutrophication
Survey in 1973 (U.S. Environ. Prot. Agency, 1976). Con-
sequently, an  investigation was  conducted during
1982 to assess water quality in the lower end of Ken-
tucky Reservoir. This paper briefly describes the 1982
investigation and summarizes the water chemistry
and  biological results.  A more  complete  report is
available from TVA (Carriker and Taylor, 1983). Results
of a complementary limiting  nutrient  bioassay con-
ducted by biologists from the Murray State University
Hancock Biological Station  using  200-liter  polye-
thylene bags in situ also are available from TVA (King
and Houser, 1984).
                                                  PLAN OF STUDY

                                                  The 1982 investigation consisted of a series of month-
                                                  ly water chemistry surveys and biomonthly biological
                                                  surveys conducted from February to September 1982
                            Table 1.—Principal features of Kentucky reservoir.
Water surface elevation (ft MSL)
Surface area (acres)
Volume (acre-ft)
Mean depth (ft)
Maximum depth (ft)
Controlled flood storage capacity (acre-ft)
Hydraulic retention time (days)
Shoreline (miles)
354.0*
130,000
2,121,000
16.3
S55
4,008,000


359.0b
160,300
2,839,000
17.7
£60
3,290,000
££2d
2025
375.0°

6,129,000

=75
0
see

b Normal maximum pool

c Top of flood gates

d At mean discharge of 65,290 ofs

e At maximum discharge of record (452,000 cfs)
                                                53

-------
LAKE AND RESERVOIR MANAGEMENT

(Table 2). Samples were collected on the same dates
from five  stations located in  the  main body of the
reservoir at river miles 24.0, 42.0, 65.6, 91.5, and 100.4
(Fig. 2) using two field crews working from opposite
ends of the study area. The crews met at the Hancock
Biological Station (river  mile 44.5 L) for field process-
ing of the samples and  for  incubation  of limiting
nutrient and primary productivity bioassays. Samples
were preserved  at the  time of collection and were
transported on wet ice to the  field processing pcint,
and  from  there to either the biological  laboratory in
Muscle Shoals,  Ala., or the chemical laboratory  in
Chattanooga, Tenn.
  Samples for algal identification and  enumeration
were collected at four depths in the euphotic  zone.
Enumeration was based on counts obtained using an
inverted microscope at 320X magnification, counting
either one or two perpendicular strips across the ma-
jor axis of a modified Uthermohl chamber. Two strips
were counted  for samples for which less than 500
organisms were  encountered in the first strip.
  Primary productivity and limiting nutrient bioassay
samples were transferred to 125 ml  pyrex bottles;  at
the point of collection, then placed in a dark container
for transport to the Hancock Biological Station. There
they were  inoculated with 2 ^ci of C14 - sodium bicar-
bonate; limiting nutrient samples  also received
                               nutrient inoculations as shown in Table 2. The micro-
                               nutrient cocktail used was that specified for the EPA
                               algal assay procedure (U.S. EPA, 1973). Primary pro-
                               ductivity samples were run in duplicate and were in-
                               cubated for 3 hours at the depth of collection. Limiting
                               nutrient samples were collected and  incubated for 6
                               hours  at 1  m  depth. Subsequent  treatment  and
                               analyses of these and all other samples were  con-
                               ducted in accordance with procedures specified in
                               TVA biological  and chemical laboratory  and  field
                               manuals (TVA, 1980, 1983a, b).


                               RESULTS AND DISCUSSION

                               Aquatic Chemistry

                               Water in Kentucky Reservoir, as in  the other main
                               stem Tennessee River  reservoirs, is moderately hard
                               (total hardness 60 to 70 mg/l) and slightly alkaline (pH
                               7.0 to 8.2, total alkalinity 50 to 65 mg/l). Water clarity
                               varies both temporally and spatially. Winter and early
                               spring runoff increases turbidity to  25 to 40 NTU and
                               decreases Secchi depth to as little  as 0.25 m. Clarity
                               improves  in  late  spring   and early  summer, then
                               decreases again as algal  blooms progress upstream
                               during the summer.  Secchi depth  rarely  exceeds 2
                               meters. Clarity usually increases  in  a downstream
        Table 2.—Tennessee Valley Authority Kentucky Reservoir water quality study February-September 1982.

                                             Survey Design
Frequency
           Depth (m)
                   Parameters
Monthly
Monthly
Monthly



Bimonthly


Bimonthly
0.3,1,3,5,7,9,12


Surface, middepth, 1 m off bottom



0.3,1.0,3.0,5.0
     pH, temperature, DO, conductivity (Hydrolab)
     Secchi
     Total alkalinity. NO2+NO3-N, NH3-N, total
     N, total P, dissolved ortho PO4- P, organic N,
     turbidity
     Phytoplankton enumeration (composite sam-
     ple), chlorophyll-a, C-14 primary productivity
     (two bottles each depth)
     Limiting nutrient algal assay bottle test (C-14
     primary productivity measurement, five
     replicates per treatment. Treatments: control,
     + 100ng/l N, +300 ^g/l N, +50^g/\ P, +200
     H9/I P, + 300 nQ/l N and + 200 ng/l P,
     + micronutrients)
                                               Stations
STORET
Code
210014
202833
476142
003610
475799
River
mile
24.0
42.0
65.6
91.5
100.4
Horizontal location (% left bank
looking downstream)
50
50
50
50
20
Maximum
depth (m)
22
17
17
17
15.5
 Date
               Table 3.—Occurrence of algal blooms during the 1982 TVA Kentucky Reservoir surveys.
     Stations
Genus
Cell Counts (x 1Q6/L)
3/17/82
5/25/82
7/27/82
9/21/82
TRM 24.0, 42.0, 65.6
TRM 24.0
TRM 91. 5
TRM 24.0
TRM 24.0
TRM 24.0, 42.0, 65.6
All stations
TRM 24.0, 42.0
TRM 24.0, 42.0
TRM 24.0, 42.0
All stations
Stephanodiscus
Oscillatoria
Melosira
Synedia
Melosira
Oscillatoria
Merismopedia
Melosira
Scenedesmus
Merismopedia
Oscillatoria
1.1-1.7
1.0
1.4
1.3
1.0
2.0-19.0
1.2-6.9
2.7-2.9
1.0-1.9
7.5-8.8
1.2-11.8
                                                  54

-------
                                                                       WATER QUALITY ASSESSMENT METHODS
direction as stream velocities decrease, but this trend
is occasionally reversed when high algal populations
develop in the area near Kentucky Dam.
   Surface water temperatures varied  from 7.5°C  to
30.8°C and bottom temperatures were within 2 to 3°C
of surface temperatures throughout most of the  in-
vestigation.  Although thermal stratification was never
discernible,  dissolved  oxygen (DO) concentrations
decreased with depth at all stations from May through
September.  Pronounced oxyclines (A DO^LO mg/l be-
tween two points of measurement) were observed at
river mile  65.6 on May 25,1982, and at river miles 24.0
and 42.0 on  June 23 and July 27. The greatest change
in DO occurred at river mile 24.0 on June 23 (Fig.  3).
The  upper  5 m of  the  water  column was  well-
oxygenated  at river  miles 24.0 and 42.0 throughout the
study.
   At the three upstream stations, DO values in the up-
per 5 m generally ranged between 5.0 and 6.0 mg/l dur-
ing July and August, while minimum DO levels of 3.6
to 5.0 mg/l  were present at depths greater than 5
meters. Values less than 5.0 mg/l also occurred at river
mile 65.6 at  depths  of 15 to 18 m on May 25,1982 and
at river mile 42.0 at depths of 11 to 14 m on July 27,
1982. At river mile 24.0, the lowest DO value observed
was  5.0 mg/l at depths of 7 to 12 m on  July 27.
   The effect of the  short hydraulic residence time on
DO dynamics in Kentucky Reservoir is illustrated  by
the May, June, and July data for river mile 65.6. In con-
trast to values less than 5.0 mg/l which occurred in the
             BARKLEY DAM
                                 HOPKINSVILLE
bottom water in May and almost throughout the water
column in July, the June data at river mile 65.6 showed
DO ranging from 8.5 mg/l at the bottom to 10.8 mg/l at
the surface (Fig. 4). Slightly higher stream flows the
last week of May, coupled with the  Big Sandy River
entering as a thermal interflow at river mile 67.0 during
June, temporarily disrupted the oxycline observed in
the May and July surveys.
  The February  and  July data  (Fig. 5) illustrate
seasonal differences in nutrient levels. Higher concen-
trations  occurring  in  the winter and early  spring
samples were associated with  surface runoff and high
streamflows. Nitrite plus nitrate comprised the major
fraction of total nitrogen in those samples, and most
of the phosphorus was present as particulate matter.
  By  midsummer,  total  nutrient levels were  lower,
with organic nitrogen accounting for  most of the total
nitrogen   present,   and  dissolved phosphorus  con-
stituting  a larger  fraction of  the total phosphorus.
However, as the July data for river miles 24.0 and 42.0
show, total  nutrient  concentrations,  particularly
nitrogen, may reach higher levels during summer if an
algal  bloom is in  progress or has just occurred. In-
organic nitrogen levels were uniformly low throughout
the summer, reflecting rapid uptake by phytoplankton.
  Previous studies have attributed high phosphorus
values in the reach of the reservoir immediately down-
stream from the Duck River to the nutrient load from
its inflow at river mile 110.8 R (U.S. Environ. Prot. Agen-
cy,  1976; Brye,  1970). Decreases in total phosphorus
from river miles 100.4 to 24.0 observed for the March,
                                                                25.0
                                                                        Temperature (°C)

                                                                                 26.O
                                              27.0
                                             	I
                                                                      Dissolved Oxygen (mg/l)
                                                                          7        8
                                                                                                    10
                                                      12
                                                            Temperature
Figure 2.—Location of Kentucky Reservoir and 1982 survey
stations.
Figure 3.—Dissolved oxygen and temperature profiles at
TRM 24.0 on June 23,1982.
                                                  55

-------
LAKE AND RESERVOIR MANAGEMENT
April, and August samples are consistent with those!
results. Nutrient concentration variations with depth
were minor throughout this investigation, indicating
the  significance  of  wind  mixing  and  the shor:
hydraulic retention time.
   Results of the In situ limiting nutrient bioassays in-
dicated  phosphorus  limitation,  or perhaps phos-
phorus and nitrogen colimitation of algal growth in
Figure 4.—Dissolved oxygen profiles at TRM 65.6 on May 25,
June 23, and July 27,1982.
May and July. Additions of 200 ^g/l phosphorus and of
200 ^g/l phosphorus plus 500 ^g/l nitrogen produced
significant  growth  responses  in  those  months.
Although the September algal growth response pat-
terns were similar to May and July, the magnitudes of
increase were greatly diminished. In March there was
little, if any, response to added macronutrients. One
very interesting result  was a consistent toxic or in-
hibitory effect of added micronutrients observed in
each bioassay.
  Algal community structure shifted  from  chryso-
phyte   dominance  (principally  Stephanodiscus,
Melosira,  and Chaetocerous sp.)  at all stations in
March  to  cyanophyte  dominance  (principally
Osclllatoria, Merismopedia,  Anacystis, and  Lyngbya
sp.) in  July. Blooms (>1Q6 cells per liter) of at least
one, and  as many as four different genera  were
observed at river mile 24.0 on each of the four phyto-
plankton surveys. In July and September, blooms of
Oscillatorla and Merismopedia occurred throughout
the study area (Table 3). Except for the May 25 survey,
algal cell counts were significantly higher at the three
downstream stations than at river miles 100.4 or 91.5.
The largest algal communities observed  during the
study were 58 million cells per liter at river mile 24.0 on
July 27; 39 and 33 million cells per liter at river mile
42.0 and 24.0, respectively, on Sept. 21; 27 million cells
per liter at river mile 24.0 on March 17; and 25 million
cells per liter at river mile 91.5 on May 25.
SUMMARY

The failure to detect a thermocline in the lower 77
miles of Kentucky Reservoir during this  study in-
dicates that if thermal stratification occurs, it must be
very weak and transient. Oxygen depletion at depths
greater than 5 m produced significant oxyclines in late
                                                           LEGEND

                                      Total-N   Org-N   NH3-NH4 NO2&NO3  Total P   Diss. P
    2.0-1
                                                                           I
        g   FEBRUARY 23,  1982
    1.5-E
    1.0- =
                                                                                   ttt
                                                                                             I- .20
                                                                                             -.15
                                                                                             h.io
                                                                                              .05
    2.0-,
    1.5-
    1.0-
            JULY 27,  1982
                                                                 ni
                                                                                              .20
                                                                                             -.15
                                                                                              .10
                                                                                              .05
                              T,
                                              O)

                                              I
                                              a.
         TRM  100.5          TRM 91.5          TRM 65.6          TRM 42.0

Figure 5.—Surface nutrient concentrations observed in the 1982 Kentucky Reservoir surveys.
                             TRM 24.0
                                                  56

-------
                                                                          WATER QUALITY ASSESSMENT METHODS
spring and summer, but DO levels were never depleted
below 3.6 mg/l. The short hydraulic residence time (<
22 days) and the seasonal variation of flows have pro-
nounced  effects  on  nutrient,  oxygen,  and  algal
dynamics in Kentucky Reservoir. Most of the nitrogen
was present as nitrate plus nitrite during  winter and
early spring, but by late spring organic nitrogen com-
prised  the  major fraction.  Phosphorus  showed  a
similar shift from suspended particulate phosphorus
in winter to dissolved forms in summer.
   In situ bioassays indicated that available light may
limit  algal  growth  in  March,  and that phosphorus
alone, or phosphorus and nitrogen together limit algal
growth in late spring and summer. Algal community
structure shifted  from  chrysophyte  dominance  in
winter and  early spring to cyanophyte dominance in
summer. Algal standing  crops generally were higher
at river miles 24.0 and 42.0, where 25 to 50 million cells
per liter were commonly observed in late  spring and
summer, but  blooms were occasionally observed at
the three upstream stations.
REFERENCES

Brye, B.A. 1970. Summary of observed nutrient concentra-
  tions and nutrient entrapment in TVA reservoirs. Pages
  34-51 in TVA Activities Related to Study and Control of
  Eutrophication in the Tennessee Valley—Papers Discuss-
  ed at the Meeting of the Joint Industry/Government Task
  Force on Eutrophication, Nat. Fertilizer  Develop. Center,
  Muscle Shoals, Ala., April 29-30.
Carriker, N.E., J.P. Cox, and M.L Taylor. 1984. Kentucky res-
  ervoir water quality—1982.  Tenn.  Valley Auth.  Chatta-
  nooga, Tenn.

Carriker,  N.E., and M.L Taylor. 1984. Water  quality and
  aquatic biological  conditions  in  Kentucky  Reservoir
  -February-September  1982. Tenn. Valley Author.  Chat-
  tanooga, Tenn.

King, J.M., and  G. Houser. 1983. In situ Limiting Nutrient
  Bioassays in  Kentucky Reservoir Using  200-Liter Polye-
  thylene Bags.  Hancock Biolog. Sta., Dep. Biolog. Sci., Mur-
  ray, Ky. (In press).

Tennessee Valley  Authority. 1974. Quality of water in Ken-
  tucky Reservoir. Div. Environ. Plan., Water Qual. Br. Rep.
  E-WQ-74-3.
	1980. Laboratory Branch Quality Manual. Vol. 1, 2,
  3. Div. Nat. Resour.  Oper. Chattanooga, Tenn.
       _. 1983a.  Field Operations Natural  Resource Engi-
  neering Procedures Manual. Div. Nat. Resour. Oper., Chat-
  tanooga, Tenn.
	1983b. Field Operations Biological Resources Pro-
  cedures Manual. Div. Nat.  Resour. Oper., Chattanooga,
  Tenn.

U.S. Environmental Protection Agency. 1973. Biological Field
  and Laboratory Methods for Measuring the Quality of Sur-
  face Waters and  Effluents. EPA-670/4-73-001.  Off. Res.
  Develop., Cincinnati,  Ohio.

	1976. Report  on Kentucky Lake,  Hardin, Decatur,
  Wayne, Perry, Benton, Humphreys,  Houston, Henry, and
  Stewart Countries, Tennessee, and Callaway, Trigg, Mar-
  shall, Lyons, and  Livingstone Counties,  Ky. Work. Pap.
  354, National Eutrophication Survey.
                                                     57

-------
 IRON, MANGANESE, AND SULFIDE TRANSFORMATIONS
 DOWNSTREAM FROM  NORMANDY  DAM
 JOHN  A. GORDON
 W. PAUL  BONNER
 Department of Civil Engineering
 Tennessee Technological University
 Cookeville, Tennessee

 JACK D.  MILLIGAN
 Water  Quality Branch
 Tennessee Valley Authority
 Chattanooga, Tennessee
            ABSTRACT

            During recent hearings on a unit of the Duck River Project, Columbia Dam and Reservoir, the ques-
            tion "How far downstream can water quality problems related to iron and manganese be expected
            to occur, and why?" arose. The two most prominantly unknown variables were time-of-travel below
            the dam and oxidation-precipitation rates. No rates were found for field studies and laboratory rates
            were either very high or very low. Most laboratory studies involved considerable pH shifts. Since the
            literature produced little information useful for predicting the oxidation rates of iron and manganese
            in tailrace streams, a study of iron, manganese, and sulfide kinetics was designed and performed
            at Normandy Dam, a TVA multipurpose proiect on the Duck River upstream  of the Columbia Dam
            near Tullahoma, Tenn. The study found that manganese in the Normandy tailrace exists largely in
            the Mn++ form which passes a O.V filter. Only a small percentage of the total manganese is par-
            ticulate. Mn++ is oxidized as a linear function of time-of-travel at a rate of 0.041 mg/l per hour at a
            pH of 7.1 and a temperature of 17°C. The oxidized Mn precipitate is quickly  settled and/or sorbed
            upon rocks and debris resulting m a a linear loss of total Mn with time-of-travel. The total Mn loss
            rate is 0 035 mg/l per hour at the previously stated conditions. Precipitation rates in the Duck River
            below Normandy Dam are as much as 50 times greater than rates determined in laboratory studies.
            Iron in the Normandy tailrace exists in three' forms: paniculate, colloidal, and dissolved. Exchange
            between  the colloidal  fraction and the dissolved Fe++ fraction occurs m  the river. The total iron
            decreases only slightly with time-of-travel in the Duck River. The colloidal fraction will  not settle and
            evidently the particulate fraction is too buoyant to settle Only 27 percent of the  total iron was revoved
            during 29.5 hours of travel time. The presence of Fe++ apparently keeps the  S' concentration very
            low at less than 0.025 mg S=/l both in the lake and the tailrace stream due to formation of FeS which
            is insoluble. This research was conducted during the time period of June through December 1982.
INTRODUCTION

The releases from stratified reservoirs frequently con-
tain objectionable  quantities of iron and manganese
and give rise to hydrogen sulfide odors. Clark  et al.
(1980) reported that  25  Tennessee Valley Authority
reservoirs occasionally   release concentrations  of
total iron and manganese that exceed  U.S. Environ-
mental Protection  Agency standards. The effects of
these concentrations are minimal unless a potable or
industrial water supply is located downstream  from
the dam. Then, well-known effects caused by iron and
manganese such as staining, taste problems, increas-
ed treatment costs, increased chemical usage, and
operational problems may result.
  A review of the literature by Gordon (1983) revealed
that no field studies of the transport and fate of iron,
manganese,  and sulfides in tailrace streams had  been
conducted. Laboratory studies of oxidation processes
showed  that long travel  times should be required to
oxidize manganese while sulfides and iron should be
oxidized much more rapidly. Removal mechanisms in-
clude  oxidation, autocatalytic oxidation, adsorpticn,
sedimentation, and bacterial oxidation.
  The purposes of the studies reported here were to
observe the transport and fate of iron, manganese,
and  sulfides in the  Duck River immediately  down-
stream from Normandy Dam, Tenn. The studies were
conducted at three steady-state flow  rates so that the
reaction order could be determined.
LITERATURE

Gordon (1983) summarized the literature on iron oxida-
tion as follows. The oxidation of ferrous iron in natural
waters is complex, but understandable. Iron oxidation
is rapid  at  neutral  pH values.  The conditions of
temperature, pH, and initial iron concentrations found
in trailrace  streams should  lead  to  such  rapid
chemical oxidation that bacteria cannot become in-
volved  in the oxidation process. Pankow and Morgan
(1981) presented Table  1 which shows the oxidation
half-lives for first order reactions and shows the ef-
fects of pH and autocatalysis. The Corps of Engineers
(1978) reported taking samples from the tailrace of J.
                                                 58

-------
                                                                      WATER QUALITY ASSESSMENT METHODS
Percy Priest Dam which were then stirred and oxida-
tion observed. The data show rapid oxidation wherein
Fe+ + dropped from 0.60 mg/l to<0.1 mg/l in less than
3 hours. According to Table 1, this rate would require a
pH of 7.3 to 7.5. A pH of 7.4 was observed. (Assume 10
half-lives for complete oxidation).
   Rapid chemical oxidation of H2S can be expected in
an aerobic environment.  EPA (1976) pointed out that
H2S is readily oxidized to sulfate  in aerated water. The
process minimizes the effect upon aquatic life and
rules out bacterial oxidation. The presence of Fe+ +
will preclude the presence of H2S by forming the in-
soluble FeS. However, the odor of H2S can be smelled
at a concentration of 0.0011 mg/l  according to Krenkel
and Novotny (1980).
   A review of  the literature by Gordon (1983) showed
that the mechanics of Mn + + oxidation and removal in
lakes during fall overturn and in  tailrace streams are
not completely understood. Controlled studies in the
laboratory have shown that the reaction is controlled
by temperature, pH,  autocatalysis, and  sorptive sur-
faces. Because the Mn+ + oxidation rate is chemical-
ly slow at neutral pH values, bacterial oxidation may
be an important process.
   Wilson (1980)  presented Figure 1 which shows the
"quarter-life" (125%) values for Mn+ + , assuming first
order kinetics, as a function of pH and pMn + +. A con-
centration of 2 mg/l Mn+ + at a pH of 7.0 would take
25 to 30 years to oxidize to a level of 1.5 mg/l! Wilson
(1980) included only chemical oxidation rates in Figure
1.
   Delfino and Lee (1968) aerated Lake Mendota bot-
tom  water with compressed air. Dissolved oxygen
reached saturation in 2 hours and the pH increased to
8.5 because of  CO2 scrubbing.  Figure 2  shows the
results.  The reaction appears to be linear,  but has
been assessed as first order also. Wilson (1980) used
the Delfino and Lee (1968) data from Figure 2 to check
his work. They had assumed first order kinetics, used
a pH of 8.5 and a pMn+ + of 4.95 (0.6 mg/l) and  got a
t25%  value of 10.5 days. They thought that this  value
agreed well with the value of 7 days obtained from the
graph.
   The Corps of  Engineers (1978) stirred a sample of
manganese-containing water from the tailrace of  J.
Percy Priest Dam and produced the data of Figure 3.
Mn+ + oxidation was fairly slow at the 7.4 pH values
and a t.25% value  of 180  hours  was produced. This
value is much less than the 5 years required according
to Figure 1.
   Since Mn++  is somewhat stable at neutral  pH
values, bacterial oxidation is a good possibility.  Many
species of Mn + + oxidizing bacteria are known and all
forms seem to be attached organisms which deposit
Mn02. These bacteria are not well-studied but appear
to be aerobic  autotrophs  which use an anaerobic
   Table 1.—Effect of pH and initial iron concentration on
          ferrous oxidation half-lives (seconds).
              (Pankow and Morgan, 1981).

              Initial Molar Concentration
 PH
10-3M
          10-«M
10 -7
8
7
6
1 sec.
102sec.
104 sec.
102 sec.
10" sec.
106 sec.
10" sec.
106 sec.
108 sec.
 Notes: (1) Half-life is time required for the reactant to be consumed by Vi
     assuming first-order kinetics
     (2) M x 56,000 = mg Fe/l
                                        resource (Mn+ +). As such, they would find the tail-
                                        race stream to be a good environment for growth.
                                        EXPERIMENTAL METHODS

                                        Three field studies were conducted with steady-state
                                        flows at the dam of 43,119 and 180 cfs. Several sampl-
                                        ing stations were located along a 21 km (13-mile)reach
                                        Figure 1.—Values of t25% calculated by Wilson (1980) and
                                        plotted as functions of pH and pMn+ +. The lines pH 9 and
                                        pMn+ + = 6 delineate the limits of common environmental
                                        conditions.
                                                                                    • 9
                                                                                    . 7
                                          0.0
                                                            12   16   20    24

                                                         Aeration Time, Days
                                                                          28
                                       Figure 2.—Laboratory aeration experiment.  Anoxic  Lake
                                       Mendota, Wise., bottom water taken from 22 m depth in deep
                                       hole region Sept. 30,1967 (from Delfino and Lee, 1968).
                                                                     30   90    50   60    70

                                                                         Stirring Time (hours)
     OlFe"1
Fe1
                                       Figure 3.—J. P. Priest study on Mn+ + oxidation kinetics,
                                       1973 (from Corps of Engineers, 1978).
                                                  59

-------
 LAKE AND RESERVOIR MANAGEMENT



    1.0




    0.9




    0.8
    0.7
 I"  0.6
    0.5
    0.4
   0.3
   0.2
   0.1
                              Total  Iron
             Particulate Fe
                      0.45)i Filtered
            10    20     30    40    50     60    70

                   Time of Travel (hours)
Figure 4.—Iron forms found in the Duck River below Norman-
dy Dam, Tenn., at 43 cfs, 8/9/83.
      - Participate Fe
 o 0.4
       0.45u Filtered
 of the Duck River. Time-of-travel, flow, and constituent
 concentrations were measured at each station. In ad-
 dition, routine analysis of the stream for temperature,
 dissolved  oxygen, conductivity, pH, and  oxidation-
 reduction  potential was performed. Constituent con-
 centrations were corrected for dilution by ratioing bas-
 ed upon the flow increase along the stream. Metals
 were measured by atomic absorption methods follow-
 ing filtration through 0.45^ and 0.1^ filters.


 RESULTS

 Sulfides. Sulfides were not detected in the Normandy
 Lake hypolimnion nor in the tailrace stream during the
 first study. (The detectable limit was less than 0.025
 mg/l). These data agreed with the literature in that
 Fe+ + ions will react with sulfides to form the insolu-
 ble FeS. Sulfide  odors were noted  at the dam site
 where turbulent conditions stripped H2S from large
 volumes of water. (As noted earlier, H2S is detectable
 at 0.0011 mg/l by human smell.) CRC (1974) reported
 the solubility of FeS to be 6.2 mg/l in cold water. Thus,
 it became evident from the first that sulfides were not
 present in the Duck River downstream from Normandy
 Dam. Sulfides were deleted from the two subsequent
 studies.
   Iron. The iron data exhibited some rather surprising
 trends which should form the basis of a rethinking of
 current methods  for measuring soluble  iron. During
 the first two studies at 43 and 119 cfs, little Fe+ + was
 present as measured in the filtrate from the 0.V filter.
 These two data sets had fluctuating Fe+ +  levels at
 less than  0.2 mg/l. All three surveys found a sizable
 colloidal  fraction between the 0.1/* and 0.45^ size
 range. Neither the total nor particulate iron fractions
 had a significant tendency to settle out of the flowing
 water. These data sets are shown by Figures 4, 5, and
 6.
   Essentially, these iron studies show that iron in the
 Duck River has a significant colloidal fraction which is
                   10             20
                    Time of Travel (hours)
                   10              20

                    Time of Travel (hours)
Figure 5.—Iron forms found in the Duck River below Norman-
dy Dam, Tenn., at 119 cfs, 9/9/82.
Figure 6.—Iron forms found in the Duck River below Norman-
dy Dam, Tenn., at 180 cfs, 8/16/83.
                                                   60

-------
                                                                         WATER QUALITY ASSESSMENT METHODS
slow to agglomerate. If the true dissolved fraction is
desired, samples must be filtered through 0.1 (V filters.
  At the highest flow rate, soluble Fe+ + was present
at the dam at an initial level of 0.89 mg/l. It rapidly ox-
idized to a level of 0.30 mg/l before beginning the fluc-
tuations shown by Figure 6. The soluble iron oxidation
data showed a first order tendency (Fig. 7). If only the
first four data points are used, the half-life of Fe + + in
the Duck River is 4 hours.  This  is longer than  the 20
minutes  predicted by Table 1. Probably the data  are
somewhat inaccurate because  of the tendency  for
some iron to  redissolve due  to dilution as the flow
goes downstream. In addition, if the reaction is both
exponential and autocatalytic, then one would expect
the oxidation rate to continually decrease at constant
temperature and  pH. The  reach-by-reach  oxidation
times did increase as shown by Figure 7, but were still
higher than theoretical half-lives as shown by Table 2.
Thus, in the Duck River, ferrous oxidation occurred
rapidly but at  lesser rates than those of Pankow and
Morgan (1981).
  Manganese.  Manganese oxidation  was clearly
observed in all three surveys. All three yielded oxida-
tion rates which could be explained by first-order reac-
tion rates. Manganese appeared as either a dissolved
fraction which passed a O.V filter or a particulate frac-
tion. Very little  colloidal  manganese  was present.
Total manganese was removed at about the same rate
as  Mn++  was oxidized.  The  results of  the  three
surveys are shown by Figures 8, 9, and 10.
  Correlation  coefficients  based  upon a  log-trans-
formation of the dissolved  manganese concentration
(passing a O.V filter) were 0.99 for the 43 cfs study;


  Table 2.—Iron oxidation half-lives in Duck River, Tenn.,
          in contrast to predictions from Table 1.

 Reach   pH   Initial Fe(M)  Measured t'/>   Predicted t**
0-1
1-2
2-3
3-4
0-4
7.3 1.59x10-5
7.5 1.18x10-5
7.6 8.04x10-6
7.7 5.36x10-6
7.5 1.59x10-5
1.2 hours
6.0 hours
3.5 hours
9.0 hours
4.0 hours
0.60 hours
0.24 hours
0.50 hours
0.36 hours
0.30 hours
 'Values interpolated from Table 1
    0.9<
    0.8

    0.7

    0.6


    0.5
Statistics (First four points only)

Intercept = 0.816 mg/l
Slope = -0.743 hr-'
Corr  Coef. = -0.980
                             0.97 for the 119 cfs study; and 0.97 for the 180 cfs
                             study. Quarter-life (t25% values) were 5.95 hours at 43
                             cfs; 4.79 hours at 119 cfs; and 4.53 hours at 180 cfs.
                             These short hourly values are clearly in contrast to the
                             long t25% times reported by Wilson (1980), Delfino and
                             Lee (1968) and the Corps of Engineers (1978).
                               There are two possible explanations for the rapid
                             rate  of manganese oxidation and  removal in the
                             stream as contrasted to the laboratory studies: (1) bio-
                             logical oxidation and  (2)  sorption  on manganese =
                             coated surfaces. Both  of these processes  are known
                             to operate according to Gordon (1983). Further studies
                             will be required to identify these mechanisms.
                              1.0
                              0.9
                              0.8
                              0 7

                              0 6
                            ^ 0.3
                            f
                            „ 0 10
                            £ 0.09
                            5 0.08

                            I °-07
                            * 0.06

                             0.05

                             0.04
                                                                            Ti«e of Travel (hours)
                                                      Figure 8.—Manganese forms found in the Duck River below
                                                      Normandy Dam, Tenn., at 43 cfs, 8/9/83.
                     20      30

                  Time of Travel  (hours)
                                             10      15     20
                                                  Time of Travel (hours)
Figure 7.—Fe+ + oxidation in the Duck River below Norman-    Figure 9.—Manganese forms found in the Duck River below
dy Dam, Tenn., at 180 cfs, 8/16/83.                          Normandy Dam, Tenn., at 119 cfs, 9/9/82.
                                                    61

-------
 LAKE AND RESERVOIR MANAGEMENT
                        15     20      25
                       Time of Travel (hours)
Figure 10.—Manganese forms found in the Duck River below
Normandy Dam, Tenn., at 180 cfs, 8/16/83.
    0.10


    0 OS



    0 06

  ~ 0.05

  S o 04
  >
  o
  " 0.03
Statistics
Intercept = 0.010 hr-1
Slope = 0.192
Corr. Coef. - 0.996
                           60  80 100
                          : Nunlraiidy yarn in cfs
Figure 11.—The log-log relationship between manganese ox-
idation  rate and flow in the Duck River below Normandy
Dam, Tenn.

  The relationship between  the oxidation rate for
manganese and flow at the dam was found to be a log-
log relationship. Figure 11 shows the virtually perfect
fit of these data to the model. This figure will allow
forecasts fo the Mn+ + oxidation rate at other flows.
                              CONCLUSIONS

                              Based upon the data taken during these three surveys,
                              the following conclusions were drawn:
                                1. Iron in  the releases from Normandy Dam was
                              mostly in a  particulate or  colloidal  form—very little
                              was truly dissolved in the  Fe++ form. In the single
                              study when Fe+ + was present at about 1  mg/l, oxida-
                              tion was very rapid but did not occur as rapidly as
                              some lab studies had predicted. Fe+ + did appear to
                              follow a pH-dependent, autocatalytic model during  ox-
                              idation.  The  use of 0.45^ filters for  measurement of
                              Fe+ +  is not realistic.
                                2. Manganese is removed at a very rapid rate com-
                              pared to lab-derived rates. First-order kinetics describ-
                              ed both the oxidation and removal of manganese. The
                              first-order  reaction  rate  was  flow  dependent  in a
                              perfect  log-log relationship.  This  will allow good
                              predictive capability. Manganese is most likely being
                              removed by sorption onto MnO2 coated surfaces or is
                              being oxidized  by  autotrophic bacteria.  Manganese
                             does  not exhibit a pronounced colloidal form and
                              Mn+ +  may  be assessed using  either 0.45 or 01u
                             filters.
 REFERENCES

 Clark, L.R., et al. 1980. Is the water getting cleaner, a survey
  of water quality in the Tennessee Valley. Water Qual. Br.,
  Tenn. Valley Author., Chattanooga, Tenn.

 CRC. 1974. Handbook of Chemistry and Physics. CRC Press.
 Delfino, J.J., and G.F. Lee. 1968. Chemistry of manganese in
  Lake Mendota, Wis. Environ. Sci. Technol. 2:12.

 U.S.  Environmental Protection Agency. 1976. Quality Criteria
  for Water. Washington D.C.

 Gordon, J.A.  1983. Iron, manganese and sulfide mechanics
  in  streams and lakes — a literature review. Rep. No. TTU-
  CE-83-2. Tenn. Technological Univ.,  Cookeville.

 Krenkel, P.A., and V. Novotny. 1980. Water Quality Manage-
  ment. Academic Press.

 Pankow, J.F. and J.J. Morgan. 1981. Kinetics for the aquatic
  environment. Environ. Sci. Technol.  15:10.
 U.S. Army Corps of Engineers. 1978. Water quality conditions
  in J. Percy  Priest Reservoir. Nashville District, Tenn.
Wilson, D.E. 1980. Surface and complexation effects on the
  rate  of  Mn(ll)  oxidation in  natural  waters. Geochim
  Cosmochim Acta. 44:1311.

-------
APPLICATION OF MULTISPECTRAL  DIGITAL IMAGERY TO THE
ASSESSMENT OF PRIMARY PRODUCTIVITY IN FLAMING GORGE
RESERVOIR
JAMES VERDIN
Bureau of  Reclamation
Denver, Colorado

DAVID WEGNER
Bureau of  Reclamation
Salt  Lake City, Utah
            ABSTRACT

            In support of the Bureau of Reclamation effort to manage eutrophication trends in Flaming
            Gorge Reservoir, Utah and Wyoming, remote sensing studies employing multispectral digital im-
            agery were undertaken. Specifically, the imagery was used to extrapolate point measurements
            of chlorophyll concentration and light penetration to characterize conditions throughout the
            reservoir. In 1981 and 1982, aircraft and satellites acquired digital imagery of the reservoir con-
            current with  surface sampling. Sampling stations were identified within the imagery and the
            digital values at these points recorded. The surface measurements of chlorophyll and light pene-
            tration were regressed against the digital image values to obtain predictor equations for these
            indicators of primary productivity. The equations were then applied to the images of the whole
            reservoir to obtain maps of the distribution of chlorophyll and light penetration. A unique set of
            equations was defined for each data-gathering date. The favorable results obtained in the date-
            specific studies prompted a search of data archives for other dates for which surface sampling
            and satellite  imagery were both available. An atmospheric radiative-transfer model was also ob-
            tained to allow for correction of imagery for sun angle and atmospheric effects. This effort ser-
            ved  two purposes: (1) to permit development of  chlorophyll and  light penetration predictor
            models from  as many surface observations as possible, regardless of date of collection, and (2)
            to permit  estimation  of reservoir conditions from imagery for which no concurrent surface
            sampling was available. Seven satellite scenes were processed to estimate and map chlorophyll
            and light penetration, although concurrent surface sampling data were available for only four of
            them. The maps produced were used by reservoir managers to follow year-to-year trends in pri-
            mary productivity.
INTRODUCTION

Flaming Gorge Reservoir in Utah and Wyoming, is an
integral component of the U.S. Bureau of Reclamation
Colorado River Storage Project. Its primary function is
water storage for hydropower, irrigation, and down-
stream water commitments. The reservoir is also used
extensively for boating, camping, and fishing. Since
initial closure of  Flaming  Gorge Dam in 1963,  the
reservoir developing behind the dam has been subject
to variable trends  in eutrophication. During initial fill-
ing of the reservoir, extensive eutrophication and sub-
sequent algal blooms  resulted as nutrients were
leached from  the reservoir bottom and became avail-
able for biological productivity. As the reservoir aged
over the past 20 years, the eutrophication in the upper
end of the reservoir has been maintained at high levels
primarily as a result of inflowing nutrient loads and
nutrient recycling  from  the sediments. The size and
hydrodynamic complexity of the reservoir result in a
continuum of water quality conditions at any one time.
  The U.S. Bureau of Reclamation initiated studies to
evaluate these water quality conditions in response to
two main areas: salinity trends in the Colorado River
Basin and determination of the impact of  reservoir
management  policies on fishery and recreation both
upstream and downstream  of Flaming Gorge Dam.
Limnological  surveys of water quality profiles and
plankton  trends were initiated. To better define  the
surface water quality gradients, a remote sensing pro-
ject using surface chlorophyll a and transparency was
initiated.
  Designing a surface sampling scheme for Flaming
Gorge  Reservoir  poses  a  considerable  challenge
because of its large size, remote location, and hydro-
dynamic complexities. Enough sites must be included
to identify all zones of limnological significance while
staying within the limits imposed  by available time,
funding,  personnel,  and equipment.  Multispectral
remote sensing  in  the  sampling  effort  was  incor-
porated to extrapolate and increase the information
gained by direct measurements of reservoir water pro-
perties.
  Multispectral  remote sensing has been used to ad-
vantage in a number of lake and reservoir water quali-
ty studies. Martin et  al. (1983) and Lillesand et al.
(1983)  used  Landsat imagery to  perform trophic
classification of large numbers of lakes in Wisconsin
and Minnesota.  Meinert et al. (1980) and Grimshaw et
al. (1980) performed similar studies of large reservoirs,
and also used  the  data to portray  within-reservoir
variations  of  water  quality parameters.  Mace (1982)
successfully used airborne MSS (multispectral scan-
ner)  imagery to map water  quality parameters  in
Flathead Lake,  Mont. Witzig  and Whitehurst (1981)
provide a useful review of the literature describing the
                                                 63

-------
 LAKE AND RESERVOIR MANAGEMENT
 application of remote sensing to surface water quality
 studies.
 DESCRIPTION OF STUDY AREA

 Flaming Gorge Reservoir is a large impoundment at
 an elevation of 2,000 meters on the Green River, a ma-
 jor tributary of the Upper Colorado River Basin. The
 dam, located in northeastern Utah, was completed in
 1963. The reservoir extends 125 kilometers upstream
 of the damsite, well into southwestern Wyoming. An-
 nual filling of the reservoir is achieved primarily by Ihe
 spring snowmelt in the Uinta, Wind River, and Wyom-
 ing ranges of the Rocky Mountains. The stored water
 produces hydroelectric power and meets downstream
 water supply  commitments through the year. The
 reservoir and  adjacent  lands comprise the Flaming
 Gorge National Recreation  Area.
   Flaming Gorge  Reservoir can  be separated into
 three hydrodynamic zones:  the riverine  inflow area, a
 transition zone, and the lacustrine main body of Ihe
 reservoir (see Fig. 1). The inflow area is  characterized
 by high turbidity, extensive mixing, and high  nutrient
 concentrations, whereas the transition  zone exhibits
 reduced sediment turbidity and increased primary pro-
 ductivity. Waters of the main body of the reservoir are
 very clear with conditions of relatively low primary pro-
 ductivity.
   Wegner (1982) has documented and analyzed soa-
 sonal  changes in  the spatial distribution of chloro-
 phyll a concentrations and Secchi transparencies
 since the initial filling of the reservoir. His analysis
 showed a general decrease in primary productivity as
 one moves down reservoir, although peak chlorophyll
 concentrations were usually seen in  the transition
 zone.  Secchi  transparency increases  in a  regular
 fashion as one moves downstream from the turbid up-
 per arms of the reservoir.
DATA ACQUISITION FOR DATE-SPECIFIC
ANALYSES

Reservoir Survey of September 1981. The Bureau of
Reclamation conducted a comprehensive  limnolog-
ical survey of Flaming Gorge Reservoir on Sept. 9 and
10, 1981, visiting 18 sites in the upper three fourths of
the reservoir on the first day, and seven more in the
lower reaches  on the second day. Manifestations of
eutrophication are normally at a peak at this time of
year. At  11:20  a.m.  MDT on Sept. 9, the Landsat  2
satellite acquired a multispectral digital image of the
reservoir. The  spectral  bands sensed by this instru-
ment are listed in Table  1. Between 9:40 and 10:50 a.m.
that  same  date, the U.S.  Environmental Protection
Agency acquired digital imagery of the reservoir with
its airborne eleven channel MSS (bands listed in Table
2). The instrument was flown at an altitude of 6,560
meters MSL, yielding an 11.5 meter nominal pixel size.
Color aerial photography was acquired simultaneous-
ly.
  Reservoir Survey  of September 1982. The Bureau
again sampled Flaming Gorge Reservoir on Sept. 24,
1982, along with Wyoming Game and Fish, and the
Utah Division  of Wildlife Resources. Secchi  trans-
parency  measurements  were made at  56  sites
throughout  the length of the reservoir. Chlorophyll  a
concentrations were determined for samples drawn
from 0.1  foot depth for 16 sites in the upper end of the
reservoir (Campbell,  1983). Very significant blooms of
 blue-green algae were present, and therefore very high
 concentrations of chlorophyll a were observed.
   The U.S. EPA airborne MSS was again used to ob-
 tain digital imagery of Flaming Gorge  Reservoir  on
 Sept. 24, 1982. Two flights lines were flown at 6,250
 meters MSL between 10:40 and 11:10 a.m., MDT, with a
 nominal pixel size of 11 meters square resulting. Im-
 agery was acquired only for the upper third of the

 Table 1.—Band passes sensed by the Landsat multispectral
                      scanner.
Channel
4
5
6
7
Wavelength band (^m)
0.50 - 0.60
0.60 - 0.70
0.70 - 0.80
0.80- 1.10
Color/spectrum
Green
Red
Near infrared
Near infrared
Figure 1.—Generalized map of the Flaming Gorge Reservoir,
Utah and Wyoming.
                                                64

-------
                                                                        WATER QUALITY ASSESSMENT METHODS
reservoir. Landsat  4 passed over the reservoir  the
following day, but cloud cover prevented  the acquisi-
tion of usable imagery.
  Correlation of Surface Sample and Remote Sensing
Data. A computer-driven color video display with cur-
sor identified surface sample sites in the digital im-
agery. Digital counts at these points were  recorded for
use in regression procedures with the surface mea-
surements.  Clouds and cloud shadows obscured the
lower end of the reservoir in the  Landsat scene, so on-
ly 20 of the  25 sites could be identified.
  Measurements of Secchi transparency were avail-
able for all  surface sample sites. Chlorophyll  a con-
centrations  had been obtained  for 11 of  these sites,
and at  6 of them  detailed  measurements of  con-
ductivity,  dissolved  oxygen,  and  nutrient   con-
centrations  were made.
  Using stepwise linear  regression,  equations were
developed with  water quality parameters as depen-

Table 2.—Band passes  sensed by  the EPA airborne multi-
                  spectral scanner.
 Channel
Wavelength band (^m)     Color/spectrum
1
2
3
4
5
6
7
8
9
10
11
0.38
0.42
0.45
0.50
0.55
0.60
0.65
0.70
0.80
0.92
8.0
-0.42
-0.45
-0.50
-0.55
-0.60
-0.65
-0.69
•0.79
-0.89
-1.10
-14.0
Near ultraviolet
Blue
Blue
Green
Green
Red
Red
Near infrared
Near infrared
Near infrared
Thermal infrared
dent variables, but before this mean digital counts
were converted to radiance values (mW  cm2.sr) by
means  of  the calibration  constants  tabulated by
Richardson et al. (1980). Candidate independent vari-
ables were radiance  ratios and chromaticity  ratios.
(The band 4 chromaticity ratio for a point would be the
band 4 radiance divided by the sum of bands 4, 5, and
6 radiances.) Useful equations were identified for Sec-
chi transparency, chlorophyll  a concentration, and
total phosphorous concentration. The equations are
summarized in Table 3. Water quality maps were then
prepared by applying  the equations to digital images
of the reservoir. In this fashion, the water quality mea-
surements  made at sample sites were extrapolated to
characterize the entire water surface appearing in the
image.  Color  coding was used to  make the water
quality images easier to interpret.
   Stepwise regression was applied in  an  identical
manner for the airborne MSS data of Sept. 9,1981. The
equations obtained appear in Table 4. They were used
to produce color coded  images of chlorophyll a and
Secchi transparency spatial variability for  four cloud-
free reaches of the reservoir. Sixteen chlorophyll a sta-
tions appeared in the  Sept. 24,1982, airborne MSS im-
agery. Observations from 17 Secchi stations above the
confluence of the Green River and Blacks Fork Arm
were used  to establish a comparison with  the remote
sensing data.  MSS data  in the form of eight bit, 0-255,
digital  counts were used in the regression procedure.
Conversion to radiance values prior to  regression
analysis is  advisable if spectral ratios are to be includ-
ed as  candidate independent  variables. However,
analysis of the Sept.  9,  1981, airborne MSS data did
not yield  any equations  with spectral   ratios as
variables, so they were dispensed with in the analysis
of the 1982 imagery. The predictor equations obtained
are summarized  in Table 5. The 1982 chlorophyll a
                  Table 3.—Water quality predictors based on Landsat radiance values, Sept. 9,1981.
           Equation
                                                 Standard Deviation
                                                    About Mean
                                         Number of
                                           Sites
 SD = 85.6 e -604 R4

 CHLA = 24.8 e (603 R6 - 292 CR5)

 TP = 0.270 - 0.667^
                              0.93
                              0.94
                              0.89
        2.4ft
        1.2 mg/m3
        0.0087 mg/l
20
11
 SD - Secchi transparency
 CHLA - chlorophyll a concentration
 R4, R5, and R6 refers to radiance values for bands 4, 5, and 6 of Table 1
 CR5 - Chromaticity ratio for band 5 of Table 1


               Table 4.—Water quality predictors based on airborne MSS radiance values, Sept. 9,1981.
           Equation
                                                 Standard Deviation
                                                   About the Mean
                                         Number of
                                           Sites
 SD = 24.5 e -OSSRS
 CHLA = 0.482 e <637 R7 - 1 35 R4)
                              0.94
                              0.91
        2.1 ft
        1.5 mg/m3
19
10
 SD - Secchi transparency
 CHLA - chlorophyll a concentration
 R4, R5, R7 refer to radiances for bands 4, 5, 7 of Table 2
                Table 5.—Water quality predictors based on airborne MSS digital counts, Sept. 24,1982.

1
— = o
SD
CHLA =
Equation
.156 + 0.0393 DC7
0.02065 e ° 9°? DCS
(2
0.63
0.88
Standard Deviation
About the Mean
1.9m
30.2 mg/m3
Number of
Sites
17
16
 SD - Secchi transparency
 CHLA - chlorophyll a concentration
 DC7 and DC8 refer to digital counts for bands 7 and 8 of Table 2
                                                   65

-------
LAKE AND RESERVOIR MANAGEMENT
equation is seen to have a fit, as measured by the co-
efficient of determination (r2), nearly as good as that
obtained for the 1981 mission. However, the standard
deviation about the mean is twenty times greater. This
is due to the difference in the ranges of chlorophyll a
concentration observed on the two dates. On Sept. 9,
1981, the observed range was  0.88-20.00 mg/m3; on
Sept. 24, 1982, the range was 4.99-2307.53 mg/m3


MULTIDATE LANDSAT ANALYSES

Encouraged  by  the   results   of  the date-specific
analyses, it  became  apparent to the authors  that
Landsat MSS imagery, with  its  16-day  repetitive
coverage, could be of even greater utility if sun angle
and atmospheric effects in the data could be removed.
This would permit the development of water quality
predictive regression equations from concurrent sur-
face and satellite observations for all available dates
at once. Such regressions could be used (along with
the atmospheric  correction  procedures) to estimate
water  quality  conditions with  imagery obtained on
dates  when there were no  sampling crews on the
reservoir.
   A search of Landsat and water quality data archives
revealed that  in  addition to Sept.  9, 1981, surface
sampling had been concurrent with Landsat image ac-
quisition on Sept. 22, 1975, Aug. 2, 1978, and Aug. 24,
1982. Scenes showing the reservoir on Sept. 17, 1976,
Oct. 18, 1977, and Oct.  13, 1978, were also purchased
from the EROS Data Center, Sioux Falls, S.D. No sur-
face sampling had been carried out during the 1976
through 1978 algal bloom seasons, and it was hoped
that analysis of  these  images would fill an  informa-
tional gap. The choice of dates for these seasons was
based on the availability of cloud-free imagery and the
judgment of those  familiar with the reservoir during
the period in question.
   Ahern et al's. method (1977) was selected for correc-
tion of the imagery  since it  requires  no ground
measurements of  solar radiation. This approach re-
quires that an oligotrophic standard reflector appear
in the scene. An area in the main body of the reservoir
(see Fig. 1) was selected for this purpose. This area
was chosen because it  is consistently characterized
by extremely  clean, deep waters free of bottom  and
edge effects. The technique also requires the applica-
tion of the deterministic atmospheric radiative trans-
fer model of Turner and Spencer (1972). A full descrip-
tion of the application of the technique is provided by
Verdin (1983). Airborne MSS images were not included
in  the  multidate  analyses because their scene
geometry is much more complicated than that of
Landsat images.
   Least-squares regression procedures were used to
produce estimator equations for Secchi transparency
and chlorophyll a concentration in terms of Landsat-
derived reflectances.  For Secchi transparency, the
equation found was:
      1
    	 = 0.0665 + 35.6 p5                   (1)
     SD
              where SD is Secchi transparency (meters), and p5 is
              the  apparent Lambertian reflectance for band  5 of
              Table 1. This equation was derived from data for 69
              observations, with a coefficient of determination (r2) of
              0.94 and standard deviation about the mean of about
              1.5 meters. The goodness of fit is illustrated in Figure
              2. The chlorophyll a predictor obtained was:
                  CHLA = 1.37e(i07p6)
                                                            (2)
             where CHLA is chlorophyll a concentration mg/m3)
             and p3 is reflectance for band 6 of Table 1. Forty-four
             observations were used to develop this equation with
             r2 = o.74 and standard deviation about the  mean of
             about 3.0  mg/m3. Figure 3 illustrates its goodness of
             fit.
               Development of equations 1 and  2 made  possible
             the production  of color-colded water quality images
             for  all seven dates of interest. From an examination of
                              0030     OOTB     0.100      0.123     01 BO
                            REFLECTANCE I/)) IN THE 0 60 - 0 70 /ifn BAND
             Figure 2.—Secchi depth goodness-of-fit  for the data col-
             lected at Flaming Gorge Reservoir.
             Figure 3.—Chlorophyll a goodness-of-fit for the data col-
             lected at Flaming Gorge Reservoir.
      Table 6.—Intervals of Secchi transparency and chlorophyll a concentration and corresponding trophic states.


  Trophic State
    Secchi
Transparency (m)
    Chlorophyll a
Concentration (mg/m3)
 Oligotrophic
 Mesotrophic
 Eutrophic
     >3
      1-3
        >4
        4-10
                                                 66

-------
                                                                     WATER QUALITY ASSESSMENT METHODS
historical records of Secchi transparency and chloro-
phyll a concentration and the digital imagery, it was
determined that the trophic intervals of Table 6 would
yield meaningful maps of trophic spatial variability.
For each scene of interest, then, digital count color
coding schemes were used to define two trophic state
maps,  one based  on  Secchi  transparency and one
based  on chlorophyll a concentrations. Color photo-
graphic prints of these digital images were produced
for limnological interpretation and analysis of the an-
nual trends portrayed.


DATA INTERPRETATION  CONSIDERATIONS

One point of consideration in  the production of color
coded  water quality maps is the need to distinguish
between areas of organic and  inorganic turbidity. For-
tunately, these two forms of turbidity area are normal-
ly spatially discrete in Flaming Gorge Reservoir. Areas
of high suspended sediment  concentrations exhibit
relatively high radiances and  as such would be por-
trayed  as  areas of very high  chlorophyll a  con-
centration if the regressions were applied here. In fact,
chlorophyll a is normally quite  low  in these  areas
because available light for photosynthesis is limited.
In order to avoid confusion, the Secchi relationships
were used to mask out areas of high suspended sedi-
ments  prior to  the application of  the chlorophyll a
regression. This was done by consulting meteorologi-
cal  records for the days  preceding image acquisition,
people who had been in the field and by inspecting
aerial photographs and digital imagery to interpret a
threshold Secchi depth, as determined by regression,
to segregate organic and inorganic turbidity. In 1981,
all waters with Secchi transparency less than or equal
to 1.0 meter were treated as areas of inorganic turbidi-
ty.  In  1982 the threshold was interpreted to be 2.3
meters and for multidate analysis  it was  0.5 meter.
The exact value used seemed to depend on the fit of
the Secchi regression.
  Erroneously high estimates of chlorophyll a concen-
tration were also obtained in some  areas due to edge
and bottom effects. Pixels with radiance contributions
from   adjacent  shoreland  and/or  bright  bottom
material had this problem. Fortunately, the areas af-
fected  were limited. The availability of natural color
aerial photographs obtained simultaneously with the
airborne MSS data was a  great help in identifying
problem areas.
  Clouds  and their shadows  eliminated large areas
from analysis. In September 1981 the lower fourth of
the reservoir could not be imaged for this reason. On
Sept. 25, 1982,  clouds obscured the entire reservoir
from the view of the Landsat 4 MSS.
  Specular reflectance (sun glitter)  in one case made
a portion of the airborne MSS data for Sept. 9, 1981,
unusable for water quality analysis. A flight line made
at 9:05 a.m. MDT heading south suffered from this ef-
fect. If it had not been eliminated  from analysis, er-
roneous estimates of Secchi depth and chlorophyll a
concentration would have resulted.  Landsat MSS data
are only rarely subject to this effect, thanks to its nar-
row scan angle (relative  to that of the airborne MSS).
LIMNOLOGICAL INTERPRETATIONS

Historically, remote sensing techniques have primari-
ly  made one-time assessments of a body of water,
usually a lake. Two items make this study unusual: (1)
it was applied over a variety of dates  to assess
seasonal and annual variability and (2) it was applied
to a highly complex reservoir environment.
  Remote sensing  quantification over a variety of
seasonal conditions, allows for the identification and
evaluation of the extent of productivity levels and sur-
face hydrodynamic  characteristics. The definition of
surface water quality conditions  can be correlated
with hydrodynamic  relationships occurring  through-
out the water column. The inflow zone is characterized
by riverine hydraulics and flexible water quality condi-
tions. The location and extent of the transition zone
varies with seasonal and hydrologic  conditions. High
flow years typically  expand and push the riverine and
transition zones  farther down the  reservoir  as  a
response to increased volumes and flow of water. Low
water years typically allow the transition zone to move
farther upstream into the main inflows. The impor-
tance of defining this transition zone is that it repre-
sents the area  where  reservoir processes become
dominant over riverine  processes, resulting  in initial
development of  stratification, definition  of density
currrents, and usually is the zone of extensive sedi-
ment deposition. The transition zone in Flaming Gorge
is seasonally characterized as the area where shifts in
algal  dynamics  occur  as  populations shift from a
diatom/green algae  dominance to a  green/blue-green
characterization.
  Secchi  depth   and chlorophyll  a  parameters as
definers  of  the  water  quality gradients in  Flaming
Gorge show strong  correlation with the reservoir pro-
ductivity. Secchi  depth levels have historically directly
measured eutrophication  levels and water clarity.
Chlorophyll  a measurements, while  not defining the
specific algal species, do give an estimate of the over-
all  productivity.   These two  factors  can  estimate
general water conditions and define  areas where
specific components of the  fishery may be segre-
gated. Seasonal  separation of the cold  and  warm
water components of the Flaming Gorge fishery can
be seasonally related to the productivity estimate es-
tablished through remote sensing.
SUMMARY AND CONCLUSIONS
We have reported here on what we feel has been the
successful application of multispectral remote sens-
ing to the characterization of water quality in a large
and hydrodynamically  complex Western reservoir.
Given the costs of conducting surface sampling of
such a water body, we believe the results presented
here justify the additional expenditure required to ac-
quire and process remote sensing imagery in order to
extrapolate point measurements over the entire reser-
voir. The contribution  to a  better understanding of
reservoir limnology is  significant.  We are confident
that the  techniques discussed here will work equally
well on other large Reclamation reservoirs.
   A long-range objective of the remote sensing work
being conducted in the Upper Colorado Region  is to
develop baseline water quality estimates for the reser-
voirs being managed by the USBR. Once baseline con-
ditions are established, seasonal and  annual trends or
shifts  in basic water quality relationships can  be
monitored. Remote  sensing analysis of the entire
reservoir surface allows for quantitative estimation of
water quality conditions on a large scale and at more
frequent intervals.
                                                 67

-------
LAKE AND RESERVOIR MANAGEMENT
REFERENCES

Ahern, F., et al. 1977. Use of clear lakes as standard reflectors
  for atmospheric measurements. Pages 731-755 In Proc. 11th
  Int. Symp. Remote Sensing of the Environment Environ. Res.
  Inst. Mich. Vol. 1.

Campbell, S. 1983. Phytoplankton and chlorophyll relationships
  in Flaming Gorge Reservoir—1982. Applied Sci.  Refer.  Mo.
  83-2-11. Div. Res. Bur. Reclam., Denver, Jan. 27.
Grimshaw, H., et al. 1980. Classification of Oklahoma reservoirs
  using  Landsat  multispectral  scanner  data. Okla. Wa:er
  Resour. Board Publ. 104. Oklahoma City.
Lillesand, T., et al. 1983. Use of Landsat data to predict the
  trophic state of Minnesota lakes. Photogramm. Eng. Remote
  Sens.  49(2): 291-329.
Mace, T. 1982. Characterization of lake water quality parameters
  with airborne multispectral scanner data: Flathead Lake,
  Mont. Pages 375S7 In  Proc. 1982 Annu. Con.  Am.  Soc.
  Photogramm., Denver.

Marton,  R.,  et al. 1983. Wisconsin's Lakes: A Trophic Assess-
  ment Using Landsat Digital Data. Inland Lake Renewal Sec-
  tion, Wis. Dep. Nat. Resour., Madison.
Meinert, D., D. Malone, A. Voss, and F. Scarpace. 1980. Trophic
   classification of Tennessee Valley area reservoirs derived
   from  Landsat  multispectral  scanner  data.  Tenn.  Valley
   Author.

Richardson, A., D. Escobar, H. Gausman, J. Everitt. 1980. Com-
   parison of Landsat  2 and field spectrometer reflectance
   signature  of  south  Texas rangeland  plant communities.
   Pages 88-97 in Proc. Symp Machine Processing of Remotely
   Sensed Data.

Turner, R., and M. Spencer. 1972. Atmospheric model for correc-
   tion of spacecraft data. Pages 895-934 in Proc. 8th Int. Symp.
   Remote Sensing of Environment. Environ. Res. Inst. Mich.
Verdin, J. 1983. Monitoring water quality conditions in a large
   western reservoir with Landsat imagery. Subm. to Photogram.
   Eng. Remote Sens.

Wegner, D. 1982. Limnological environment of Flaming Gorge
   Reservoir.  Proc. 1982 Western Div. Meet. Am. Fish. Soc., Las
   Vegas.

Witzig, A., and C. Whitehurst. 1981. Literature review of the cur-
   rent use and technology of MSS digital data for lake trophic
   classification. Pages 1-20 in Proc. 1981 Fall  Meet. Am. Soc.
   Photogramm., San Francisco.
                                                        68

-------
                                          Agricultural  Runoff
                                             and  Water  Quality
SPATIAL AND SEASONAL PATTERN  OF NUTRIENT AVAILABILITY IN
LA PLATA LAKE, PUERTO RICO
JORGE R. GARCIA
LAURENCE J. TILLY
University of Puerto Rico
Mayaguez, Puerto Rico



           ABSTRACT

           As part of a diagnostic restoration/feasibility study of La Plata Lake in Puerto Rico, we have examined
           seasonal, vertical, and (some) horizontal patterns of nutrient distribution. The eutrophic state of this
           reservoir is evidenced in a series of characteristics including epilimnetic (0-4 m) chlorophyll concen-
           trations averaging 15 mg m~3,  virtually total absence of dissolved oxygen  below 4m, and water
           hyacinth areal coverage estimated at 40 percent. Annual loadings of nitrogen and phosphorus have
           been calculated in 107 and 30 gm~2 yr~1 respectively, giving N:P loading ratios of 8.1. The ratio of
           N to P in the entire water column crop averaged 10.1 during the study period. The epilimnetic (0-4m)
           ratio averaged only 6 1. During a stratified dry period N:P ratios as low as 0.9:1 were observed whereas
           during periods following heavy runoff and lake mixing, the ratios approached values of 16:1. This pat-
           tern of inorganic nitrogen distribution suggests surface depletion from assimilation by hyacinths and
           algae The suggestions currently being examined are that nitrogen is limiting to primary productivity
           in the La Plata system and that the principal controller of lake function is the pattern of hydrological
           events Any plan to manage this and similar reservoir systems in Puerto Rico must take these factors
           into account
INTRODUCTION

Most lakes in Puerto Rico are manmade impound-
ments of major rivers. These systems were originally
constructed (since 1913) for hydroelectric power gen-
eration and agricultural irrigation, but now are primari-
ly water sources for domestic and industrial consump-
tion. As elsewhere, a major concern in reservoir man-
agement  in  Puerto  Rico is  eutrophication, which
sometimes  manifests itself  in  extensive stands of
water  hyacinth,  Eichhomia crassipes. It has been
shown that hyacinth productivity rates are significant-
ly affected by nutrient availability, both in natural envi-
ronments and artificial situations (Wooten and Dodd,
1976; Lugo et al. 1978; Wolverton  and  McDonald,
1979).  The present consideration of using hyacinth
harvest for nutrient removal  in  lake restoration  re-
quires the best possible understanding of the  nutrient
dynamics, especially of the temporal patterns of nu-
trient availability in the surface layers.
  Water reservoirs differ significantly from natural
lakes in structure and function. In general, reservoirs
are advectively dominated systems which usually pre-
sent pronounced horizontal gradients in water quality
(Thornton et al. 1980). The  impoundment  of river
waters usually leads to substantial modifications of
the physical, chemical,  and biological regimes in
association  with the increases in depth and surface
area  and the reduction of velocity (Markofsky and
Harlemann,  1971). Gloss et al. (1980) have shown that
the temporal fluctuations of nutrient availability in the
epilimnion of large temperate zone reserviors depend
upon advective delivery by tributary rivers. Tropical Af-
rican reservoirs  such as Lake Mcllwaine (Thornton
                                              69

-------
LAKE AND RESERVOIR MANAGEMENT
 and Nduku, 1982) have a potentially polymictic pattern
 of lake turnover given the proper conditions of surface
 cooling and river inflow in which nutrient enrichment
 of surface waters occurs. Polymixis in Lake Mcllwaine
 is, however, superimposed on a well-defined seasonal
 monomictic cycle, a mixing pattern similar to that re-
 ported for large tropical lakes by Lewis (1973,1983) for
 Lakes Lanao (Philippines) and Valencia (Venezuela),
 and also  related  to the availability of  higher nutrient
 concentration in the epilimnion of the lake. The speci-
 fic time  frame for  mixing  is determined by  local
 seasonal  variations of air temperature, wind velocity,
 and rainfall.
   Natural lakes generally have a smaller ratio of drain-
 age area  to lake  volume than reservoirs (Thornton et
 al.  1980)  and are thus  less  influenced by tributary
 rivers.  Nutrient  availability  in temperate lakes is
 primarily  related to convective and wind-induced mix-
 ing.
   Limnological data on  Puerto Rican  reservoirs are
 scarce. Phytoplankton populations have been evalua-
 ted by Candelas (1956), water hyacinth productivity by
 Nevarez and Villamil (1981), phytoplankton primary
 productivity by Gomez and Gonzalez (1978), Brown et
 al. (1979), and Martinez (1979), and general limnology
 of Lake Loiza by Quinonez-Marquez (1980). Of these,
 only the study of Lake Loiza provides some insights
 about temporal patterns of nutrient crops. The prin-
 cipal objectives of our study are  to describe the an-
 nual trajectory of nutrient distribution in one tropical
 reservoir  and to infer the main aspects of lake func-
 tion determining the nutrient dynamics of the system.
 Study Site

 La Plata Reservoir (Fig.  1) is  located in the  interior
 mountainous region of Puerto Rico (18° 20' N, 66° 13'
 W) at an elevation of 47 m above sea level. The res-
 ervoir has maximum extensions of 0.5 km width and
 9.6 km length, covering a surface area of 3.07 km2. The
 general configuration is long and narrow with a rela-
 tively  low  surface to volume  relationship; average
 depth is 10 m. The volume is approximately 3.08 x 107
 m3. Average theoretical replacement time is 28 days.


 METHODS

 The sampling approach consisted of a routine, month-
 ly  collection   of  water  samples   and  field
                                 PUERTO RICO
 measurements  at  three  watershed stations in  the
 main tributaries to La Plata Lake (La Plata River: W-1);
 Guadiana River: W-2; and Canas River: W-3) and at one
 station in the lake proper (L-l). Six additional  stations
 were also established in a gradient along the river axis
 of the lake.
   Water  samples were taken with Niskin bottles at
 depth intervals of 4 meters from the surface down to a
 depth of 20 meters. Analytical  procedures followed
 standard U.S. Environmental Protection Agency (1979)
 or equivalent methods. Water temperature,  pH, dis-
 solved oxygen, and conductivity were measured with a
 Hydrolab model 4041 multiprobe apparatus at depth
 intervals of 1 meter from the surface down to a depth
 of 20 meters. Round-the-clock measurements of water
 temperature and dissolved oxygen were performed
 during April 6-7 and October 5-7, 1982. A detailed ex-
 position of  methods, instrument precision, sampling
 and analytical sources of  error, and additional com-
 ments  are  presented  elsewhere  (Garcia and Tilly
 1983).                                          X'
 RESULTS

 Watershed Characteristics Related to
 Nutrient Loading by Major Tributaries

 The La Plata watershed (448 km2) is characterized by
 moderate to very steep slopes, with well-drained soils
 and rounded hilltops of strongly dissected uplands
 (Boccheciamp,  1978). The maximum  elevation of the
 basin is 980 meters (Pico, 1975). Approximately 90 per-
 cent of the total  drainage area  corresponds to La
 Plata River basin. The sub-basins  from Guadiana and
 Canas Rivers (former tributaries of La Plata River) ac-
 count for 6.2 and 3.5 percent of the total drainage area
                                                               10  11  12  1   23
                                                          20
                                                          12-
Figure 1.—La Plata Lake and sampling stations.
          10 11  12  1   2   3  4  5   6   7  8  9

                       MONTHS

Figure 2.—Monthly variation in loading of (A) Nitrogen and
(B) Phosphorus forms to the lake.
                                                  70

-------
                                                                  AGRICULTURAL RUNOFF AND WATER QUALITY
of the lake, respectively. A runoff coefficient of 50 per-
cent of the total precipitation has been calculated by
Giusti  and  Lopez  (1967)  from  streamflow
measurements in  relation to the total precipitation in
the basin.
  The single highest category of watershed land use
(61.5 percent)  is  agricultural, followed by forested
areas (25 percent) and residential (7.6 percent). Table 1
presents a crude estimate of the annual budget of po-
tential  nutrient  sources to the basin. Information on
the natural fertility of the soils is  not available. Ferti-
lizer application on pasture lands appears to be the
highest source of phosphorus and nitrogen in the ba-
sin representing approximately  49 and 42 percent of
the total estimated N and P crops, respectively.
  The annual phosphorus loading  to the lake via tribu-
tary rivers was  1.003 x 108 gP/lake/yr (30 g/m2/yr.) La
Plata River was the main loading  vehicle with 92 per-
cent of the total load. Total nitrogen loading  by tribu-
taries was also observed to be very high (5.12 x 108
g/lake/yr) equivalent to 167 g/m2/yr. Based on annual N
and P loading figures and the estimated potential nu-
trient stocks in the basin, a  leakage coefficient of 15
and 10 percent resulted for N and  P, respectively. Sig-
nificant positive correlations (p~.05) were found be-
tween monthly nutrient loading and total monthly pre-
cipitation  in the basin, both for phosphorus  (r = 0.64)
and nitrogen (r = 0.75).
  The monthly  distribution of phosphorus and nitro-
gen loading to the lake is presented  in Figure 2. Nutri-
ent inputs from tributary rivers were characterized by
relatively  high  concentrations of  soluble  reactive
phosphorus (SRP)  and  nitrite/nitrate  (NO3/NO2-N)
forms. Organic nitrogen was also substantial, repre-
senting 38 percent of the total  N  concentration. Am-
monia-nitrogen  was detected only in very low concen-
trations at tributary stations.
Temporal Pattern of Lake Stability and
Mechanisms of Mixing

Lake stability (Fig. 3) changed seasonally with higher
values during the summer (warmer) months and rela-
tively lower values during the winter (colder) months
(range 67-281 g-cm/cm2). A strong positive correlation
(r = 0.99) was found between surface water tempera-
tures and water column stabilities at La Plata Lake.
The mean temperature differential between zero and
20  meters  was 3.8 °C  (range  2.2-6.9 °C). Water
             temperature gradually declined with depth during the
             period  between  November 1981  and  April  1982;
             although a classical thermocline (i.e., A T1.0 °C/m) was
             not found during this period, a weak thermal gradient
             was maintained (AT = 0.16°C/m). A thermocline which
             fluctuated between the 3-8 meter  depth interval ap-
             peared between May and September 1982 (see Fig. 4).
             The maximum vertical gradient  injvater temperature
             occurred during June 1982 with A T = 2.0°C/m at the
             depth interval of 3-4 meters.
               Nearly complete mixing of the water  column was
             associated with discharges of relatively large and cold
             tributary flows into the reservoir  during  December
             1981.  During this period thermal stabilities were re-
             duced to the minimum recorded during the study (67
             g-cm/cm2). The mechanism for mixing appears to be
             turbulence  associated with  intrusion  of inflowing
             water.  In December intruded water reached  thermal
             and density equilibrium almost  at  the bottom of the
             water column (see Fig. 5) displacing the waters of less
             density toward the surface to be discharged over the
             dam. Inflows of warm, less dense  water float above
             the water  mass below the thermocline,  allowing for
             the maintenance of stratification.
                300
                250

             CM
              E
             0200
                                                     ca
                 100'
                 50
                        *12   1   23456789

                               MONTHS

             Figure 3.—Temporal pattern of lake stability.
                         Table 1.—Annual potential nutrient sources in La Plata basin.
Source
Human (P)
Swine (P)
Dairy Cattle (P)
Fertilizer (A)
Application
Number of Nitrogen Phosphorus
Units m. tons/yr. m. tons/yr. Coefficients
187,000 387
7,000 162
2,321,000 1,160
26,102 1,800
187 N: 3.0 Kg/per capita/yr.
P: 1.0 Kg/per capita/yr.
56 N: 23 Kg/an imal/yr.
P: 25 Kg/aminal/yr.
464 N: 0.5 Kg/animal/yr.
P: 0.2 Kg/animal/yr
600 N: P: K = 15:5:10
0.45 m. tons/
Reference
Quinonez-
Marquez,
1980
Uttormark
et al.
1974
Uttormark
et al.
1974
Local
Farmers
                                                                            s. acre/
                                                                           yr.
 Totals
 P = Total Population
 A = Total Cultivated Area
3,703
1,435
                                                  71

-------
LAKE AND RESERVOIR MANAGEMENT
   We looked for evidence of convective or wind in-
 duced vertical displacements (atelomixis) as has been
 demonstrated by Lewis (1983) for large tropical lakes.
 Density instabilities or isothermal profiles were not
 evident from our monthly sampling. Finer time scale
 observations of water  column temperature  over
 periods of 48 hours also failed to demonstrate any
 clear indication of atelomixis. The clearest indication
 of  mixing during the entire year was the disap-
 pearance of the typical  clinograde  02  profile  in
 December. At this time, higher 02 concentrations were
 found below lower ones reflecting the intrusion of
 well-oxygenated tributary water into previously anoxic
 water (see Fig. 6). This higher oxygen water was too
 deep (below the compensation point) to have  been
 photosynthetically induced.
 Horizontal Gradients Along the River
 Axis of the Lake

 Figure 7 presents the horizontal distribution of a set of
 limnological  parameters sampled in  a grid fashion
 along the river axis of La Plata Lake. Important allooh-
 thonous inputs were dissolved  oxygen, total  phos-
 phorus, soluble reactive phosphorus,  and nitrite/ni-
 trate concentrations which decreased gradually from
 the tributary station toward the dam site. Secchi disk
 transparency, macrophyte densities, and ammonia.-N
 concentrations increased in the  same direction, indi-
 cating their endogenous character.
   The peak in chlorophyll a near the middle of the lake
 can  be interpreted as  an optimization response of
 phytoplankton  to high  nutrients and reduced  in-
 organic turbidity. Further down the lake nutrients are
 consumed and biological shading probably becomes
 important. Water hyacinth growths occurred as nar-
 row  bands at both sides of the river axis, increasing to
 extensive mats approximately 1.5 km  from the dam
 site  in the  direction  of  the river  flow, covering
 approximately  50 to 60 percent  of the space in this
 section of the lake. Hyacinths tend to be concentrated
                                   in this location as a combined result of the water flow
                                   and prevailing wind direction.
                                   Temporal Patterns of Vertical Nutrient
                                   Distribution in the Lake

                                   Nutrient profiles were of two different forms related to
                                   marked variations in nutrient inputs from tributary
                                   rivers, hydrodynamic mixing events, and stratification.
                                   Thus, we will refer to a period  when  the lake was
                                   physically driven (November-January)  characterized
                                   by  relatively  high precipitation  in  the basin, large
                                   nutrient  loading, and  low thermal stability  and to
                                   another period when the  lake was biologically con-
                                   trolled (May- September)  characterized by relatively
                                   higher thermal  stability, lower triburary loading,  and
                                   persistent stratification.
                                     Dissolved oxidized forms of phosphorus and nitro-
                                   gen such as SRP and NO;i/NO2-N were in substantially
                                   higher concentrations during the period of physically
                                   driven conditions and  reflected a gradient of higher

                                                      MONTHS
                                   Figure 5.—Monthly variation  in lake volume replacements
                                   and depth of thermal equilibrium for La  Plata River dis-
                                   charges into the lake.
         OCT
NOV

1981
DEC   JAN     FEE    MAR    APR

                  MONTHS
                                                                MAY    JUN     JUL    AUG   SEPT

                                                                       1982
Figure 4.—Isolines of water temperature versus depth at s.tation L-l. Contour interval: 0.2°C.
                                                  72

-------
                                                                 AGRICULTURAL RUNOFF AND WATER QUALITY
concentrations  toward the  bottom  (Fig. 8).  The
average volume-weighed crops during this period were
7.06 and 1.3 g/m2 for NOa/NO^N and SRP, respective-
ly.  During the more lentic stratified conditions the
volume-weighed crops for NO3/NO2-N and SRP were of
0.46 and 1.0 g/m2, evidencing  profiles of a  strong
dichotomic character  with maximum concentrations
generally found at 8 meters (Fig. 9). This type of distri-
bution has been described for highly productive lakes
(Hutchinson,  1957) where clinograde  oxygen struc-
tures  prevail in the water column.
  Low concentrations result in the trophogenic layer
from biological assimilation by plants  and in the tro-
pholytic layer from biochemical reduction  of oxidized
forms, for example, nitrate to ammonia. As a result, an
inverse relationship (r=-.56,  n = 13, sig .05)  was
found  between monthly crops of ammonia-nitrogen
and nitrite/nitrate during the year. The overall ammo-
nia-N   concentrations  increased  (vol.-weighed  ave.
crop = 12.9 g/m2) during the period of biological con-
trol and decreased to a minimum vol.-weighed crop of
2.67 g/m2, during the period of physical control.
  Representative profiles of average concentrations
of N forms evidenced during physically driven and bio-
logically controlled conditions are presented in Figure
10.  The relative proportions of nitrogen forms reflect
the presence of  high composition of ammonia-N be-
low the oxygen chemocline during the stratified per-
                      D.O.  (mg/l)
iod. At the same time, above the chemocline organic N
was the most  abundant species, probably reflecting
the influence of dense algal or macrophyte  crops in
that layer. In the more physically driven scenario, large
tributary inputs such as NO2/NO3 and  organic N  are
the dominant forms, while ammonia was only found in
trace concentrations.
  Total phosphorus crops peaked during the period
between  November  and  January  as a result   of
tributary loading while SRP crops were less variable
throughout the study (Fig. 11).
Annual Budgets of Phosphorus and Nitrogen

The  major pathways of phosphorus and  nitrogen
crops are  outlined in Figure 12. As previously noted,
tributary  loading  represented  the largest  input of
nutrients to the system with calculated loads of 167
and 30 g/m2/yr for N and P, respectively. Spillage over
the dam appears to be the major loss of nutrients from
the lake.  This loss  is related to the large pulses of
tributary loading,  the occasional mixing of the water
column, and short average residence times of incom-
ing waters. Organic sedimentation appears high in
relation to the nutrient income. The annual turnover of
water hyacinth biomass requires  incorporation of N
and P of the order of 102 and 39 g/m2/yr, respectively.
This estimate  is based  on  hyacinth  productivity
measurements  of 9.68 g  DW/m2/day reported  by
Nevarez and Villamil (1981) for nearby Loiza Lake and
nutrient content  determinations of plants  sampled
from La Plata which were respectively .011 ±.002 gP
and  .029 ±0.14 gN per gram  dry weight of  plant.
Assuming hyacinths are not nutrient-limited and pro-
ductivities maintained at 9.68 g  DW/m2/day an internal
nutrient loading rate of 48 and 33 g/m2/yr for nitrogen
and  phosphorus was  calculated  assuming  no
changes in the standing crop of hyacinths.
                                                     WATER

                                                     HYACINTHB-
                                                             L A PLATA
                                                             RIVER 	
Figure 6.—Vertical distribution of dissolved oxygen during
December 1981 at La Plata Lake (Station L-l).
Figure 7.—Horizontal gradients of water quality along the
river axis of La Plata Lake.
                                                 73

-------
 LAKE AND RESERVOIR MANAGEMENT
 Nutrient Ratios in the Water Column

 The occurrence of an N:P loading  ratio of 10:1 has
 been used to suggest nitrogen limitation in temperate
 lakes (Forsberg, 1980). The annual loading at La Plata
 averaged 12:1. In-lake crop ratios were 10:1 and 11:1
 for the epilimnion  and hypolimnion,  respectively. A
 large portion of the nitrogen in the  epilimnion of the
 lake was found to be organic. Available nitrogen (am-
 monia   or  nitrite/nitrate)  was  depleted   in  the
 trophogenic zone relative to available P (SRP)  with
 average N:P of 6:1. During  stratified  periods of low
 rainfall, monthly standing crop ratios of dissolved N:P
 as low as 0.9:1 were observed in the epilimnion of the
 lake. The highest standing  crop ratios of available
 nitrogen to phosphorus in the epilimnion of the lake
 (16:1) resulted during and after periods of high rainfall
 and watershed runoff (November 1981-February 1982;
 see Fig. 13).
 SUMMARY AND DISCUSSION

 From the data reviewed, a generalized description of
 the La Plata Lake's trajectory can be attempted. Annu-
 ally, La Plata Lake receives large amounts of phos-
 phorus and nitrogen from the rain-runoff regime of n
              SRP   (mi/I)

               .05   .10    .15
.20
      20-
                 N02/N03  (mg/l)

             .2    .4     .6     .8
 M
                    basin in which substantial domestic and agricultural
                    developments exist. The influence of La Plata River,
                    which accounts for more than 90 percent of the water
                    input to the lake, is evident in a series  of horizontal
                    gradients related to the gradual sedimentation of par-
                    ticulate materials and  biological assimilation  and
                                  05
                                                                                     0   7 /  82

                                                                                     *  6 /  82
                    Figure 9.—Representative profiles of SRP and N/N concen-
                    trations during periods of biologically controlled conditions
                    in the lake.
                                                           12-

                                                           10-

                                                           8-
Figure 8.—Representative profiles Of SRP and N/N concen-
trations during periods of physically driven conditions in the
lake.
                              10  II  II   I  2   3   4   5   6   7   8   9
                               1981       MONTHS        1982


                    Figure 10.—Profiles of average concentrations of N forms
                    during periods of (A) physical and (B) biological control.
                                                  74

-------
                                                                   AGRICULTURAL RUNOFF AND WATER QUALITY
transformation of nutrients along the river axis of the
lake. A progressive increase in depth and volume re-
duces the velocity of incoming waters to a point in
which a significant degree of physical, chemical, and
biological  stratification  occurs.  The  stratification
leads to a pronounced clinograde O2 distribution and
nutrient depletion  in the trophogenic layers resulting
largely from assimilation by floating macrophytes and
algae. Of the dissolved "available" nutrients, nitrogen
forms appear to be limiting to primary production in
the superficial layers.
   Mixing  of the water  column  appears related to
severe rainstorm events when incoming waters of low
temperature attain  density equilibrium  with  deep
layers of the lake and replace "old" lake water with
"new"  tributary waters of lower transparency, higher
dissolved oxygen (DO), and higher available nutrient
concentrations. With the reduction of rainfall, strati-
fied conditions rapidly reestablish the clinograde pro-
file and below the chemocline nitrite/nitrate concen-
trations are transformed (reduced) to ammonia-nitro-
gen and accumulated, while the dissolved nutrient
stock above the chemocline is progressively metabo-
lized by macrophytes  and algae. Similar patterns of
nutrient stress during prolonged periods of stratifica-
tion have  been reported by  Lewis (1983)  for  Lake
Valencia  as a  result of nutrient depletion in the
euphotic zone.
   On the basis of the limited data available, Tilly and
Garcia (1983, this volume)  provisionally generalized
that  for   Puerto  Rican  reservoirs  high   nutrient
            availability is related to mixing events in which large
            tributary loading also occurs. The study of Lake Loiza
            (Quinonez-Marquez, 1980) and the present study have
            shown that  very weak thermal  gradients and even
            isothermal profiles occur in the water column during
            the period between October and February  making
            these systems more susceptible to mixing during this
            time.  Rainfall associated with cold fronts, which  are
            the dominant weather  systems of the winter season
            (October-February), would  be expected to enter  the
            lake as relatively cold, dense water masses capable of
            plunging deeper into the water column and mixing the
            lake. More observations are needed to determine the
            role of other recurrent climatological conditions such
            as tropical depressions  in lake  mixing and  conse-
            quent epilimnetic nutrient availability.
            REFERENCES

            Boccheciamp, R.A. 1978. Soil survey of San Juan of Puerto
              Rico. Soil Conserv. Surv. U.S. Dep. Agric.
            Brown, R.A., et al.  1979. Preliminary results from a survey of
              water quality in  some Puerto Rican lakes. Center Energy
              Environ.  Res. Human Ecol. Div. Univ. Puerto  Rico, San
              Juan.
            Calvesbert, J.R. 1961 Climate of Puerto Rico and U.S. Virgin
              Islands. U.S. Gov. Print. Off. Washington, D.C.
            Candelas, G.A. 1956. Studies on the freshwater plankton of
              Puerto Rico. Ph.D. Thesis. Univ. Minnesota.
                                      N  (mg/l)
                                     N  (mg/l)
                           0   .2
-6   .8   1.0   1.2    1.4
Figure 11.—Monthly composition of SRP into the TP concentrations during the year at station L-l in La Plata Lake.
                                                   75

-------
 LAKE AND RESERVOIR MANAGEMENT
                      PHOSPHORUS    BUDGET
                          G/MVYR
                       NITROGEN   BUDGET
Figure 12.—Phosphorus and nitrogen budgets for La Plata
Lake.
          10  11  ie
                         e  3   
-------
A SIMULATION MODEL FOR ASSESSING THE SUCCESS OF
AGRICULTURAL BEST MANAGEMENT PRACTICES ON
SURFACE WATER QUALITY
JAMES MADIGAN
Bureau of Water Research
New York State Department of Environmental Conservation
Albany, New York

DOUGLAS  HAITH
Cornell University
Ithaca, New York

SCOTT 0. QUINN
JAY BLOOM FIELD
Bureau of Water Research
New York State Department of Environmental Conservation
Albany, New York


            ABSTRACT

            A simulation model was used to assess how agricultural practices used to reduce soil erosion, such
            as no-till cropping, affect the plant nutrient and suspended sediment levels in streams. The model
            is based on principles developed by the Soil Conservation Service, such as the Universal Soil Loss
            Equation and the Hydrologic Curve Number Equation. Annual model predictions for discharge and
            the mass loadings of phosphorus, nitrogen, and suspended sediment compare favorably to actual
            data collected for Irondequoit Creek, in western New York State during 1980 and 1981. The model
            was then run, driven by 20 years of meterological data, in order to assess various agricultural Best
            Management Practices. The effect of selected practices on discharge, nutrient loadings, and sedi-
            ment loadings is presented.
 INTRODUCTION

 Irondequoit Bay, located 6 km northeast of Rochester,
 N.Y., is separated from Lake Ontario  by a 90-meter-
 wide sand bar and has the general characteristics of a
 freshwater lake of similar dimension (6.7 km x 1 km).
 A dense algal crop occupies its surface water continu-
 ously from  May  to mid-October  and  its  deep sedi-
 ments  have been characterized  as "black muck."
 Decomposition of organic matter during  periods of
 winter stratification depletes dissolved oxygen in the
 bottom water and generates high concentrations of
 hydrogen sulfide, ammonia, and phosphate. Water
 quality data indicate that the bay is not nutrient
 saturated, and is probably phosphorus limited (N.Y.
 State Dep. Environ. Conserv. 1982).
  The phosphorus control strategy, which is expected
 to restore primary productivity to that of a borderline
 eutrophic-mesotrophic  status,  consists  of four
 phases:

  I. Statewide ban on detergent phosphates ....  ef-
 fected 1973,
  II. Interception of STP flows .... $130 million spent,
  III. Interception of combined sewer overflows.... $80
 million spent, and
  IV. Reduction of nonpoint sources (NPS).... present-
 ly being addressed.

  The Irondequoit Bay National Urban Runoff Program
addressed modeling of urban nonpoint sources, as well
as the monitoring of water quality  throughout the
watershed.  This  report  is  a  synopsis  of  the
methodologies used to assess nonpoint source pollu-
tion from the agricultural portion of the  watershed,
specifically the 11,508 ha Thornell Road subwatershed
—the study area.
  The study area is small and well defined, allowing
for reliable and accurate pollutant runoff determina-
tions.  Major land use  categories are agricultural,
forested, and developed, representing 49 percent, 32
percent, and 19 percent of the study area, respectively.
Row crops represent 37 percent  of the agricultural
land and 18 percent of the total study area. Half the
land in row crops has highly erosive soil types. Taking
into account the erosive potential  associated with
conventional row crop tillage practices, these areas
warranted  close  scrutiny as  significant nonpoint
sources of  nutrient runoff. After extensive ground-
truth work, it was determined sediment export would
be predicted from the agricultural  portion only.

Modeling Approach

A simulation model (Haith and Tubbs, 1979) was used
to assess agricultural  nonpoint  contributions  of
suspended sediment, runoff, and both soluble and
particulate phase P and N. The model is based on prin-
ciples developed by the U.S. Soil Conservation Serv-
ice, namely the Universal Soil  Loss Equation (USLE)
                                              77

-------
 LAKE AND RESERVOIR MANAGEMENT
 and the Curve Number Equation (CNE). Perhaps it is
 easiest to conceptualize this as a bookkeeping opera-
 tion in which 500 1  ha cells were randomly selected
 and assigned tabulated curve numbers and soil loss
 parameters.  Both  soluble and solid-phase  com-
 ponents are computed and summed on a daily basis—
 thus the "bookkeeping" perspective. Note Figure 2 for
 a general description of this process. Contributions
 from each unit source area are based on the following
 equations and assumptions:
 Runoff
(Rkt + Mw - 0.2Skt)2

(Rkt +  MM +  0.8Skt)
                                             (1)
 Kt  is the  unit  source  area per  time  interval.
 Rkt and Mkt = the rainfall and snowmelt (cm) on clay t
 and source area k; and Skt =  a detention parameter
 (cm) which is a tabulated function of soil hydrologic
 group, crop, management,  hydrologic condition, and
 antecedent soil moisture. The detention parameter,
 Skt, is determined from a curve number  CNkt, accor-
 dingly:
           2540
    Skt =	25.4
           CN
(2)
              kt
The onset of direct runoff does not begin until Rt - Kt
>.2Skt. Otherwise, Qkt = 0 and the rain and snowmelt
values are used for baseflow and antecedent moisture
sums only.  During winter, low evaporation rates and
      Lake Ontario
                            Irondequoit  Bay
                                           N

                                   0            I.5 km

                                  contour interval =IOft
      Figure 1.—
                                              TRANSPORT  AND

                                                ATTENUATION
                                                                LOADINGS FROM  OTHER

                                                                UNIT SOURCE AREAS
 UNIT  SOURCE
     AREA
Crop , soil ,
management ,

topography,

weather
                                     EDGE-OF-FIELD

                                   POLLUTANT  LOSSES
             DELIVERY TO

              WATERSHED

                OUTLET
                                            dissolved
                                            pollutants
                                                                          WATERSHED

                                                                            EXPORT
                                                                                        solid- phase
                                                                                        pollutants
                                                                 LOADINGS FROM OTHER

                                                                  UNIT  SOURCE  AREAS
Figure 2.—Illustration of General Methodology.
                                                78

-------
frozen soil produce high runoff conditions, correlating
to the wettest antecedent moisture condition used for
the CNE.  When daily temperature records indicate
subfreezing  conditions,  precipitation is considered
snow and accumulates until melt occurs, according to
a degree-day equation (Woolhiser, 1976).

Sediment

The USLE was designed for use in estimating annual
averages for cropland soil loss. To determine erosion
from individual rainstorms, it was modified according-
ly:
      kt = 0.013(Etl,30)Kk(LS)kCkPk
(3)
Parameters, Kk,  (LS)k, Ck, Pk =  standard tabulated
values for soil erodibility, topographic, and cover prac-
tice  factors, respectively (Wischmeier  and Smith,
1978). The remaining portion of the equation, 0.013 (Et
lt30), is the rainfall erosivity factor for the rainstorm on
day t, in which lt30 = maximum 30-minute storm inten-
sity (cm/hr); and Et = storm kinetic energy (joules/m2).
Kinetic energy associated with various  rainfall  pat-
terns  can be calculated from a regression equation
(Wischmeier, 1976).
  To better  predict  hydrologic variability, the follow-
ing refinements  supplemented the basic USLE  and
CNE computations:
  (A) Evapotranspiration component
  (B) Groundwater discharge component
  (C) Determination of hydrologic curve number as a
continuous  linear  function  of 5-day  antecedent
moisture condition
  Although these equations are not presented within
this report, the equations and computer algorithms
are readily available from the authors.

MODEL INPUTS
Meteorological Data

The CNE and USLE  equations require two distinctive
sets of meteorological inputs. The CNE requires daily
             AGRICULTURAL RUNOFF AND WATER QUALITY

 temperatures, daily precipitation values, 5-day antece-
 dent moisture totals, and 30-day running averages for
 temperature. Requirements of the USLE equation are
 somewhat more complicated,  with hourly rainfall
 values required  for the  mathematical description  of
 various rainfall patterns—the storm  erosivity factors.


Nutrient Concentrations

Direct runoff and soil loss values obtained from the
CNE and USLE are multiplied by nutrient  concentra-
tions to obtain loadings  of P and N.  Soluble nutrient
concentrations represent literature values  for various
land use/soil combinations (Dornbush et al. 1974). Par-
ticulate phase concentrations of N and P were derived
after analyzing soil samples for total Kjeldahl nitrogen
(TKN), percent volatile solids, and total phosphorus
via an acid extraction (HCI, CdB, NaOH).


Cell Specific Information

Table 1 specifies the information needed from each 1
ha cell in order to develop factors needed to satisfy
the  USLE and CNE equations.
     Simulation of Best Management
     Practices (BMP)

     Five distinctive BMP strategies were evaluated: con-
     touring, strip cropping, reduced (conservation) tillage,
     no-till, and sod-based rotations. Brief definitions of
     each of these BMP's appear in Appendix 1.  Each of
     the five BMP's is simulated by changing soil loss and
     runoff parameters for the row crop areas of the water-
     shed. The process involves identifying the 1-hectare
     cells that contain row crops and adjustment of cover
     (c), the supporting practice (sp), and the curve number
     (en), as appropriate. The simulation model is then run
     for a 20-year period, merging the new parameters with
Table 1.— Cell-specific information.
Land Use Categories:
Factors
Soil mapping unit
Soil hydrologic group
* Soil erodibility factor (K)
Soil sample
* Slope length/gradient (LS)
Crop type
* Cover factor (c)
Crop rotation information
* Practice factor (CP)
Conservation practice info.
* Runoff curve number (AMC II)
* Detention coefficient (AMC II)
* Downslope distance to nearest
identifiable drainage channel (DK)
Litter depth
Humus type (mull/mor)
Forest type (soft vs. hard)
Tree size (DBH)
% exposed mineral soil
Type management
% impervious surface
* Standard Soil Conservation Service tabulated values
Agricultural

X
X
X
X
X
X
X
X
X
X
X
X

X








Forest

X
X
X
X

X




X
X

X
X
X
X
X
X
X


Residential

X
X
X
X

X




X
X

X




X

X

                                                 79

-------
 LAKE AND RESERVOIR MANAGEMENT
 a  20-year  meteorological  data  set.  Appendix  2
 describes the parameter changes for each BMP.
 Results

 Figure  3 illustrates the most  important  findings—
 those reductions in nutrient loadings attributable to
 the 20-year  simulation of various BMP  strategies;.
 Study period land use conditions were merged with
 the 20-year  meteorological  record,  serving  as  a
 baseline  for comparison.  During execution at the
 model,  it was possible to summarize nutrient loadings
 from agricultural areas independently from the entire
 study area. Likewise, it was possible to calculate per-
 cent reductions according to the same scheme.
 CONCLUSIONS

 Monthly  stream  discharge,  sediment  yield,  and
 nutrient fluxes in streamflow  from  large  hetero-
 geneous watersheds can be predicted by a relative!/
 simple simulation model. The model requires dail/
 temperature and precipitation data and does not re-
 quire calibration. However, stream discharge data are
 needed for estimation of a ground water recession
 constant. Watershed soils and cover data are deter-
 mined from land use, topographic, and soil maps.
  The general  watershed model was tested  on tho
 11,508 ha  Thornell Road subwatershed. The model
 was run for the period Aug. 1, 1980, to July 31, 1981.
 Monthly predictions were compared with measure-
 ments from flow guaging and water quality sampling
 for all  months during  which  measurements were
 made.
  The model accurately predicted  total  nutrient  and
 sediment  loads despite problems  with some of tho
 monthly streamflow predictions. A significant error
                      Soluble P
2  -40-
    20-


     0

   -20-


   '-40-
                Ki_i
                      Total N
                                    I   I Total watershed
                                    I	I  (Ag. included)
                                       j Agricultural pcrtion
                                       I  of watershed
                      Total P
                      Sediment Export

       K  h  I!   I
Figure 3.—Percent reductions from 20 year simulations ol
agricultural best management strategies.
was detected  in several of the high-flow  months.
These errors resulted from the precipitation data from
Thornell Road used in the simulation runs, which ap-
parently are not representative of watershed precipita-
tion.  Mendon  Ponds  precipitation data were deter-
mined to be more representative.
   Long-term estimates of sediment export, and phos-
phorus  and nitrogen annual  loadings were developed
based  on  20-year  simulations.  Agricultural  direct
runoff contributed 89 percent of the suspended sedi-
ment annual  load and 76 percent of the total phos-
phorus  annual  load. Adjustments were made to the
model to evaluate  the significance of agricultural
BMP's.  BMP's were applied only to row crop  portions
of the study area.  Twenty-year  simulations of agri-
cultural BMP's, identified no-till and strip cropping as
most effective  at reducing suspended sediment and
total  phosphorus loading. No-till reduced total sub-
watershed  annual loading of suspended sediment by
43 percent  and total phosphorus by 29 percent. Strip
cropping reduced total subwatershed annual loading
of suspended sediment by 40 percent and total phos-
phorus by 31 percent.
   Until   the  predicted  loadings  and  mitigative
strategies have been put into full perspective, it is dif-
ficult to assay the immediate benefits to water quality.
Throughout the modeling effort, valuable hydrologic
information was compiled, leading  to  a  greatly
enhanced understanding  of the agricultural nonpoint
source pollution impact on Irondequoit Bay.
 APPENDIX 1
 Definition of BMP's

 Best Management Practices Presented
 in This Study

 Five distinctive BMP strategies were  evaluated for
 purposes of this study. Four of the  practices—con-
 touring, strip cropping, reduced (conservation) tillage,
 and no-till—are cost-shared by Agricultural Stabiliza-
 tion and Conservation Service in both Monroe and On-
 tario Counties. The fifth BMP, sod-based rotation, is
 both  a soil and  water  conservation  and  nutrient
 management practice; however, it is not cost-shared.
 Descriptions of these practices follow.
  Contouring.  In  contouring,  tillage  and planting
 operations follow  cross-slope field contours. Down-
 slope water movement is slowed, thus reducing soil
 detachment and  increasing  infiltration.  On  certain
 topographies,  this practice is difficult  to implement
 because of  the hazards associated with large multi-
 row equipment. Keep in mind, contouring may ag-
 gravate wetness problems on poorly drained soils.
  Strip cropping. This practice alternates sod and row
 crop strips which are planted cross-slope. Strip cropp-
 ing combines the benefits of contouring with the soil
 cover and enhanced infiltration of  hay or a close-
 seeded legume (alfalfa or clover). The practice implies
 replacing a continuous row crop with a 50-50 sod and
 row crop rotation.  Neglecting to consider long-range
 benefits associated with the preservation of valuable
top soil, one can note a short-term decrease in profits.
  Reduced or conservation  tillage.  Reduced  tillage
eliminates routine soil inversion by moldboard plow-
ing. Rather,  chisel or disk plowing is used to  loosen
the soil. Much  of the crop residue remains on the soil
surface to provide erosion protection.
                                                 80

-------
                                                                     AGRICULTURAL RUNOFF AND WATER QUALITY
   No-till. This is an extension of reduced tillage which
eliminates all soil disturbance by tillage. Weeds are
controlled chemically and fertilizer is left on the soil
surface  to be  leached into the soil.  No-till  will
dramatically reduce erosion over conventional tillage.
However, since fertilizer phosphorus remains close to
the  soil  surface,  dissolved  phosphorus  losses in
runoff will increase. A summary of field studies by
Logan and Adams (1981)  indicates  that dissolved
phosphorus concentrations in runoff from no-till sites
will be at least double the conventional tillage situa-
tion. Since no-till may impede the drying and warming
of soil in spring, it  is most  suitable for  well-drained
soils (soil hydrologic classes A and B).
   Sod-based  rotations.  In this practice,  continuous
row  crops such as corn are replaced with a rotation
that is at least 50 percent sod (hay, clover, or alfalfa).
Improved filtration and reduced erosion result; but, as
in the case of strip cropping, an economic impact can
be attributed to the reduction in  row crop area.
APPENDIX 2


Adjustment of Model Parameters for the
Simulation of BMP's.
     BMP         Model Parameter Adjustments

Contouring    Contouring affects both the supporting
              practice factor and curve number for row
              crops. The former is changed to 0.60,0.50,
              0.60,0.80, and 0.90 for slopes of 1.1-2 per-
              cent, 2.1-7 percent, 7.1-12 percent, 12.1-
              18 percent, and 18.1-24 percent, respec-
              tively  (Stewart et  al.  1975). The curve
              number is changed from the straight row
              to the  contoured value.

Strip          Row crop SP factors are 0.30, 0.25, 0.30,
cropping      0.40, and 0.45 for  the same  respective
              slope classes given for contouring (Stew-
              art et al. 1975). Curve number adjustments
              must reflect the 50-50 row/  sod  distribu-
              tion and the improved soil infiltration.
              Referencing  the  cells' soil hydrologic
              group, the appropriate curve number is
              computed as the average of curve num-
              bers  for row crop  and   close-seeded
              legume.  In  choosing the  strip cropping
              curve number, contour plowing and good
              soil hydrologic condition is assumed.

Reduced      Effects of reduced tillage on row crops
tillage         are seen in the cover factor, which is ad-
              justed  to the average C for the reduced til-
              lage land use cells in Western New York
              State (C = 0.11) (Perritt, pers. comm.)
 No-till        Three changes are required to simulate
              the effect of changing all row crop cells to
              the no-till practice.  First, curve numbers
              are changed to reflect "good hydrologic
              condition."  Second,  dissolved  phos-
              phorus, to be multiplied by the new runoff
              volumes, ;'s Increased from 0.26 to 0.52
              img/l. Also, the C factor is adjusted to the
              average  C for  no-till land  use  cells  in
              Western New York State (C = 0.05) (Perritt,
              pers. comm.)

Sod-based    The average C factor for sod-based rota-
rotations      tions with corn is 0.12 for the watershed.
              Curve number adjustments  are similar to
              strip cropping with  good soil hydrologic
              condition assumed. Taking  into account
              the cells' hydrologic soil group, the curve
              number is computed as the average of
              curve numbers for (1) row crops planted in
              straight row and (2) close-seeded legume.
REFERENCES

Dornbush, J.N., J.R. Anderson, and LL Harms. 1974. Quanti-
  fication   of  Pollutants  in  Agricultural  Runoff.
  EPA-660/2-74-005. U.S. Environ. Prot. Agency, Washington,
  D.C.

Haith, D.A., and LJ. Tubbs. 1979. Modeling nutrient export in
  rainfall and snowmelt runoff. Pages 665-85 in R.C. Loehr,
  D.A. Haith,  M.F.  Walter,  and C.S. Martin, eds. Best
  Management Practices for Agriculture and Silviculture.
  Ann Arbor Science Publishers. Ann Arbor, Mich.

New York State Department of Environmental Conservation.
  1982. Work Plan  Irondequoit Basin Agricultural Non-Point
  Source Study. Bur. Water Res., Albany.

Perritt, R. Pers. comm. Soil Conserv. Serv. Syracuse,  N.Y.

Stewart, B.A., et al. 1975. Control of Water Pollution from
  Cropland.  Vol.  I. EPA-600/3-75-026a, U.S.  Environ. Prot.
  Agency. Washington, D.C.

Wischmeier,  W.H.  1976. Cropland erosion and  sedimenta-
  tion. Pages 31-57  in Control of Water  Pollution from
  Cropland.  Vol. II. EPA-600/2-75-026b. U.S.  Environ. Prot.
  Agency. Washington, D.C.

Wischmeier, W.H.,  and D.D. Smith. 1978. Predicting Rainfall
  Erosion Losses  — A Guide  to  Conservation Planning.
  Agric.  Handbook  No.   437.   U.S.  Gov.   Print.  Office,
  Washington, D.C.

Woolhiser, D.A. 1976. Simulation of daily potential direct run-
  off.  Pages  123-48  in Control of Water  Pollution from
  Cropland. Vol. II. EPA-600/2-75-026b. U.S.  Environ. Prot.
  Agency, Washington, D.C.
                                                    81

-------
 THE  EFFECTIVENESS OF BMP'S AND  SEDIMENT
 CONTROL STRUCTURES  AND THEIR
 RELATIONSHIP TO  IN-LAKE WATER QUALITY
FORREST E. PAYNE
TIMOTHY M. BJORK
South  Dakota Department of Water and Natural  Resources
Pierre,  South Dakota
            ABSTRACT

            Lake Herman is a nitrogen limited, hypereutrophic, shallow (mean depth 1,7m), warm water lake located
            in Lake County, S D. The lake has experienced extensive blue-green algal blooms, fish winterkills,
            and receives a high sediment load. Algal and macrophyte growth has diminished the open water sur-
            face area. The predominant land  use is agriculture with permanent homes, small businesses, and
            public recreation areas surrounding the lake's shoreline. On January 13, 1978, Lake Herman and
            its associated watershed were selected to participate in the U.S. Department of Agriculture and U.S.
            Environmental Protection Agency  sponsored Model Implementation Program. The primary objective
            of the project was to improve the  water quality of Lake Herman by reducing phosphorus, nitrogen,
            and sediment loads through voluntary application of Best Management Practices (BMP's) and con-
            struction of sediment control structures. Approximately 5 years of water quality data from the lake,
            the tributaries, the outlet, and above and below the BMP's, and the sediment control structures are
            available from this project. Although the results of t-test analyses indicated that the sediment control
            structures are reducing the sediment and nut lent load, a corresponding reduction has not been observ-
            ed in the lake. Therefore, two phosphorus mass budget models were used to predict the phosphorus
            concentrations in the lake: neither predicted the high phosphorus concentrations observed in the lake.
            It is assumed that another nonpoint sourco of phosphorus affects the lake (internal loading from
            resuspended sediments or aquatic macrophytes). The next step in the Lake Herman restoration pro-
            ject is to dredge selected areas to deepen them and reduce the resuspension of sediments.
INTRODUCTION

Lake Herman is  a nitrogen limited, hypereutrophic,
shallow, warmwater lake located in Lake County, S.D.
(Churchill and Brashier,  1972; S. Dak. Dep. Water Nat.
Resour. 1980; Rast and Lee, 1979). Table 1 summarizes
the  morphometric,  hydrologic,   and  phosphorus
budget parameters for Lake Herman. The lake has ex-
perienced  blue-green  algae blooms, winter  fishkills,
and algae  and aquatic macrophytes have diminished
the open water surface area. The  predominant la.nd
use (75 percent) in the watershed is corn and small
grain farming in support of livestock operations. The
remaining  25 percent  is  comprised  of haylands,
pasture, or wetlands. Approximately 60 percent of 1 he
shoreline is in private ownership  and 40 percent is
public or semipublic.
  On Jan.  13, 1978, Lake Herman  and its associaled
watershed were selected  to participate in  the U.S.
Department of Agriculture's (USDA)/U.S. Environmen-
tal Protection Agency's (EPA) Model Implementat on
Program (MIP). The MIP was designed to demonstrate
the effectiveness of concentrating and coordinat ng
the various water quality  management programs of
the USDA and the EPA.  Although the primary em-
phasis of the project  was to investigate interagency
cooperation, the USDA was concerned with the loss of
soil from croplands and the EPA was concerned with
improving the water quality of the lake. Through inler-
agency participation it was determined that the most
effective  methods  for  treating  soil  loss and
associated water quality problems were to apply Bast
Management Practices (BMPs) and construct several
drawdown  type sediment control structures.
  The primary purpose of the BMP's was to reduce
soil erosion from agricultural lands and the primary
purpose of the sediment control structures was to
reduce sediment delivery to  the lake.  Although im-
plicit, the actual reduction of  nutrient concentrations
in the lake was a secondary consideration. Because
the  EPA and the South Dakota Department of Water
and Natural Resources (DWNR) were concerned with
the  water quality  of  Lake Herman,  water quality
parameters were monitored from 1978 until July 1983
to determine if the BMP's and the sediment control
structures were effective in reducing the nutrient con-
centrations into Lake Herman via the tributaries. The
monitoring program was established to accomplish
the  following objectives:
  1. To monitor various water  quality parameters
associated with  the tributaries, the lake, and the
outlet.
  2. To determine  if  the nutrient content in the
tributaries decreased over the duration of the project.
  3. To determine if the sediment control structures
and the BMP's were effective  in reducing nutrients.
  4. To attempt to determine if existing phosphorus
mass  budget  models  adequately  explained
phosphorus concentrations observed in the lake.
  Although the EPA and the  DWNR were concerned
with a  variety  of water quality parameters, the em-
phasis  in  this paper  is  on  total  phosphorus and
suspended solids. Total phosphorus is not the limiting
nutrient in Lake Herman but is considered the nutrient
which,  if  reduced, would possibly decrease noxious
blue-green blooms  by  allowing other algal forms to
                                                82

-------
                                                                 AGRICULTURAL RUNOFF AND WATER QUALITY
compete more  effectively with  the blue-greens. The
suspended solids were considered to be indicative of
soil loss from the surrounding watershed. A reduction
in these parameters  was expected  to  improve  the
water quality in the lake (i.e., increase clarity, reduce
blue-green blooms, and upgrade the trophic status of
the lake from hypereutrophic to eutrophic).
THE EFFECTIVENESS OF  BMP'S AND THE
SEDIMENT CONTROL STRUCTURES

A variety of BMP's was applied to agricultural land in
the Lake  Herman  watershed. They included perma-
nent seeding, terraces,  livestock watering facilities,
windbreaks  and  shelterbelts,  conservation tillage,
water impoundment reservoirs,  wildlife cover plan-
tings, sod waterways, animal  waste control struc-
tures, and rotation seedings. By the end of 1980, ap-
proximately 87 percent of the land had been adequate-
ly treated with BMP's (Fig. 1).
  Four tributary sites were established to monitor the
quality of water draining various sections of the water-
shed (Fig. 2). The water quality data from these sites
were  analyzed using the Tukey-Kramer  method  of
multiple comparisons among pairs of means based on
unequal sample sizes (Sokal and Rohlf, 1981). Table 2
summarizes the results from  the analyses. The con-
centrations  of  total  phosphorus  decreased
significantly (a< .05) in 1980 from the concentrations
observed  in 1979  at tributary Sites 1, 3,  and 4. In-
organic  nitrogen  concentrations  decreased
significantly when compared to the concentrations
observed at all four tributary sites in 1980 and organic
nitrogen only declined significantly at tributary Site 1.
Total suspended solids did not decrease significantly
at any site. After 1981, the results at Sites 1 and  2
could not  be  attributed to the  BMP's but were at-
tributed to the BMP's plus the sediment control struc-
tures. Therefore, the results from Sites 1 and 2 will be
discussed later. In 1983, the total phosphorus and in-
organic  nitrogen  concentrations  did  decrease
significantly (a <.05) from the concentrations observ-
ed in 1979 at Site  3. However, significant reductions
were not observed at Site 4.
  As stated previously, after 1981 any observed reduc-
tion  of nutrients at Sites 1 and 2 would have to be at-
tributed to the BMP's plus the sediment control struc-
tures. One sediment control  structure was located
above Site 1 (i.e., Sediment Control Structure 3) and
two  sediment  control structures were located above
Site  2 (i.e., Sediment Control  Structures 1  and 2). In
1983, the  concentrations  of total phosphorus, in-
organic nitrogen  and organic nitrogen observed at
Site  1 and 2 were significantly less than those observ-
ed in 1979.
  Table 3 presents the results of t-test analysis bet-
ween the means of various  water quality parameters
above and below the sediment control structures. In
Figure 1.—The hatched areas represent areas which have
not  been treated with Best Management  Practices as of
1980. Approximately 87 percent of the Lake Herman water-
shed had been treated with BMP's by 1980.
Figure 2.—The in-lake and tributary water quality sampling
sites. (Sites 1, 2, 3 and 4 are tributary sites; sites 5, 6 and 7
are in-lake sites;  S1 is sediment control structure 1; S2 is
sediment  control structure 2; and S3 is sediment control
structure 3.)
                                                 83

-------
LAKE AND RESERVOIR MANAGEMENT
1982, there were no significant differences between
the upstream and downstream water quality at Sedi-
ment Control Structure 1. Water did not flow into and
out of Sediment Control Structure 2 in 1982 and there-
fore there were no data to compare. Sediment Control
Structure 3 showed a significant (a <.05) reduction in
total phosphorus, total solids, total dissolved solid;;,
and suspended  solids.  In  1983,  Sediment Control
Structure 1  significantly  reduced (a <  .05)  total
phosphorus and  suspended  solids,  and Sediment
Control Structure 3 significantly reduced (a< .05) total
phosphorus. Sediment Control Structure  2 did  not
show any significant  differences. The probable ex-
planation for the nonsignificance observed  at  the
structure is that the water level control valves were
opened before  the material was allowed  to settle
because  of prolonged flooding  of land  behind  the
structure.
  Because the EPA and the DWNR were interested h
reducing  the phosphorus load to Lake Herman and be-
cause the Sediment Control Structures 1 and 3 were
reducing  total phosphorus concentrations, an attempt
was made  to determine the effectiveness of these
structures in reducing the potential phosphorus load
to the lake. Although the associated flows above and
below the structures were not measured, the flows at
Sites  1 and 2 were assumed to be sufficiently repre.-
sentative of flows above and below the structures. Us-
ing  these flows and the total phosphorus concentra-
tions observed above and below the structures a load
was calculated into and  out of the sediment controf
     structures. Sediment Control Structure 1 reduced total
     phosphorus loads approximately 21 percent (standard
     deviation ±11.6) and Sediment Control Structure  3
     reduced total phosphorus approximately 40 percent
     (standard deviation ±25.0) in 1983.


     RESULTS FROM  PHOSPHORUS MASS
     BUDGET MODELS AND INTERNAL
     LOADING

     Although  nutrients  and suspended  solids  were
     decreased by the BMP's and the sediment control
     structures, a corresponding reduction was not observ-
     ed in the lake.  Therefore,  two  phosphorus  mass
     budget models were used to predict the in-lake phos-
     phorus concentration to determine if the in-lake con-
     centrations reflected the phosphorus load (Dillon,
     1975; Reckhow et al. 1980; U.S. EPA, 1980; S. Dak. Dep.
     Water Nat. Resour. 1983). Tables 4 and 5 describe the
     models in their most useful form and their associated
     assumptions. In  general, the predicted values were
     less than the observed values. Assuming the models
     do have some predictive valve, then it must be assum-
     ed that there is  another unaccounted for source of
     phosphorus (internal loading).
       The  possibility of internal loading was  explored
     using a method similar to the methodology described
     by DeGroot (1981). The average total phosphorus load
     per month was determined from 1978 through 1982.
     These loads were designated the gross external loads.
        Table 1.—Summary of morphometric, hydrologic, and phosphorus budget parameters for Lake Herman.
                                 Symbol
1978
1979
                       1980
                                   1981
                                  1982
Drainage area (m-2 10-6)
Lake surface area (m2 106)
Drainage area: lake area
Volume (m3 106)
Mean depth (m)
Maximum depth (m)
Discharge (m3 106)
Flushing rate (yr~1)
Retention
Areal loading (g m-2yr-i)
Volumetric loading (g m-3 yr~1)
Areal water load (m.yr-1)
Ad
Ao
Ad/Ao
V
Z
Z max.
Q
p = Q/V
R
L
LV
qs = Q/A0
1 73.80
5.37
32.4
9.28
1.7
3
6.41
.69
.82
1.362
.788
1.194


15.12
1.63
.76
3.367
1.948
2.816


1.05
.11
.98
.136
.079
.196


0 0
0 0
— —
.138 .074
.080 .043
— —
  Table 2.—Summary of Lake Herman water quality data analyzed by the Tukey-Kramer method of multiple comparisons
       among pairs of means based on unequal sample sizes (T.P. = total phosphorus; INN = Inorganic nitrogen;
         ORN = organic nitrogen; TSS = total suspended solids; TDS = total dissolved solids; and ' =(a< .05).
Parameter T.P. INN ORN
Site 1234 1234 1234
1979 > 1980 *_*« .... *___
1979 > 1982 ___- _ _ * _ ___*
1979 > 1983 ***_ ***_ * • » _
1980 > 1979 _-_- _ _ _ _ _ _ _ _
1980 > 1982 ____ ____ _ * _ _
1980>1983 -_-_ ____ _**_
1982 > 1979 ____ ____ ____
1982 > 1980 *-*_ ____ *___
1982 > 1983 *-*- ___._ _ _ _ _
1983 > 1979 _ _ _ _ ____ ____
1983 > 1980 ____ ____ ____
1983 > 1982 --_- ____ ____
TSS TDS
1234 1234
---_ _____
____ _____
---_ _____
_ _ _ _ * * _ *
_ _ _. * * *
_ _ _ _ * * .. *
____ _____
____ _____
_ _ _ _ _ _ _ _
_ ___ _____
____ ____
_ _ _ _ _ * _ _
                                                84

-------
A net external load was calculated by subtracting the
total phosphorus load leaving the lake from the gross
external load for each month. Using the net external
load and the average  lake concentration, an in-lake
phosphorus concentration was predicted for the next
month. If the predicted in-lake concentration exceed-
ed the observed in-lake concentration, the phosphorus
was assumed to have  settled  to  the bottom. If the
predicted  in-lake  concentration was  less than the
measured  in-lake concentration,  then  the  excess
phosphorus was assumed to be a result of internal
loading.
  Figure 4 illustrates the results of the calculations.
In the figure, the solid line is the gross external load,
the dotted line is the net external load (gross external
load minus the outflow load), and the hatched area is
the phosphorus  flux. This flux  in a positive (upward)
direction is assumed to represent internal loading and
in a  negative (downward) direction  is assumed to
represent settling. The results indicate that internal
loading is important during the years when the gross
external load is low, implying that,  even if phosphorus
is eliminated from external nonpoint sources, internal
loading can be a significant factor for several years.

DISCUSSION

The sediment control structures and/or the BMP's did
reduce the nutrients and suspended solids concentra-
tions. However,  the manner of the reduction was not
consistent.  It was expected that the BMP's and/or
sediment control structures would reduce the concen-
trations of nutrients and this reduction would remain
relatively  constant  as  long as new disruptive in-
              AGRICULTURAL RUNOFF AND WATER QUALITY

fluences did not occur in the watershed. However, at
tributary Sites 1 and 3, a significant increase (a<.05)
in total phosphorus was observed in 1982 as opposed
to the concentrations observed in 1980 and in 1983.
  There are several possible explanations for the lack
of a relatively constant decrease in nutrients and
solids resulting from the presence of sediment control
structures. The most plausible explanation is the lack
of proper water level valve operation. According to the
operational procedures, all valves are to remain clos-
ed until  water stops flowing behind the  structure.
Once the flow  has stopped, the material in the water is
allowed to settle for 72 hours and then the upper water
level control valve is opened. After the upper layer of
water has been evacuated, the remaining water plus
associated particles is allowed to settle for another 72
hours.  After the second 72-hour settling period,  the
lower water level control valve is opened, evacuating
the second layer of water. These valves have been
discovered open when water is still flowing upstream
of the structure.
  Another possible explanation for the inconsistent
reduction could be attributed to the type of  material
delivered from the cropland. It has been observed that
sediment that is eroded  from cropland contains a
higher percentage of finer and lighter particles than
the soil from which it originates (Robillard et al. 1982).
These  lighter  particles  (i.e.,  clays  and  organic
residues)  may remain in suspension  longer than the
two 72-hour periods allocated for settling behind the
sediment control structures. If a selective erosion pro-
cess is occurring then the overall pollutant  delivery
can  increase  because these small particles have a
greater adsorption capacity for other pollutants than
        Table 3.—The results of t-test analyses between the means of water quality parameters above and below the
                                          sediment control structures.

Sediment
Control
Structure 1

Total phosphorus
Ortho-phosphate
Inorganic nitrogen
Organic nitrogen
Total solids
Dissolved solids
Suspended solids
1982
NS
NS
NS
NS
NS
NS
NS
1983
* *
NS
NS
NS
*
*
* *
Sediment Control
Structure 2
1982 1983
- NS
- NS
- NS
— NS
- NS
- NS
— NS
Sediment
Control
Structure 3
1982
* *
NS
NS
NS
* *
* *
**
1983
* *
* *
*
NS
NS
NS
NS
   • Downstream greater than upstream (a < 05)
   *' Upstream greater than downstream (o< .05)
   NS Non significant

               Table 4.—The Dillion (1975) loading concentration model based on phosphorus mass balance.


   [P] = L(1 - R)/zp
   [P] = concentration of total phosphorus (mg»m-3) in the water column at spring overturn
   L  = areal load total phosphorus (mg»m-2.yr-1)
   R  = phosphorus retention coefficient
   p  = flushing rate(yr-i)
   z  = mean depth

   Assumptions: (EPA, 1980)

   1. The lake is completely mixed.
   *2. Rate of supply of phosphorus, the flushing rate, the lake volume, and the sedimentation coefficient are constant
      through time.
   3. The outflow P concentration equals the lake p concentration.
   *4. The lake is in steady state; that is, phosphorus concentration does not change over time.
   •Violated
                                                   85

-------
LAKE A^D RESERVOIR MANAGEMENT
Table 5.—The Recknow et. al. (1980) loading concentration
       model based on phosphorus mass balance.

 P = UVS qs
 P = phosphorus concentration, mg/l
 L = areal load of phosphorus concentration (g/m2»yr)
 qs = areal water load (m/yr)
 vs  = apparent phosphorus settling coefficient (m/yr)

 Limitations: Reckhow et. al. (1980)
Variable
*1 P
**2 L
***3 qs
Minimum
.004 mg/l
.07 g/m2yr
0.75 m/yr
Maximum
.135 mg/l
31.4g/m2yr
187 m/yr
  * Violated in all years sampled
  *" Violated once out of 5 years sampled
 **' Violated three times out of 5 years sampled
>1000
800
700
600
500
400
300
200
100 ;
x 0 '
Dl
£
-100
-200
-300
-400
-500
-600
-700
-800
/v







ml
.'.}'• ^ ;









'

.
jl
fh

\
f

'








1 J
lP
y













M
3 ^
1 i











i
^
1













3
'1
^
.;
^











1





|
L






c





















1








I
f
|
!
^
'/
'//
     J  A  J  0 J  A  J  0 J A  J  0  J A  J  0  J  A J 0 J
       1978
               1979
                       1980
                                 1981
                                          1982
\ J
1983
 3.
Figure 3.—Phosphate loadings of Lake Herman in mg P/m'
month. Uninterrupted line: gross external loading. Inter-
rupted line: net external load (inputs minus outputs). Hatch-
ed:  internal  loading with  phosphorus release positive  and
phosphorus settling negative.
 do large particles (Robillard et al. 1982). Therefore, if
 selective erosion is occurring, a significant decrease
 in nutrient concentrations might not be observed.
   A third  possible explanation for the inconsistent
 decrease in nutrients could be because sediment and
 flow measurements were not collected directly above
 and below the sediment control structures. Therefore,
 much of the information needed  to determine the ef-
 fectiveness of the  BMPs and the sediment control
 structure may be missing.
   Although  consistency  has not been  observed,
 nutrients have decreased. In addition, there have been
 years (e.g., 1981) when runoff carried no nutrient loads
 into  the lake. However,  there has not been a cor-
 responding nutrient concentration  decrease in the
 lake. Phosphorus  mass budget models predict phos-
 phorus concentrations less than those observed in the
 lake. Therefore, the possibility of internal loading was
 explored and found to be an important factor in those
 years when the gross external load of phosphorus is
 low. Hence, even if the nutrient load from the surroun-
 ding watershed is reduced, the  effects  will not  be
 observed for several years in the lake.


 REFERENCES

 Churchill, C.L, and C.K. Brashier. 1972. Effect of dredging on
  nutrient levels and biological population of  a lake. Final
  rep. U.S. Dep. Inter., Off. Water Resour. Res., Proi. Number
  B-013-SDAK.
 Dillon, P.J.  1975. The phosphorus budget of Cameron Lake,
  Ontario: the Importance of flushing  rate to the degree of
  eutrophy of lakes. Limnol. Oceanogr. 20:28-39.
 DeGroot, W.T.  1981. Phosphate and wind in a shallow lake
  Arch. Hydrobiol. 91:475-89
 Rast, W., and G.F. Lee. 1979. Summary analysis of the North
  American (U.S. portion) OECD  eutrophication  project:
  Nutrient loading-lake response relationships and trophic
  state indices. EPA-600/3-78-008. U.S. Environ. Prot. Agency,
  Washington, D.C.
 Reckhow, K.H., M.N. Beaulac, and J.T. Simpson. 1980. Model-
  ing phosphorus loading and lake response under uncer-
  tainty:  A manual and compilation of export  coefficients.
  EPA 400/5-80-011. U.S. Environ. Prot.  Agency.
 Robillard, P.O.,  M.F.  Walter,  and  LM.  Bruckner.  1982.
  Planning guide for evaluating agricultural nonpoint source
  water quality controls.  EPA  600/3-82-021. U.S.  Environ.
  Prot. Agency.
 Sokal,  R.R., and F.J. Rohlf. 1981.  Biometry:  The Principles
  and  Practice of Statistics in Biological Research. W.H.
  Freeman and Co., San Francisco.
 South Dakota Department of Water and Natural Resources,
  1980. Lake Herman Model Implementation Program: Three
  Year Rep.
	1983.  Lake  Herman  Model   Implementation Pro-
  gram: Final Rep. (In prep.).
 U.S. Environmental  Protection  Agency. 1980. Clean Lakes
  Program Guidance Manual. EPA 440/5-81-003.
                                                     86

-------
                State  Program   Development:
                                   Priorities  &  Strategies
PROCESS TO IDENTIFY, SCREEN AND PRIORITIZE  RURAL WATER
RESOURCE AND LAKE REHABILITATION PROJECTS IN  ILLINOIS
THOMAS E. DAVENPORT
Illinois Environmental Protection Agency
Springfield, Illinois
           ABSTRACT

           Initial water quality management planning efforts documented that agricultural activities are a major
           source of pollution in Illinois and mandated the development of plans to control this nonpoint source
           pollution from agriculture. The most severe agriculturally-related problem is soil erosion and its ef-
           fects upon the aquatic environment. In Illinois estimated gross erosion exceeds 180 million tons an-
           nually, 88 percent of it caused by sheet and rill erosion from cropland. To control agricultural non-
           point pollution from its source, efficient and effective land management practices and programs must
           be developed. A fundamental component of this strategy is the identification of specific areas that
           significantly contribute to the problem, to permit targeting of resources. Illinois Department of Agriculture
           (IDOA) formed the Soil Erosion and Water Quality Advisory Committee (SEWQAC), which implemented
           a two-tier targeting system. Local targeting within each district is the first level. The second is statewide
           targeting to solve problems identified locally that cannot be addressed with local resources. A sub-
           committee formed by SEWQAC developed a uniform process to identify, screen, and prioritize rural
           water resource and lake rehabilitation projects within the State. The process provides a uniform and
           systematic method for local Soil and Water Conservation Districts, ASCS County Committees, and
           other interested local units of government to identify and compete for funding under three program
           authorities. Designed to set meaningful State priorities, the system provides equal access for each
           project to all the program authorities and gives the local county responsibility for identifying and prioritiz-
           ing its projects. The local Soil and Water Conservation Districts are assisted by a State Association
           Water Quality Coordinator and assistant funded through an IEPA contract. The process has been
           successful to date.
BACKGROUND

Initial water quality  management  planning efforts
documented that agricultural activities are a major
source of pollution  in Illinois and mandated  the
development of plans to control this nonpoint source
pollution. The most severe agriculturally related prob-
lem identified  was soil erosion and its effects upon
the aquatic environment. During the initial planning
process, financial and technical resources needed to
correct the identified  problem were estimated. There
has been a considerable shortfall  between these re-
source estimates and current levels of appropriated
resources.
  To control agricultural nonpoint pollution from its
source, an efficient and effective program had to be
developed. This program had to include technical,
financial and educational components; all  three are
needed to concentrate  and deliver assistance in a
timely and acceptable manner. The strategy  is to iden-
tify specific areas that significantly contribute to the
problem. The ability to identify and quantify "source"
areas allows for targeting of resources and  programs
to correct the identified problem. Therefore, to develop
a comprehensive implementation strategy and to im-
prove coordination between the affected agencies, the
                                             87

-------
 LAKE AND RESERVOIR MANAGEMENT
 Illinois Department of Agriculture formed the Soil Ero-
 sion and Water Quality Advisory Committee (SEW-
 QAC). SEWQAC represents a wide diversity of interest
 from throughout the State. Its 28 members represent
 gubernatorial appointees,  farm organizations, State
 and  Federal  agencies,  and  statewide agricultural
 organizations.
   SEWQAC was formed to advise and assist local,
 State, and Federal agencies in coordinating soil ero-
 sion and water  quality  programs, developing long-
 range planning programs,  making recommendations
 for  legislative  action,  resolving  mutual problem;;,
 meeting training needs,  and developing annual pro-
 gress reports to comply with the approved water quali-
 ty management plan for the State of Illinois.
   SEWQAC identified three broad areas of  concern:
 education, project coordination on small watershed:;,
 and  regional  targeting of manpower and  financial
 resources; all required in-depth evaluation and plann-
 ing  before the  implementation  strategy  could be
 developed (Fig. 1). Subcommittees comprised of tech-
 nical  employees of various  agencies and organiza-
 tions  were formed to make policy recommendations
 to the SEWQAC in these three areas. SEWQAC's sub-
 committees  are:  (1) State  Watershed  Priority, (2)
 Education, and (3) Long-range Planning.
   State Watershed Priority Subcommittee. The first
 subcommittee formed by SEWQAC, it has focused on
 developing a consensus for targeting resources and
 setting priorities to efficiently use available manpower
 and financial resources to solve critical soil and water
 conservation problems on a watershed scale.
  Education Subcommittee. The goal of this subcom-
 mittee is to facilitate educational information dissemi-
 nation  efforts  in  the  State. The subcommittee is
 developing a procedure to coordinate the educational
 efforts of all agencies involved in soil and water con-
 servation education  programs. It  is  developing  a
 method to target those education efforts into high-
 priority geographic areas. The education program will
 promote wiser land use and highlight current research
 findings related to soil and water conservation. The
 Cooperative Extension Service  is the lead agency in
 the soil erosion control program.
  Long-range  Planning Subcommittee. This subcom-
 mittee was formed to provide information and data for
 the development of a long-range implementation plan
 and  to establish criteria for targeting manpower and
 financial resources on a regional basis. This subcom-
 mittee has selected three resource problem areas to
 be used as indicators in setting these priorities: (1) soil
 erosion, (2) water  management,  and  (3) land use.
 Social and economic considerations are considered
 major components of each of the three major  problem
 areas.
         THE PROCESS

         In response to  the need for assistance, a  two-tier
         targeting system  was developed.  Local  targeting
         within each district is the first level. This  local pro-
         cedure   is outlined in "A  Procedure for  Settling
         Priorities on Soil and Water Resource Problems in Il-
         linois."  This procedure consists of seven activities to
         be completed in order, with a schedule for accom-
         plishments and evaluation. The Illinois Department of
         Agriculture is in the process of evaluating the status
         of this procedure in all 98 districts. To date, 93 percent
         of the districts that responded have completed the in-
         itial   prioritization  of  local  areas. The individual
         districts should  develop long-range plans to use this
         targeting effort to meet the goals of the water quality
         management plan.
            To determine  the effectiveness  of local targeting
         and the type of lands on which conservation practices
         are being applied, the Soil Conservation Service has
         developed a soil erosion  reduction accounting pro-
         cedure.  Steve Probst, SCS  resource conservationist in
         Champaign, III.,  is  the contact person  for this pro-
         cedure.  This procedure is tied to the Soil Erosion and
         Sediment Control  Guidelines' interim and  overall
         goals. This act designates the 98 Soil and Water Con-
         servation Districts as the implementation agencies for
         this program. The program's overall goal is to reduce
         erosion  on all subject lands to "T" by the year 2000.
         "T" is the average  annual tons per acre soil  loss a
         given soil may experience and still maintain  its pro-
         ductivity over an extended period of time.
           The second tier is statewide targeting to solve prob-
         lems identified locally that cannot be addressed with
         local resources, because of their magnitude. Districts
         were asked to identify these problems on a watershed
         basis.
           The subcommittee, formed by SEWQAC, developed
         a uniform process to identify, screen,  and prioritize
         rural water resource and lake rehabilitation projects
         within the State  (Fig. 2). The  process provides a uni-
         form and systematic method for local Soil and Water
         Conservation Districts, Agricultural Stabilization and
         Conservation Service County Committees, and  other
         interested local  units of government to identify and
         compete for funding under three program authorities.
         The system provides equal access for each project to
         all the available program authorities. It is intended to
         be equal and fair with the primary objective being to
         set meaningful State priorities. The system is  design-
         ed to provide the local county  with  the maximum
         amount  of responsibility to identify and prioritize their
         projects. The local Soil  and  Water  Conservation
         Districts are assisted in this effort by a State Associa-
         tion Water Quality Coordinator and  assistant who are
         funded through a contract with Illinois EPA.
                                    SOIL EROSION Af\D WATER QUALITY
                                         ADVISORY COMMITTEE
STATE WATERSHED PRIORITY
SUJ-CiMIITTEE
EDUCATION
SUd-LOMMITTEE
LONG-RANGE PLANNING
SUB-COMMITTEE
Figure 1.—Process for identifying and ranking Illinois lake rehabilitation projects.
                                                  88

-------
                                                        STATE PROGRAM DEVELOPMENT: PRIORITIES & STRATEGIES
Selection Process

1. The most important part of the screening process
begins at the county level. The local Soil and Water
Conservation District within each county will serve as
the primary contact and coordinating agency. Each
member of the SWCD and ASCS County Committee
will serve as voting members on the County Water-
shed Priority Committee to identify, screen, and prior-
itize projects.  In addition to these primary agencies,
the Committee should seek the input from other local
groups: municipalities, county or regional  planning
commissions, conservation and  park districts, farm
organizations, SCS, ASCS, Extension personnel, and
others.
  The County Watershed Priority Committee should
identify potential projects, using the critical area in-
ventory being developed by the SWCD, in cooperation
with other local agencies.  Information and data con-
cerning water resources within their jurisdiction were
forwarded to the individual districts by the appropriate
State  agencies. The Committee will evaluate these
projects to determine if they are of local significance
and have local support. They should seek the advice
from the appropriate agencies and groups, and iden-
tify projects on the basis of need and the amount of
local support.
  The County Watershed Priority Committee will be
given technical assistance by Illinois Department of
Agriculture regional coordinators and SCS area  and
district conservationists. These advisors  will work
              PROCESS SCHEMATIC TO IDENTIFY, SCREEN, AND

              PRIORITIZE RURAL WATER RESOURCE AND

              LAKE REHABILITATION PROJECTS IN ILLINOIS
                   .Land Use
                   louncil
                   Prioritization
                   State Watershed
                  -Cornni ttee 4. ~ -
                   Screening
         Non prioritized
         projects
         for
         Implementation
         or
         Planning
         Assistance
          Projects
          ready for
          planning and
          Implementation
          3 categories
          Soil E rosion and
          Water Quality
          Advisory committee
Projects i
additional data/study
          Implementation
            I
          Yearly Progress
            Report
Figure 2.—Organization of advisory committee activities.
with the County Watershed  Priority Committee  to
evaluate the projects and determine if they meet the
basic State criteria. These criteria are designed  to
quickly  identify  those projects that cannot  be con-
sidered  actively for any of the program  authorities
under this system.
  The County Watershed Priority Committee will then
screen the projects and  determine those which meet
State criteria and have sufficient local need and sup-
port. Projects that meet these criteria will be ranked 1,
2, 3 .... There is no upper limit to the number of pro-
jects. Although a county  may submit  an unlimited
number of projects, they are encouraged to make the
screening  and prioritization process meaningful  so
that the projects supported by the committee meet
State criteria, have strong local support, and can move
ahead rapidly if selected.
  Once the County Watershed Priority Committee has
selected and ranked projects, the Cooperative Exten-
sion  Service, SCS, ASCS, and Illinois  Agriculture
regional coordinators shall individually  make written
comments regarding each project and submit them to
the members of the Land Use Council Watershed
Priority Committee. Each project will submit a com-
pleted watershed application to the Land Use Council.
The application and agency comments will serve as
the written basis for the  Land  Use Council evaluation.

  2. Land Use Council-Watershed Priority Committee.
The  makeup of  the Land  Use Council-Watershed
Priority  Committee will  be one representative from
each Soil and Water Conservation District and County
Committee from that Council area. These representa-
tives  will  make up the voting membership and will
evaluate and  rank  projects. The Committee  should
also seek  the assistance of  additional  groups,  in-
cluding  farm organizations, State  and Federal agen-
cies (which are  members  of  the  State  Watershed
Priority Committee), a member of the AISWCD staff or
executive  board, members of the local planning com-
mission, or others. The Land Use Council is encourag-
ed to seek input from a wide base of interest.
  The Council shall consider each  project submitted
by the counties and determine whether  or not it war-
rants further  consideration. The  Council will then
place projects  under additional consideration  into a
priority  ranking. Again, there will be no limitation in
the number of projects that may be submitted  by the
Land Use  Council  to the State Watershed  Priority
Committee. However, projects should  be limited to
those that have the greatest local need and support.
The projects which  have  received priority ranking from
the Land  Use Council will then be submitted  to the
State  Watershed  Priority Committee  for  further
evaluation and screening.
  The Land Use Council should supply written com-
ments,  along  with  its priority ranking, to the State
Watershed  Priority  Committee.  These  comments
should explain how projects were ranked. The Council
should  also explain to each  county  submitting pro-
jects why a project was  dropped from consideration.

  3. State Watershed Priority Committee. This com-
mittee will use the application and the  comments of
the Cooperative Extension Service, the SCS area and
district staff, and the Illinois Agriculture regional staff
to  determine those  projects  that need additional
study. These will be submitted to  the Soil Conserva-
tion Service River Basin Planning  Staff for technical
data collection.
  The Watershed  Priority  Committee  will  use the
available data to divide projects into three areas:
                                                   89

-------
 LAKE AND RESERVOIR MANAGEMENT
                                 Table 1.—Eleven priority watersheds selected, 1979.
Ranking
(D
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)

Watershed
L Pittsfield*
L. Carlinville
L. Canton
Silver L*
Spring L.
L. Springfield
L Taylorville
L. Lou Yaeger
L. Bloomington
Paris L.
L. Paradise

Watershed
Size
(ha)
2,890
6,755
3940
12,317
5,236
66,973
34,029
29808
17621
5 184
4690

Lake
Size
(ha)
98
68
101
223
112
1,630
465
514
257
89
71

Project
Status
National ACP Special Water Quality
AGP Special 1980

RCWP
P.L. 83-566
RBSL
RBSL



* *

  ACP:     Agricultural Conservation Program
  RCWP    Rural Clean Water Program
  P.L 83-566. Small Watershed Program Implementation
  RBSL-    River Basin Study Level (PL:83-566)
  *       Comprehensive monitoring programs
  **'      Illinois Department of Agriculture, sediment dredging project
   1. Projects that should not be considered further for
 planning or implementation assistance,
   2.  Projects needing additional data and/or study,
 and
   3. Projects ready for planning and implementation
 assistance.
   Those projects needing additional data or study will
 be submitted to the Soil Conservation Service River
 Basin Planning  Staff for additional study,  a highe-
 degree of data collection, and the development of im-
 plementation alternatives,  a function of the River
 Basin Study  Level  program. The  reports generated
 through  this effort can easily be modified for applica-
 tion to the P.L. 83-566 implementation program. Pro-
 jects that go through this process will be  resubmittecl
 to the Watershed  Priority Committee for additiona
 evaluation and screening to determine whether they
 are ready for planning and implementation or should
 be sent back to the sponsor.
   State  Watershed  Priority Committee will submit a
 list of recommended projects by category to the State
 Soil Erosion and Water Quality Advisory Committee
 for final  prioritization. Each project will be placed intc
 one of three different categories for separate prioriti-
 zation and evaluation. These are:
   1. In-lake projects, which fall within the authority ol
 the 314 Clean Lakes Program;
   2. Out-of-lake projects, which  have off-site damages;
 from the watershed but  whose potential solution lies,
 wholly within the watershed. Examples of programs
 are the Rural Clean  Water Program, Agricultural Con
 servation Program Special, ASCS programs, and P L
 83-566 SCS Program.
   3. Projects that combine in-lake and out-of-lake pro-
 jects. These projects have off-site damages and the
 potential solution is both in-lake  rehabilitation  and
 watershed management.
  The Illinois Department of Agriculture is developing
 a water resource  project tracking system for the State.
 It will be  used as  a management tool in assisting local
 sponsors (SWCD's) and for documenting the extent of
 water resource problems in the State.
  4. State Soil  Erosion  and Water Quality Advisory
 Committee. This  committee will then review each of
the projects  within  the  various program authorities
and make final  recommendations to the appropriate
agency with the implementation or priority setting
responsibility.
  These agencies will use these recommendations to
establish their own priorities and take appropriate ac-
tion for implementing their program.
  Each  project selected  for  priority  assistance
through this process will be evaluated annually by the
Soil Erosion and Water Quality Advisory Committee.
The responsible planning and implementation agency
will submit a written  report to the Committee stating
the work items during the past  year and proposed ob-
jectives for the upcoming year. In addition, the State
Watershed  Priority  Committee  will semi-annually
review the technical  progress of each  project.  The
review will include a detailed discussion  between
responsible planning and implementation  agencies
for  each project. The soil erosion reduction accoun-
ting procedure developed by SCS will be used to see if
       Table 2.—Status of selected watersheds.

Projects Under P.L. 83-566
    1. Ash Loop (implementation)
    2. Spring Lake (implementation)
    3. Kinkaid Lake (seeking plan authorization)
    4. Raccoon Lake (recommendation from the
Governor's Office to SCS)

River Basin Study Level
    1. Kinkaid Lake
    2. Raccoon Lake
    3. Lake Sara
    4. Fountain Creek
    5. Indian Creek (Lake Shabbona)
    6. Lick-Pole Creek (Lake Springfield)
    7. Waverly Lake
    8. Waukarusa Creek
    9. North Pope Creek
        A. Vandalia Lake
        B. Stephan Forbes
        C. Long  Lake
        D. Lake Taylorville
        E. North Oakley (Lake Decatur)
                                                   90

-------
                                                       STATE PROGRAM DEVELOPMENT: PRIORITIES & STRATEGIES

the high priority areas are being treated first within    since inception; 15 of the 33 projects submitted have
each watershed.                                      been selected for various programs. Table 2 shows the
  Through the initial water quality management plan-    project selection by program (River Basin Study Level
ning process, 11 priority watersheds were selected in    and P.L 83-566) that has occurred  through  uniform
1979. Table 1 shows the status of activity within these    process.
watersheds. There have been three selection cycles
                                                  91

-------
 MANAGEMENT  PLANNING FOR 25
 NEW JERSEY  LAKES
 JOHN BRZOZOWSKI
 New Jersey Department of Environmental  Protection
 Trenton,  New Jersey


 STEPHEN J. SOUZA
 Princeton Aqua Science
 New Brunswick,  New Jersey



             ABSTRACT

             The State of New Jersey has approximately 1,100 lakes of which 345 are publicly owned. Many of
             these lakes are located in urban-suburban areas with high density residential and commercial land
             use activities on their watersheds. In 1975, the New Jersey Department of Environmental Protection
             (NJDEP) initiated a lake management program, inventorying and conducting water chemistry sampl-
             ing in over 450 of these lakes. Based on these observations, over 30 percent of these lakes were
             presumed eutrophic. Final determination of trophic status and the development of management plans,
             however, necessitated more extensive analysis. To accomplish this NJDEP secured a $100,000 Sec-
             tion 314 Lake Classification Grant on Nov. 17, 1979. Of the lakes inventoried, 25 were selected for
             intensive study. They represent a reasonable cross section of the types of lakes and lake problems
             in New Jersey. These 25 lakes were ranked on the basis of their trophic status as determined by land
             use, unit areal loading, methodology, and verified to the extent possible by intensive survey data.
             Point and nonpoint sources of nutrient loading were identified and quantified. Reduction in nutrient
             loadings required to improve their trophic status was estimated and management and restoration recom-
             mendations to accomplish nutrient reductions described. This evaluation indicated that, for the most
             part, nonpoint sources related to urban, suburban, and in some cases agricultural stormwater runoff
             were the major source of nutrient loads to these lakes. The identification of storm water as the major
             source of loadings to these lakes is timely since New Jersey has just finalized new storm water quali-
             ty management regulations.
 INTRODUCTION

 New Jersey is a state of contrasts. Located in  the
 center of the eastern megalopolis, it is known for its
 turnpikes,  oil  refineries,  and  chemical  industry.
 However, it also has more than 100 miles of seashore,
 extensive agricultural land, and the largest natural
 preserve on the east coast, the New Jersey Pinelands.
   The  State's lakes are also diverse.  They include
 such major lakes and impoundments as Lake Hopat-
 cong, Greenwood Lake, Round Valley Reservoir and
 Spruce Run Reservoir. These lakes are about 810 hec-
 tares  (2,000 acres)  in  size and are major regional
 recreation areas, and,  in  some cases,  double  as
 significant  water  supplies. Smaller,  privately
 developed lakes, several  of glacial origin, abound in
 the northern tier. In contrast, southern  New Jersey is
 known  for its reddish-brown tinged,  slightly buffered
 acidic (dystrophic) ponds, many of which once sup-
 plied water for cranberry production.
  Surprisingly, outdoor recreation is the  State's  se-
 cond largest  industry and  much of the  industry is
 associated with water. However, because it is the Na-
tion's most densely populated State, urbanization has
greatly affected water quality and recreational use of
many lakes. Lakes which 30 years ago supported only
sparse seasonal populations are now local or regional
centers, convenient to industrial/commercial centers.
  New Jersey residents and the State's Department of
Environmental  Protection (NJDEP) are facing prob-
lems on a scale not experienced by many other States.
Recognizing the need to assist local governments in
managing  their  lake resources,  the  Department
created a Lakes Management Program in its Division
of Water Resources in 1975. The objectives of the pro-
gram were to centralize and coordinate existing lake
management related programs throughout the Depart-
ment, offer technical assistance both within and out-
side of the Department, and develop a statewide lakes
management strategy to best use Federal and State
funds available for restoration activities.
  The first step was to inventory the State's public
and  private lakes. More than 1,000 have since  been
described, with over 400 of them public. It was clear
that  small impoundments (6-24 ha) (15-60 acres) form-
ed a significant  component of New Jersey's  lake
resources.  A limited water chemistry sampling pro-
gram was implemented.  However,  the final  deter-
mination of trophic state  and the  development  of
management  plans necessitated   more  extensive
chemical and biological analyses.
  Clean  Lakes Program Phase I study  grants  have
proven to be an effective vehicle for obtaining needed
information on large complex lake systems,  i.e.,  Lake
Hopatcong and Greenwood Lake. However, the infor-
mation needed to formulate sound management pro-
grams for these lakes did not justify individual, large
scale diagnostic programs.
  Therefore, to carry  out  this program, the Depart-
ment secured a $100,000 U.S. EPA Clean Lakes Pro-
gram grant to  conduct a Lake Classification  Study.
Administratively, using the  Lake Classification grant
                                                 92

-------
                                                      STATE PROGRAM DEVELOPMENT: PRIORITIES & STRATEGIES
proved to be an effective strategy for producing in-
dividual management  programs for 25 small priority
lakes and waterbodies. With only minor supplemen-
tary  information,  many  of  these projects can  now
qualify for Phase II Clean Lakes Program funding.
  Specific objectives of the  Lake Classification Study
were to:
  • Characterize the general physical, chemical, and
biological properties of selected priority lakes.
  • Determine lake trophic status.
  • Estimate  current nutrient loadings, using both
land use models and field data.
  • Recommend  appropriate  restoration/manage-
ment techniques to meet the problems and conditions
unique to each lake.
  •  Prioritize candidate lakes for restoration funding.
  •  Form the basis of a statewide lakes management
strategy.
  Areawide planning agency recommendations form-
ed the basis for selecting the 25 priority lake systems
for study. The criteria considered in the selection pro-
cess included:
  •  Size and location.
  • Water quality as indicated by existing data.
  •  Public accessibility.
  •  Potential  user population.
  •  Historical  uses of  the lake and  practicality of
restoration  to condition  to  re-establish  historical
use(s).
  The location of the 25 lakes  is indicated on Figure 1.
 Figure 1.—Location of  25 lakes  in the New Jersey Lake
 Classification Study.
GENERAL METHODS

An  intensive  1-year  sampling  program  was  im-
plemented.   Pertinent  physical,   chemical,  and
biological parameters were monitored for each of the
25 lakes and their major tributaries. These data were
used to describe existing conditions and to identify
their problems.
  The  field  data  were  further  supplemented  by
nutrient loading estimates  developed by unit areal
loading (U.A.L) methodology based on land  use  in
each watershed (U.S. EPA, 1980). Aerial photographs
of each watershed were used to develop land use
maps and from these maps the area of each land use
was determined. Loading coefficients for total phos-
phorus, total nitrogen, and suspended solids from the
U.S. EPA Program Guidance Manual (U.S. EPA, 1980)
were adapted to these land  use areas. This U.A.L in-
formation helped formulate a relationship between
watershed  development  and nutrient  loading.  The
loading from point source discharges into the lakes
and their tributaries  was also quantified.  Empirical
trophic state relationships of Dillon (1974), Kirchner
and Dillon (1975), Ostrofsky  (1978), Dillon and Rigler
(1974),  and Smith and  Shapiro (1980) were used  to
estimate the existing trophic status  and to provide a
means by which a reduction  in nutrient loading could
be related to improvement in lake trophic status.
  The data developed during the intensive survey and
the information  developed  through the use  of the
U.A.L.  methodology and  the empirical  trophic state
models were compared.
  In addition, the Trophic  State Index of Carlson
(1977)  was  calculated  using  the   mean  summer
chlorophyll a measured in each lake. The scores for
each lake were compiled, and a ranking system based
on the magnitude of scores was used to compare the
relative trophic state  of each lake.
  It should be noted that in many cases the measured
and predicted conditions in any one lake differed con-
siderably. However, specific knowledge of each lake
developed by the NJDEP staff during  the Intensive
Lake survey generally was sufficient to identify the
reasons  for the differences and  enabled  sound
judgments concerning  restoration and management
to be made.
 RESULTS

 Table  1  summarizes some of the  most  pertinent
 characteristics of the 25 lakes in the  study. For most
 of the parameters listed there is great variation within
 the set of 25. However, when these data are examined
 critically some generalizations concerning the 25
 lakes can be made.
   The lakes  in the study were, on the average, very
 small, shallow impoundments  with proportionally
 large watersheds in relation to the size and volume of
 the lake. The large watershed area  to lake surface
 area  ratio  results  in  a short  average  hydraulic
 residence time.
   Therefore,  these lakes are susceptible  to frequent
 flushing, especially during wet seasons. However, dur-
 ing periods of little rainfall stagnation occurs. In reali-
 ty, these  lakes  have  highly variable  hydraulic
 residence times depending  primarily upon meteoro-
 logical conditions.
   The watersheds of most  of the lakes are highly
 developed, especially in the  immediate vicinity of the
 lake. On the average, 28 percent of the watershed area
 is in urban and suburban type land use. In some cases
                                                  93

-------
 LAKE AND RESERVOIR MANAGEMENT
 as high as 70 to 75 percent of the watershed is highly
 developed. In 11 of the 25 watersheds agricultural land
 uses comprise significant  areas  of  the  watershed.
 These land use types tend to contribute the greatest
 nutrient and suspended solids  loads to the lakes via
 stormwater runoff.
   Point sources, sewage treatment plants, discharge
 into tributaries of  five of the 25  lakes. When point
 sources were  present  they usually  contributed a
 substantial portion of the total  load to the lake; how-
 ever, in  all cases  nonpoint sources  were also  sub-
 stantial.
   Most  of  the study lakes have elevated nutrient
 levels.  Because of excessive  nutrient enrichment,
 systems previously assumed to be phosphorus limited
 at times appear to be nitrogen limited. However, some
 lakes  are  presently so productive that physical  or
 biological  factors  may  be  limiting  rather than a
 specific nutrient. Excessive  primary production has
 led to the gradual accumulation of organic sediments
 which  serve as  a significant, although unquantified in-
 ternal  nutrient  source. This organic sediment has a
 high benthic oxygen demand  that causes oxygen
 depletion at the sediment-water interface and results
 in low  redox potential in the sediments. These condi-
 tions  create the  potential for significant internal
 phosphorus recycling.
  Additionally,  sediments have been found to be wide-
 ly contaminated with low levels of toxic substances,
 predominantly metals,  pesticides,   and RGB's.
 Although  in some cases fish flesh was found to be
 tainted with these substances, no action has yet been
 warranted to restrict the consumption of  freshwater
 species.
  The  trophic  status of each  of  the  25  lakes was
 calculated according to Dillon (1974). This  determina-
 tion was based on predicted spring total phosphorus
 concentrations  derived by Unit Areal Loading Metho-
 dology (Fig. 2). All but two of the lakes were classified
 as having eutrophic loading. The two that did not fall
 into the eutrophic category were unusual cases, and
 in at least one case, the basic assumptions of the
 Dillon relationship would have to be questioned. The
 Dillon  relationship was used to estimate how much
the nutrient  load must be  reduced to achieve the
trophic state necessary for the intended use of each
 lake.
  In addition to  the  Dillon (1974) relationship,  the
Trophic State  Index (TSI)  of  Carlson (1977) was
calculated, based on the mean summer chlorophyll a
concentration measured in each lake. The  ranking of
the lakes according to Carlson's TSI is presented in
Figure 3. The Carlson index is a continuous scale that
does not  have  any fixed numerical values for  the
                      FIGURE 2

                      LAKE TROPHIC STATUS - BASED ON PREDICTED
                      SPRING TOTAL PHOSPHORUS CONCENTRATIONS
                      COMPUTED FROM UAL DATA USING THE DILLON
                      MODEL
                    MEAN DEPTH ?(m)
Figure 2.—Lake trophic status—based on predicted spring
total phosphorus concentrations, computed from U.A L data
using the Dillon Model.


various trophic states. In this case, it is used solely as
a  ranking  tool.  The Carlson  index  when based on
measured  mean  summer  chlorophyll  a  is  not
necessarily indicative of the total potential primary
production, especially in shallow lakes where signifi-
cant benthic algae and  macrophyte populations may
also be competing for resources. Since phytoplankton
is  only one  component  of  the  primary  producer,
populations in the lake, production by other primary
producers is  not accounted for in the chlorophyll a
measurement. Thus,  in many cases it would tend to
underestimate the actual trophic state.
DISCUSSION

From the information developed in this study, it has
become apparent that nonpoint source pollution is the
major  cause of  lake degradation  in  New  Jersey.
Although, where point sources are present, they con-
stitute a significant proportion of the nutrient budget,
                 Table 1.—Morphometric and hydrologic characteristics of the 25 lakes in the survey,
Lake Characteristic/
Statistic
Surface area (ha)
Watershed area/
surface area ratio
Watershed area (ha)
Maximum depth (m)
Mean depth (n)
Volume (m3)
Mean hydraulic
residence time
(days)
Range
Low High
4.0

1.6
10.0
1.22
0.76
6x10"

1.72

- 117

- 628
-11650
5.5
2.74
-3.2x106

- 781

Mean (x)
17.2

211
2905
2.55
1.45
3.:2x1Q5

25

Standard
Deviation
22.5

188
3134.3
0.98
0.43
6.2x105

37

N
25

25
25
25
25
25

241

	 Median
9.7

141
1862.7
2.44
1.37
1.3x105

5.47

1One outlier was deleted because of questionable data.
                                                 94

-------
                                                      STATE PROGRAM DEVELOPMENT PRIORITIES & STRATEGIES
TROPHIC STATE j,!)^
INDEX VALUE
90

80

70
60
50
40
30
20

10


— — — LAKE NAME -

_ OVERPECK LAKE
	 LINCOLN PARK SPRING LAKES
— SUNSET LAME

- VERONA PARK WEEQUAHIC WOODBURV LAKES
- KIRKWOOD LAKE
MEMORIAL LAKE
ECHO LAKE
- TOPANEMUS LAKE
BETHEL CLOVE DEVOE LAKES
-

FIGURE 3
RELATIVE RANK OF 25 NEW JERSEY LAKES AS
BASED ON CARLSON'S (1977) TROPHIC STATE INDEX
— VALUES COMPUTED FROM MEASURED MEAN SUM
MER CHLOROPHYLL A CONCENTRATION'S

Figure 3.—Relative rank of 25 New Jersey lakes as based on
Carlson's (1977) trophic state index values computed from
measured  mean summer chlorophyll a concentrations.
their  presence in lake watersheds is  uncommon.
Rather, urban, suburban, and, in some instances, agri-
cultural stormwater runoff has been  found to con-
tribute the  major fraction  of nutrients  into New
Jersey's lakes.
  Of the various restoration options examined, a com-
bination of stormwater control, dredging, and, in some
situations, mechanical weed harvesting appears to of-
fer the most effective alternatives to achieve program
goals. In  isolated instances other restoration options
were recommended (i.e. bottom sealing, aeration, dilu-
tion flushing).
  To restore these waterbodies to a healthy eutrophic
state (plankton chlorophyll a equivalent of 6-10  mg
m-3), external nutrient loading must be reduced by 50
to 80 percent. In addition,  in many  cases internal
nutrient recycling must be decreased.
  Control  of pollutant  sources  is  the  preferable
restoration  alternative.  Since nonpoint  stormwater
has been identified as the major source of degrada-
tion in these 25 lakes, stormwater quality must be con-
trolled. Fortunately, New Jersey has several technical
and financial assistance programs that in the future
may provide help.
   In  1983, following the work of Whipple (1981)  and
Wanielista et al. (1982),  New  Jersey enacted Storm-
water Management  Regulations  (NJAC  7:8).  These
regulations provide a model  for local municipalities
that wish to enact local stormwater quality control or-
dinances. They establish control requirements for new
developments that are designed to offset alterations
 in the hydrologic response of the watershed  and  in
stormwater  quality  from  the undeveloped  to  the
developed condition. The regulations contain provi-
 sions for 85 percent State funding of approved  pro-
jects.
  Specifically, the regulations require municipalities
to adopt ordinances that enjoin developers to manage
and control stormwater runoff from their sites so that
they generate no greater peak  flows  as a result of
development  under conditions of the 2-year, 10-year,
and 100-year  frequency storms. In addition, they re-
quire the sedimentation of the particulate pollutants
found in  urban runoff.  Thus,  both flood  control and
stormwater quality improvements are emphasized.
This approach of dual  purpose  stormwater manage-
ment is a relatively new concept and the New Jersey
program  is one of the first statewide programs in the
Nation. The  Stormwater  Management  program  is
recommended by the New Jersey Statewide Water
Supply Master Plan as an essential component of a
comprehensive watershed and aquifer protection pro-
gram.
  The  Division of Water Resources and the Soil Con-
servation Service are conducting a pilot study whose
outputs will include a stormwater  management plan-
ning guide that will  detail  planning approaches,
management  techniques,   and  cost  efficiency
analyses of various control measures used during the
study. The guide is being developed to facilitate future
work  in  this  area by  counties and  municipalities
throughout the State.
  Retrofitting existing systems will also be necessary
to achieve water quality objectives in most developed
watersheds. A pilot program to examine the effective-
ness of an innovative underground  system that can be
incorporated  in  existing drainage systems is being
 partially funded through the Lakes Management Pro-
 gram.This  system would require minimal area for in-
 stallation compared to conventional surface retention
 basins and would improve the  quality of stormwater
 runoff from existing  systems.
   Coupled with stormwater management  controls,
 dredging  of  these  small  impoundments  will  be
 necessary to remove  accumulated  sediments and
 associated nutrients. This would effectively  deepen
 these  lakes and reduce internal recycling of nutrients.
   To implement dredging, sediment quality  is being
 closely examined prior to initiating restoration pro-
 jects.  The ubiquitous nature  of sediment contamina-
 tion found in this study necessitates that the suita-
 bility of dredged materials for disposal be determined
 for almost every proposed dredging project.
   Finally,  New Jersey will  continue to draw  upon
 State and local initiative in funding or partially funding
 lake  restoration  programs.  With  the  reduction in
 Federal  funding, alternative  sources  must  be found
 and local initiatives encouraged. At the State level, the
 New Jersey  Green Acres Program,  a  State bond pro-
 gram, will continue  to assist local governments ac-
 quire  and  develop recreation areas.  A total  of $540
 million has been spent during the last 10 years, mat-
 ching  an additional $355.4 million  in Federal and local
 funds.
   This program has specially emphasized the acquisi-
 tion and development of urban parks and has provided
 State  matching funds for two U.S.  EPA Clean Lakes
 Program Phase II grants  in urban areas. One of the
 most  successful and publicly supported programs, it
 will continue,  probably through a combination of
 grants and low interest loans, to be a reliable source
 of  lake management funding.
   From the outset, the goal of the Lakes Management
 Program was not to convert individual lakes into oligo-
 trophic or even  mesotrophic waterbodies. Because of
 lake morphometry and watershed  characteristics, this
 would be  impossible  to achieve. Instead, improving
 and maintaining lakes in the healthy eutrophic range,
                                                  95

-------
LAKE AND RESERVOIR MANAGEMENT


which, in most instances, will fully satisfy local useir
requirements, has been emphasized. This objective
will continue to be a major criteria in developing our
priority  list  for implementing future restoration  pro-
jects.


REFERENCES

Carlson, R.E.  1977.  A trophic state index  for lakes. Limnol
  Oceanogr. 22(2):361-69.

Dillon,  P.J. 1974. Application of the  phosphorus loading
  concept to eutrophication research.  NRC Tech. Rep.
Dillon,  P.J., and F.H. Rigler.  1974. The phosphorus-chloro-
  phyll relationship in lakes. Limnol. Oceanogr. 19:767-73.
Kirchner, W.G., and  P.J. Dillon. 1975. An empirical method o:
  estimating the retention of phosphorus in  lakes. Water
  Resour.  Bull. 11:181-82.
Ostrofsky, M.L 1978. Modification of phosphorus retention
  models for use with lakes with low areal water loading. J
  Fish. Res. Board Can. 35(912):1532-36.

Smith, V.H., and J. Shapiro. 1980. A retrospective look at the
  effect  of phosphorus  removal  in lakes. Pages 73-77  in
  Restoration of Lakes and Inland Waters. EPA 440/5-81-010.
  U.S. Environ. Prot. Agency, Washington,  D.C.

U.S.  Environmental Protection Agency. 1980. Clean Lakes
  Progam Guidance Manual. EPA 440/5-81-003. Washington
  D.C.

Wanielista, M.P., et al. 1982. Stormwater management to im-
  prove lake water quality.  EPA 600/2-82-048. U.S. Environ.
  Prot. Agency, Washington, D.C.
Whipple, W. Jr. 1981. Dual purpose detention basins in storm
  water management. Water Resour. Res. Bull. 17(4):642.
                                                     96

-------
INCOMPATIBILITY OF  COMMON  LAKE MANAGEMENT OBJECTIVES
KENNETH J. WAGNER
RAY T. OGLESBY
Department of Natural Resources
Cornell University
Ithaca,  New York


            ABSTRACT

            Lake management involves the formulation of objectives, not all of which are compatible in a given
            water body. Common objectives such as high quality water supply, suitability for contact recreation,
            and pleasing aesthetic properties are generally associated with low plankton standing stock, especially
            as regards phytoplankton On the other hand, production of fish for food or sportfishing is often im-
            paired by decreases in plankton biomass. Strong correlation has been found between measures of
            plankton production or standing stock and fish yield. Investigated relationships between phytoplankton,
            zooplankton, planktivores, and piscivores are consistent with an energy pyramid model of the pelagic
            food web;  greater production at the top of the pyramid is caused by a corresponding increase at its
            base or through increased internal energy transfer efficiency Management for higher energetic effi-
            ciency is in its infancy and that for greater plankton biomass conflicts with other management objec-
            tives Division of responsibility among various organizations  and agencies of government may help
            achieve single objectives, but impedes progress toward a unified systems approach to lake manage-
            ment. Management objectives may be less conflicting for poorly aerated systems or where fishery
            management is aimed at certain target species, but serious consideration should be given to poten-
            tial conflicts during goal formulation. Lake associations and managers should establish management
            priorities early in their planning processes Their decisions should use expertise from both fishery
            and water quality science.
INTRODUCTION

Effective  management  of  lakes for multiple  uses
under   budgetary  limitations  necessitates  that
management agencies scrutinize their objectives for
compatibility with each other and the aquatic system
under  consideration.  Common objectives include
water supply, contact recreation,  visual aesthetics,
and fishery optimization. While each lake is to some
extent a unique system, generalizations can be made
about desirable lake conditions with respect to each
objective. These generalizations facilitate a prelimin-
ary evaluation of the compatibility of common lake
management objectives.

GENERALIZATIONS

1. Increases in the standing stock of plankton in a
lake elevate the cost of treatment for water supply,
both where human consumption is involved  and in the
case of industrial process water.
   Removal of plankton from a water supply becomes
necessary when  plankton  standing stock  becomes
large enough to cause odor, taste, health, or aesthetic
problems or clogs the distribution  system. Costs rise
abruptly at each new level of treatment resulting from
capital  expenses  and  gradually  increase between
steps  as a function  of operational costs largely
associated  with  filtration  and chemical  additives.
Where water supply is concerned, maintenance of the
lowest possible plankton biomass is clearly desirable.
   2. Increases in the standing stock of plankton in a
lake decrease its aesthetic value  and water contact
recreational appeal to most people.
   Surface  scums or  dense  blooms of   algae are
generally considered  unattractive and in some cases
represent a health hazard. A  visible abundance of
planktonic  plants or animals may deter swimming.
Decreased water clarity reduces recreational  safety,
especially   where  subsurface  obstructions  are
prevalent. As with water supply, the lowest possible
plankton biomass is generally desirable.
  3. Increases in the standing stock of plankton in a
lake increase fish production in that system.
  This may seem  intuitively  less obvious  than  the
preceding statements,  but is  a well-substantiated
generalization. Theoretically, an increase in the food
resource base will  increase  the flow of energy to
higher trophic levels (Ryther, 1969; Kerr, 1974; Sheldon
et al. 1977). Empirically derived relationships (Ryder et
al. 1974; Oglesby, 1977; Jones  and Hoyer, 1981; Mills
and  Schiavone, 1982) indicate that increased phyto-
plankton standing  stock or production does lead to
greater fish production. Elevated fish production in
temperate lakes may result from additional  energy
flow through the benthic/detrital pathway (Golterman,
1975; Eggers et al. 1978; Hanson and Leggett, 1982) or
the pelagic  route (Wright,  1965; Schindler, 1972;  Mc-
Cauley and  Kalff, 1981).
  The relationship between phytoplankton biomass
and  fish yield varies considerably among lakes, but a
distinct relationship exists (Fig.  1). The variation is
partly explained by physical and chemical differences
among lakes (Ryder et  al. 1974),  but  may also be a
function  of differential energy transfer efficiency.
Food quality, availability, and productivity per unit
biomass may change as food quantity  increases. This
suggests that improved fish yield might be obtained
by  either  enlarging  the  food  resource  base  or
manipulating food characteristics at a fixed resource
level (Fig. 2).
  Management for maximum fish yield through  in-
creased food resource levels has long  been  practiced
in fish culture operations (Bennett, 1970), while max-
                                                   97

-------
  LAKE AND RESERVOIR MANAGEMENT
  imization  of energy  transfer  efficiency at  a fixed
  resource level is a science in its infancy, having rarely
  been tried and even then, with mixed results (Lange-
  land, 1981; Morgan et al.  1981; Webster, 1981). While
  limitations may exist and improving the system's  ef-
  ficiency remains an alternative, the greatest possible
  plankton biomass appears  desirable where fish pro-
  duction is to be maximized.
 INCOMPATIBILITY OF OBJECTIVES

 These generalizations suggest that there is a potential
 conflict  between  fishery optimization and the other
 management objectives (water supply, contact recrea-
 tion,  and aesthetics).  Stated  simply, maximum fish
 yield  is  obtained  at  large  plankton biomass  while
 lowest  treatment  costs,  greatest  visual appeal,
 highest  safety level,  and  lowest  health risk  are
 associated  with low plankton biomass (Fig. 3). With-
 out established priorities, maximized fish  production
 is incompatible with the other objectives.
    Chlorophyll a concentrations under 5 micrograms
 per liter, indicative of  low phytoplankton biomass,
 might be considered desirable for water supply, con-
 tact  recreation,   and  aesthetic  purposes.   Man/
 management agencies would  settle for levels less
 than 10 micrograms per liter (see Reckhow, 1979, for ,a
 comparison of commonly  used chlorophyll criteria).
                E
                t
                I
                2
                a
                   001 .
                                     10    20     50


                                     Chlorophyll (»g |-')
                                                    100   200
              Figure 1.—Regression  of  fish yield  on summer  phyto-
              plankton standing crop for 19 lakes; 95 percent confidence
              intervals for predicting values about the mean are within the
              dashed lines (modified from Oglesby, 1977).
                        Sunlight,
                        nutrients
                                       Original
                                       system
             Piscivorous fish


                 Planktivorous/benthivorous fish


              \    Zooplankton/benthos
                \
                                                                 Phytoplankton
Energy transfer efficiency at
phytoplankton—zooplankton/
benthos interface increased,
yielding increased energy at
higher levels without increasing
base
                    /
  Piscivorous fish
                                      Planktivorous/benthivorous fish
Zooplankton/benthos
                                       \     Phytoplaikton  /

                                       Energy at all levels increased due
                                       to increase in base size at constant
                                       energy transfer efficiency
\\
                                             \
Figure 2.—Schematic representation of potential manipulations of energy in an aquatic system. Energy represented by each
compartment is derived from the compartment below. Dashed lines represent the original system.
                                                    98

-------
                                                      STATE PROGRAM DEVELOPMENT: PRIORITIES & STRATEGIES
Figure 1 does not suggest any chlorophyll a concen-
tration above  which fish yield  levels  off,  with in-
creases in fish yield  continuing in the 100 to 200
microgram per liter range. A conflict between target
chlorophyll levels is evident.
  If zooplankton  production must  be maximized to
achieve the highest possible fish yield,  chlorophyll a
levels will  probably have to average well above the
desired 5  to 10 microgram per liter level. In Oneida
Lake, a eutrophic body of water where the zooplankter
Daphnia greatly influences fish production, maximum
Daphnia reproduction  is rarely attained (Wagner,
1983). A natural summer cyanophyte assemblage at a
chlorophyll a level of 50 to 70 micrograms per liter and
a natural spring diatom assemblage at 15 to 20 micro-
grams per liter produce the best results. Based on the
energy requirements of Daphnia species (Lampert and
Schober, 1980; Porter et al. 1983) and assuming that
all phytoplankton are available and of a suitable quali-
ty,  the minimum  chlorophyll  a concentration that
could sustain maximal Daphnia reproduction  is be-
tween 7 and 20 micrograms per liter. Natural availabili-
ty and quality limitations make this an underestimate,
again  suggesting  a  conflict  between  target
chlorophyll levels for different objectives.
 MANAGEMENT BY PRIORITY

 While conflicts between  objectives may be involved,
 multiple use management should not be discouraged.
 To make it effective a set of priorities on which to base
 management   decisions   and  strategy  must  be
 developed. The top  priority objective will  place con-
 straints on those of lesser priority, potentially limiting
 alternatives but clearly defining the realm of opera-
 tion. Setting priorities for objectives facilitates the
 restatement of  goals so that incompatibilities are
 minimized.
   Where  water  supply,  contact recreation,  or
 aesthetics is accorded the highest priority the manag-
 ing organization must realistically evaluate the fishery
                 5   10   20    50  100  200



              Mean Summer Chlorophyll a_ ( vg I'1)
 Figure  3.—Relationship between  lake  properties and
 management objectives.
potential. It  should  be remembered that fishery op-
timization and  maximum fish  production are not
necessarily synonomous. A good fishing experience
includes more than just catching fish. The  size, type,
and number of fish and the visual appeal of the fishing
grounds should be considered as well as other factors
not directly linked to the fish themselves. Anglers may
happily settle for fewer fish in return for greater
aesthetic  pleasure  or an  occasional trophy fish.
Fishery optimization may then take  the form of stock-
ing policy  and fishing  regulation, while  plankton
biomass is kept low to satisfy other objectives.
  There is also merit in  trying to remove  the social
stigma of planktonladen waters. An  intermediate level
of biological productivity does not significantly  im-
pede contact recreation and can be  quite appealing to
the educated eye; it  will also expand the range  of
fishery management alternatives.
  The limitations of  each  system  must  be kept  in
mind. Fish production in poorly aerated lakes may not
respond as  positively to an increase in plankton
biomass, and certain target  species may fail  to  in-
crease in number or  size  in response to elevated
plankton standing stock. A system assessment pro-
gram should always precede and accompany manage-
ment actions.
   Establishing a priority of  objectives is useful  in
other areas of  lake management. Conflicts between
water  supply  and  recreational  activities, usually
centering around access  to the water body or the tim-
ing and severity of drawdowns, may be resolved in this
manner. It is a highly workable system if priorities can
be established.
 ESTABLISHING PRIORITIES

 Setting lake management  priorities  can be difficult
 when  competing interest  groups are  involved. An
 arena for resolving  conflicts and  reaching  com-
 promises becomes necessary. However, initial agree-
 ment on a set of priorities  may eliminate many later
 conflicts, making the effort  worthwhile. Obtaining this
 agreement is the key, the optimum end product being
 consensus.
   The situation is exacerbated by the fragmentation
 of management  or regulatory responsibility among
 various organizations and agencies of government. In
 the  early  1970's  Federal  involvement   in  water
 resources was spread over 11 agencies (N. Atlantic
 Reg. Study Group,  1972; Great Lakes  Basin Comm.
 1974), with little consolidation since that time. Since
 the 1960's most  States have had at  least four water
 resources related agencies, including a general water
 resources agency, a health  department, and a fish and
 game unit (Dworsky, 1964).  The result is  that as many
 as 14 organizations could have some responsibility for
 a given water body (Kennedy  and  Cook, 1980). While
 this  is  not  an inherently  bad  arrangement,  a
 cooperative relationship among these organizations
 and leadership on the part of  one  or more is needed.
 Such relationships appear to be relatively uncommon.
   Further complications arise  under multiple or public
 ownership of a  lake, a common  occurrence. Water
 bodies and watersheds rarely coincide  with political
 boundaries, frustrating many  management  attempts
 and  making  a  unified  systems  approach difficult
 (O'Riordan, 1966; Goetze, 1980).
   The establishment  of watershed  districts or lake
 associations can foster communication and coopera-
 tion among existing agencies  and  can create a forum
                                                  99

-------
 LAKE AND RESERVOIR MANAGEMENT
 for objective formulation. Such an organization is in-
 adequate,  however, without the priority setting pro-
 cess. Any organization  charged with this function
 should actively solicit expert input from professionals
 in fishery  science, water  quality engineering,  arid
 water supply facilities planning as well as polling af-
 fected  interest  groups.  Establishing management
 priorities should take  place early in the planning pro-
 cess, facilitating a concerted management effort and
 minimizing conflicts.


 CONCLUSION

 Current engineering principles, ecological theory, and
 empirical evidence indicate that fishery optimization
 can be incompatible with lake management objectives
 such  as  water  supply,  contact  recreation,   and
 aesthetics. Cooperation between numerous organiza-
 tions is  called for to minimize conflicts and maximize
 management  effectiveness. The  establishment  of
 management   priorities  incorporating  input   from
 technical experts and concerned interest  groups is
 recommended early in the planning process.
 REFERENCES

 Bennett, G.W. 1970. Management of Lakes and Ponds. Van
   Nostrand Reinhold Co., New York.
 Dworsky, LB. 1964. Making the State an effective partner in
   water development. Address to the III. Water Resour. Dev.
   Cornell Water Resour. Center,  Ithaca, N.Y.
 Eggers, D.M. et al. 1978. The Lake Washington ecosystem:
   The perspective from the fish community production and
   foraging base. J. Fish. Res. Board Can. 35:1553-71.
 Goetze, D. 1980. A strategy for empirical evaluation of river
   basin institutions. Pages 438-58 in North, L Dworsky, and
   D. Allee, eds. Unified River Basin Management. Am. Water
   Resour. Ass., Minneapolis, Minn.
 Golterman,  H.T.  1975.  Physiological  Limnology.  Elsevier
   Scientific Publishing Co., New York.
 Great Lakes Basin Commission.  1974. Great Lakes Basil
   Framework Study. App. F20. Federal laws, policies and  in-
   stitutional arrangements. Ann Arbor, Mich.
 Hanson, J.M., and W.C. Leggett. 1982. Empirical prediction of
   fish biomass and yield. Can. J.  Fish. Aquat. Sci. 39:257-6o.
Jones, J.R., and M.V. Hoyer. 1981.Sportfish harvest predicted
   as a function of summer algal standing crop in midwest
   lakes. Mo. Agric. Exp. Sta. J. Ser. 8939.
Kennedy, R.D., and J.R. Cook.  1980.  Managing water  re-
   sources through natural resource districts. Pages 248-57
  in North, Dworsky and Allee, eds.  Unified River Basin
   Management. Am. Water Resour. Ass., Minneapolis, Minn.
Kerr, S.R. 1974. Theory of size distribution in ecological com-
  munities. J. Fish. Res. Board Can. 31:1859-62.
 Lampert, W., and U. Schober. 1980. The importance of thresh-
   o'd food concentrations. Pages 264-67 in W.C. Kerfoot, ed.
   Evolution and Ecology of Zoopfankton Communities. Univ.
   Press New England, Hanover, N.H.

 Langeland,  A. 1981. Decreased zooplankton density in two
   Norwegian lakes caused by predation of recently introduc-
   ed Mysis relicta. Int. Ver. Theor. Angew.  Limnol. Verh
   21:926-37.

 McCauley, E.,  and  J. Kalff 1981.  Empirical relationships
   between  phytoplankton  and  zooplankton  biomass  in
   lakes. Can. J. Fish. Aquat. Sci. 38:458-63.
 Mills, E.L, and A. Schiavone. 1982.  Evaluation of fish com-
   munities through assessment of zooplankton populations
   and measures of lake productivity. N. Am. J. Fish. Manaae
   2:14-27.                                            '

 Morgan, M.D., C.R. Goldman, and R.C. Richards. 1981. Im-
   pact of introduced populations of Mysis relicta on zoo-
   plankton in  oligotrophic subalpine lakes. Int. Ver. Theor
   Angew. Limnol. Verh. 21:339-45.

 North Atlantic Regional Study  Group.  1972. North  Atlantic
   Regional  Water Resources  Study.  App.  S.  Legal  and
   institutional environment. N. Atlantic Reg. Water Resour.
   Study Coor. Comm., U.S. Army  Corps Eng., New York.
 Oglesby, R.T. 1977. Relationships of fish yield to lake phyto-
   plankton standing crops,  production and morphoedaphic
   factors. J. Fish. Res. Board Can. 34:2271-79.
 O'Riordan, T. 1966. Small River Basin  Development in Eastern
   North America. Cornell Water Resour. Center, Ithaca, N.Y.
 Porter, K.G., J.D. Orcutt, and J. Gerritsen. 1983. Functional
   response and fitness in a  generalist filter feeder, Daphnia
   magna (Cladocera: Crustacea).  Ecology 64:735-42.
 Rechkow,  K.H. 1979. Quantitative Techniques for the Assess-
   ment of Lake Quality. EPA-440/5-79-015.  U.S. Environ. Prot.
   Agency, Washington, D.C.

 Ryder, R.A., S.R. Kerr, K.H. Loftus,  and H.A. Regier. 1974. The
   morphoedaphic index, a fish yield estimator—review and
   evaluation. J. Fish. Res. Board Can. 31:663-88.
 Ryther, J.H. 1969. Photosynthesis  and fish production in the
   sea. Science 166:72-76.

 Schindler,  D.W. 1972. Production of phytoplankton and zoo-
   plankton in Canadian Shield lakes. Pages 311-332 in Z. Ka-
   jak and  A. Hillbricht-llkowska, eds. Productivity Problems
   of Freshwater. PWN  Pol. Sci.  Publ., Warsaw, Poland.
 Sheldon,  R.W.,  W.H. Sutcliffe,  and  M.A.  Paranjape. 1977.
   Structure of  pelagic food chain and relationship between
   plankton and fish production. J.  Fish.  Res.  Board Can
   34:2344-53.

Wagner, K.J. 1983. The  impact of  natural phytoplankton
   assemblages on Daphnia pulex reproduction in  Oneida
   Lake, N.Y. M.S. Thesis, Cornell Univ.,  Ithaca, N.Y.
Webster, D.A. 1981. The scuttering sedges of Deerfly Pond.
   Rod and Reel, August 1981.

Wright, J.C. 1965. The population dynamics and production
  of Daphnia  in Canyon Ferry Reservoir, Mont.  Limnol.
  Oceanogr. 10:583-90.
                                                     100

-------
THE HISTORY OF THE  CLEAN LAKES PROGRAM
IN  TENNESSEE
FRED VAN ATTA
GREG DENTON
Division of Water Management
Tennessee  Department  of  Health
Environment
Nashville, Tennessee
and
            ABSTRACT

            The Clean Lakes staff of the Tennessee Division of Water Management have undertaken three separate
            projects since the inception of the Federal 314 program. The first project was the creation of a Trophic
            State Index and Priority Ranking System for the publicly owned lakes of Tennessee. This project was
            completed in 1979. During this inventory of public lakes, attention was drawn to Acorn Lake, a highly
            eutrophic, highly  used public lake located in Montgomery Bell State Park. In 1980, it was decided
            that the second Clean Lakes project would be a diagnostic/feasibility study of Acorn Lake. The lake
            was monitored for 23 months in an attempt to discover the source or sources of the nutrients causing
            the eutrophic conditions at the lake. The Acorn Lake study revealed that a variety of problems within
            the watershed had created a nutrient sink  in the sediments of the lake. Watershed management and
            lake drawdown were recommended as restoration techniques. While researching this project, the Clean
            Lakes staff discovered that a significant number of similar public lakes were being routinely artificially
            fertilized for a variety of reasons. In response to this problem, the Acorn Lake Report contained not
            only recommendations for Acorn  Lake, but also a new Division of Water Management policy against
            the application of artificial fertilizers in lakes.  The third Clean Lakes project, begun in 1982, is a
            diagnostic/feasibility study with major emphasis on sedimentation and macrophytes of the Upper  Buck
            Basin  of Reelfoot Lake. The multiple problems at Reelfoot Lake of siltation, excess macrophyte en-
            croachment, heavy metal and pesticide pollution, and eutrophication have been highly documented,
            but no one has yet been able to accurately predict the rate of sedimentation in the lake, although
            it is known to be filling rapidly. The Clean Lakes staff, assisting Dr. J. Roger McHenry of the Agricultural
            Research Service, U.S. Department of Agriculture, will use cesium-137 dating techniques in an at-
            tempt  to determine this rate.
 INTRODUCTION

 Section 314 of the Federal Water Pollution Control Act
 Amendments of 1972 (later revised as the Clean Water
 Act  of 1977) directed the U.S. Environmental  Protec-
 tion Agency to assist the States in controlling sources
 of pollution that affect the quality of freshwater lakes
 and in restoring lakes that had deteriorated in quality.
 In fulfillment of this mandate, EPA created the Clean
 Lakes Program  to  provide technical and  financial
 assistance to the States to:
   1. Classify publicly owned freshwater lakes accor-
 ding to trophic condition.
   2. Conduct diagnostic studies of specific publicly
 owned  lakes, and develop feasible pollution  control
 and restoration programs for them.
 TROPHIC STATE INDEX

 In 1978, the then Division of Water Quality Control be-
 gan work on a ranking system for the public lakes of
 Tennessee based on trophic level as well as other fac-
 tors such as use classification, level and  potential
 growth of public utilization, and other sources of pol-
 lution.
   The publicly-owned lakes of Tennessee were placed
 into three groups. Class I  lakes were  those publicly
 owned freshwater lakes within  Tennessee  which  (1)
 were under the management authority of a State or
 sub-State agency; (2) had  a surface area of 2.0 hec-
 tares or greater; (3) were determined to have substan-
 tial public interest and use; (4) if restored, would pro-
             vide a cost-effective increase in public health; and (5)
             offer public access.
               Class II lakes were those publicly owned freshwater
             lakes within Tennessee which (1) were under the man-
             agement authority of a Federal agency; (2) had a sur-
             face area of  2.0 hectares or greater; (3) were deter-
             mined to have substantial public interest and use; (4)
             would provide a cost-effective increase in public bene-
             fit; and (5) offer public access.
               Class III lakes were those lakes which did not meet
             the criteria for a Class I or Class II lake and,  therefore,
             were not identified or classified for the purpose of the
             survey.
               Trophic classification of the lakes included in the
             Tennessee Study was based on  a method developed
             by Robert E.  Carlson at the  University of Minnesota.
             His system used numerical values rather than descrip-
             tive terminology. In the study of Tennessee lakes, his
             system has proven to be preferable to the more gener-
             alized approach  of  using the descriptive terms of
             oligotrophic,  mesotrophic, and eutrophic states.
               The Carlson system develops a Trophic State Index
             (TSI) for  any  one of three parameters:  chlorophyll a,
             Secchi disk, or total phosphorus. All three parameters
             were used and a TSI was generated for each. Chloro-
             phyll a directly measures the algal biomass, while the
             Secchi  disk  reading measures  the light  absorbing
             characteristics of the  water body. Values obtained
             from Secchi disk reading assume that all of the light in
             the water column is related to the biomass. The total
             phosporus measurement assumes that the water
                                                  101

-------
LAKE AND RESERVOIR MANAGEMENT
body is phosphorus limited. However, in the study of
Tennessee's lakes primary emphasis was placed on
chlorophyll a, the measurement the public most readi-
ly  identifies  with  eutrophication,  namely,  algal
blooms, which frequently produce unsightly  scums
and their accompanying problems.
  Based upon the three trophic state indices, each
lake was numerically classified according to  its tro-
phic condition. The TSI value for chlorophyll  a  was
most generally used for such classification because
of  its  relation to  lake productivity. In those cases
where   chlorophyll  a   data  were  not  available,
classification was  based upon the Secchi disk TSI or
the total phosphorus TSI, whichever was less. This
was deemed reasonable because neither the  Secchi
disk TSI nor the total phosphorus TSI was expected to
be  significantly lower than the chlorophyll a TSI. The
Secchi disk TSI could be significantly higher in cases
where  light is absorbed by material other than algae;
and, the total  phosphorus TSI could be significantly
higher  in cases where phosphorus is not  limiting  or
available to the algae.
  Secchi disk  readings and samples for chlorophyll a
and total phosphorus  analyses were taken  for all
Class I lakes and those Class II lakes not managed by
the Corps of  Engineers  or the Tennessee Valley Au-
thority from July 15,1979, through Sept. 15,1979. This
period   was determined to  be the time  of peak
seasonal productivity. The samples were drawn at a
depth of one-half meter below the surface at the ap-
proximate center of the  lake. Where the lake had a
morphology of more than one finger or a surface area
of 81 hectares or  more, an alternative site or sites
were chosen.
  For Class II  lakes under the management authority
of the U.S. Army Corps of Engineers or the Tennessee
Valley  Authority, the trophic state indices were gener-
ated from data furnished by the appropriate Federal
agency.
  The results of this study were compiled into a report
entitled, Survey of Publicly Owned Lakes and Reser-
voirs, completed in July 1980. Within the report, a tro-
        Table 1.—Top five Class I and II lakes.

                     Class I
Priority
No.
1
2
3
4
5
Name
Reel foot
Acorn
Byrd
Radnor
Meadow Park
County
Obion, Lake
Dickson
Cumberland
Davidson
Cumberland
Priority
Points
5.740
3.669
3.630
3.473
3.441
                    Class I
Priority
No.
1
2
3
4
5
Name
Chickamauga
Cherokee
Cheatham
Fort Loudon
Douglas
County
Brakley, Hamilton,
Meigs, Rhea
Grainger, Hamblen,
Hawkins,
Jefferson
Cheatham,
Davidson
Blount, Knox,
Loudon
Cocke, Jefferson,
Sevier
Priotiry
Points
6.133
5.722
5.209
5.039
4.974
 phic state index ranked all surveyed lakes. The top five
 Class I and II lakes are listed in Table 1.
THE ACORN LAKE STUDY

In July 1980, the Division of Water Quality Control was
awarded a Clean  Lakes grant to study Acorn Lake, a
9.7 hectare lake in Montgomery Bell State Park. The
objectives were to (1) establish current  limnological
data, with an emphasis on discovering the sources of
nutrients creating the high trophic levels discovered at
the lake in 1979; and (2) develop a restoration plan.
   Ever  since its  opening in 1942, Montgomery Bell
State Park has been one of the most popular recrea-
tional facilities in  Tennessee, no doubt because of the
presence of Acorn Lake, a small manmade lake that is
the  focal  point for many park activities. However,
since the  early 1960's visitor use of Acorn Lake has
declined dramatically even though the total number of
park visitors has increased during that same period.
   The most drastic decline occurred in the number of
recreational swimmers using Acorn Lake each year.
There is little doubt that a deterioration of water quali-
ty was a major cause for this decline.
   Sampling for the project began in August 1980. Sta-
tion selection  was predicated  upon the desire to lo-
cate stations where they would  best represent the lim-
nological properties of the lake. Station Acorn Lake L1
was an  in-lake station located at the deepest point of
the lake. Acorn L2, L3, and L4 were located in an area
which would best describe the impact of the seven
tributaries  which  feed the lake. Monitoring stations
were established  at each tributary entering the lake
(Acorn 12 through Acorn 18). A monitoring station was
located  at the outflow just south of the earthen dam
(Acorn 01). A station was also established at the sew-
age treatment plant (Acorn STP) located on the bank
of the lake along the southern finger which is directly
fed by the tributary inflow designated as Acorn 18. To
gain insight into the quality of the ground water of the
watershed, a station was also located at a well (Acorn
W1) to which access was permitted (see Fig. 1).
   Sampling frequency, when possible, followed these
guidelines: Sampling was done at least monthly from
September through April and bi-weekly from May
through August.  This  provided intensive sampling
coverage during periods  of high biological activity.
Sampling times were confined to the daylight hours
from 0800 to 1600. In-lake samples were taken at one-
half meter  below  the surface, one-half meter off the
bottom,  and then at every one and one-half meter
interval  through the water column. Water transparen-
cy was  measured by  Secchi disk. Limiting nutrient
was determined to  be phosphorus by AGPT  tests
conducted by EPA-Athens Laboratories.
  Sampling continued for 23 months and was com-
piled into  a report entitled, The Acorn Lake Report,
completed in March 1983. The report contained a hy-
drologic and nutrient budget. Table 2 lists sources of
phosphorus revealed by the study.
  The monitoring  of the outflow station revealed that
78 percent of the phosphorus was retained in the lake
in an average year.

  The report concluded:

  It is the finding of  this study that the limiting nutrient at
  Acorn Lake is phosphorus and that the major source of
  this nutrient in any given year is internal loading from
  the sediments during anaerobic  conditions. The sedi-
  ments have absorbed a constant inflow of this nutrient
                                               ;102

-------
                                                         STATE PROGRAM DEVELOPMENT. PRIORITIES & STRATEGIES
  from several sources as shown by the nutrient budget.
  The end result of this condition has been the creation of
  a nutrient "sink". This is a situation in which the sedi-
  ments absorb excess phosphorus and then release this
  nutrient in large amounts during times of low dissolved
  oxygen caused  by the increase of BOD. Due to the ther-
  mocline, very little, if any, of the released phosphorus
  mixes during the year of release. However, it is available
  for the next year's growth after the lake overturns.

  Another source of nutrients and pollution to Acorn
Lake  was found to  be the park's sewage treatment
plant  which appeared to have a history of overload
and  malfunction. On  Dec.  16, 1980, a  break was
discovered in the pipe crossing under the tributary ap-
proximately 20  yards upstream of the sampling sta-
tion, designated in the Acorn  Lake study as ACORN
18. This pipe originates from the small sewage treat-
ment  plant  on  the  south  side of the lake near the
restaurant and inn. Effluent  from the treatment  plant
is carried  along  the  bank of the lake where it crosses
under the tributary and then  is piped out of the water-
shed. The break in this effluent pipe was discovered
during the normal monitoring run for  that particular
month. The  entire flow from the treatment plant was
entering the tributary at the  point of breakage.
  This condition was brought to the attention of the
park officials so that corrective action could be taken.
Monitoring for the Acorn Clean Lakes project began in
August 1980. Not until December 1980 did the break
become of such magnitude that Clean Lakes person-
nel  could locate  and identify the problem. It is not
known  how long  the pipe had been leaking smaller
quantities.
  In studying Acorn Lake and other similar publicly
owned lakes across Tennessee, it became apparent to
the Clean  Lakes staff that  a significant number of
these lakes were being routinely artificially fertilized
for a variety of reasons. In response to this problem,
the Acorn  Lake Report contained a section entitled,
"On  the Artificial  Fertilization of  Publicly  Owned
Lakes." The following is an excerpt from that section:

  Eutrophication is a natural process that will proceed to-
  wards the eventual extinction of any lake. Nevertheless,
  it is the,mandate of the Clean Lakes program to  make
  recommendations  based on the desire to prevent the
  premature demise  of public lakes through improper or
  ill-advised management practices.

  It is the belief of the Division of Water Quality Control
  that the artificial fertilization of  public lakes is a poor
  management practice which is in clear conflict with the
  intentions of the Clean Water Act of 1977 and the Anti-
  degradation Statement of the General Water Quality Cri-
  teria of the State of Tennessee. It is a practice which is

     Table 2.—Sources of phosphorus to Acorn Lake.
Inputs


A. Precipitation (direct)
B. Inflows
C. Net groundwater plus surface runoff
D. Geese
E. Septic tanks
F. Sediments
    from hypolimnion
    from epilimnion and metalimnion
        Total inputs
   Loading
(Ib/yr.)      (%)
            .6
   8
 240
  95
  32
   5

 831
  52
1,263
          19.0
           7.5
           2.5
            .4

          65.8
           4.1
         100.0
Factors were used to determine the value of A, C, D, E, and F
                                                                                        In-lake  stations

                                                                                          L-l -  L-4

                                                                                        Inlet stations

                                                                                          1-2 -  I-

                                                                                        Outlet station

                                                                                            0-1
                                                                                        1" * 280 feet
Figure 1.—Acorn Lake stations.
                                                    103

-------
 LAKE AND RESERVOIR MANAGEMENT
   believed by some to promote fish production; yet low
   dissolved oxygen levels, caused by eutrophic condi-
   tions, are one of the major causes of fish kills. This prac-
   tice severely damages a lake's potential for other uses.
   especially water contact recreation. Thus, it trades the
   short-term benefit of some for the long-term degrada-
   tion of water quality for all, at a time when long-term
   water resource management is desperately needed.
       In conclusion, The Acorn Lake Report recommen-
     ded lake drawdown, watershed management, and re-
     placement of the sewage treatment plant as prelimi-
     nary  restoration techniques. Dredging of the sedi-
     ments was also mentioned as a possibility. The Ten-
     nessee Department of Conservation is actively seek-
     ing funds to replace the treatment plant.
                                                                            RMlfoet
                                                                            Notional
                                                                            Wildlife
                                                                            Rcfug*
*£''•-, v>> *:f5SriS
&--':'-;'^l:5:ilf
                                               SE*-T.--;-X- -)h-
    Samburg
Figure 2.—Upper Buck Basin - Reelfoot Lake
                                  Cypr»t« Tr««t

                            	Transect

                              X  Sampling Point
                                  (approx. location)
                                                104

-------
                                                      STATE PROGRAM DEVELOPMENT: PRIORITIES & STRATEGIES
REELFOOT LAKE STUDY

In October 1982, work began on the third Clean Lakes
project, a sedimentation and macrophyte study on the
Upper Buck Basin of Reelfoot Lake. In addition to U.S.
Environmental Protection Agency funds, the Soil Con-
servation Service and the Tennessee  Department of
Conservation also made monies available for this pro-
ject.
  Reelfoot Lake is one of the most valuable water re-
sources of the State of Tennessee and of the South-
east.  Located in the  Great  Mississippi  flyway, the
Lake  and its environs  are used  by hundreds  of
thousands of migratory birds; it is the winter home of
an endangered species, the American Bald Eagle. The
Reelfoot  Lake area contains the  Reelfoot  National
Wildlife Management Area. In addition, Reelfoot Lake
State Park provides fishing and camping recreational
facilities  to  thousands of regional and  out-of-state
visitors each year. Possessing  what is known as the
most diverse aquatic habitat in Tennessee,  the lake
and adjacent area are the outdoor classroom for multi-
ple scientific studies.
  Groups of interrelated  factors have combined to
seriously affect water quality in Reelfoot Lake. The in-
creased rate of sedimentation has been the most pub-
licized problem, but in recent years the increased rate
of pesticide  accumulation and  nuisance aquatic ma-
crophyte encroachment have also generated  concern.
  Reelfoot Lake has one of the most colorful  histories
of any area of Tennessee. It is thought to have been
formed  by  a  series of earthquakes that  centered
around New Madrid, Mo., in 1811-12. The area was
sparsely  populated and there were no eyewitness  ac-
counts to the formation of the lake.
  Settlers slowly began  moving into  the area; the
most famous of these was Davy Crockett. Most relied
upon the lake as a source of food or employment.
  An often violent struggle for control of the lake  be-
gan around 1899 and was not resolved until the State
of Tennessee assumed control in 1914. It was during
this period that the "Night Riders" were formed. This
group began as  a vigilante organization whose pur-
pose was to protect the interests of the hunters and
fishermen who drew their livelihood from the  lake, but
the Night Riders soon began to enforce other areas of
morality.
  It was recognized very early in the lake's history that
the lake was highly eutrophic and silting in very rapid-
ly. As the lake became more and  more shallow, the
growth of aquatic macrophytes accelerated  and they
soon became a hindrance to boat travel. Many propo-
sals were forwarded to remedy these  problems, but
few plans were implemented.
  When the price and demand for soybeans increased
around 1960,  upland  areas to the east  of  the lake
which in the past had not been tilled were soon being
intensively  farmed.  The  agricultural  practices
employed increased the rate of erosion and severely
affected the lake. Although the lake is known to be fill-
ing in  rapidly, no  one has  been able to precisely
calculate the rate of this sedimentation. One of the ob-
jectives of the current Reelfoot Lake-Clean Lakes pro-
ject is to supply this missing information and explore
the relationship between two of Reelfoot Lake's more
serious problems—sedimentation  and   nuisance
aquatic plants.
  The  first  phase of the  project, the sedimentation
study, was undertaken May 2-12, 1983.  This study us-
ed techniques and personnel from the U.S. Depart-
ment of Agriculture's  Water Quality and Watershed
Research Laboratory in Durant, Okla. The techniques
were developed by Dr. J. Roger McHenry. The field
work on Reelfoot Lake was directed by Dr. Sherwood
Mclntyre. He  was assisted by staff members  of the
Division of Water  Management and the Division of
Laboratory Services.
  Sediment cores were taken at 33 sites throughout
the basin. The cores obtained from each sampling site
were measured, sectioned, and bagged for protection
during  transport to the lab in Oklahoma (see Fig. 2).
  In  Oklahoma, the amounts of radioactive Cesium-
137 will be measured, which will allow researchers to
date the sediment layers within each core. Cesium-137
is an isotope which does not occur naturally. Its first
appearance in the  sediments resulted  from radioac-
tive  fallout  associated  with  atmospheric  atomic
testing which occurred in the 1950's and early 1960's.
Although found in very minute quantities, this isotope
has been successfully used to indicate the rate of sed-
imentation in  lakes since the atomic tests.
  Standard water tests also were run at each site and
four sediment cores were analyzed for  pesticides.
  The  second phase of the project, the macrophyte
study, was performed Aug. 15-27,1983. This time was
chosen as  the period  of  peak biological  production
and maximum macrophyte encroachment on the lake.
Aerial photographs were analyzed to locate the major
macrophyte beds, then a survey team actually deter-
mined  the  extent  and species composition of the
aquatic plant coverage on the lake. At points along
transects, rooted plants were raked up and  identified.
All information was plotted on an aerial photograph.
  If all work proceeds according to schedule, a report
of the  Reelfoot Lake-Clean  Lakes study  will  be  re-
leased in March 1984.
  The Clean Lakes Program has been a valuable tool
in the  Division of Water Management environmental
protection efforts.  The program has provided  assis-
tance at specific eutrophic lakes and has gathered in-
formation which has been and will continue to be valu-
able to lake management efforts across the State. The
State of Tennessee is  committed to the continued vi-
ability of the Clean Lakes program.
                                                105

-------
 BACKGROUND FOR MANAGEMENT OF TROPICAL
 RESERVOIRS IN  PUERTO RICO
LAURENCE J. TILLY
JORGE R. GARCIA
University  of Puerto Rico
Mayaguez, Puerto Rico
            ABSTRACT

            An EPA sponsored lake restoration feasibility study afforded the opportunity to examine the ap-
            plicability of conventional temperate zone indices and management approaches to tropical
            Puerto Rican lakes. Major problems in using temperate zone guidelines are anticipated in Puerto
            Rico because lake events are relatively aseasonal, nitrogen  rather than phosphorus may be
            limiting, heavy rainfall and steep topography result in naturally high loading, and mountainous
            terrain results in protection from wind driven mixing. Data for 26 lakes were examined; two of
            these lakes had been studied in detail. Trophic state indices using chlorophyll,  nitrate, total
            phosphorus, transparency, net productivity, and other parameters were calculated and used in
            lake ranking. Some comparisons with  Florida lakes suggest that Puerto Rican lakes receive
            more nutrients without developing objectionable blooms. We suggest a series of recommenda-
            tions for study and management of Puerto Rican lakes.
INTRODUCTION

In the process of performing a Clean Lakes diagnostic
study for Lake La  Plata  in  Puerto  Rico  we  have
become increasingly aware of difficulties in assigning
trophic   status  and  in  recommending  remedial
measures for such subtropical  lakes. As we developed
the baseline information  for La Plata we tried to place
it in the context of general U.S. Environmental Protec-
tion Agency and  limnological conventions, but en-
countered significant problems in doing so. The lakes
in  Puerto  Rico do  not  fit  the usual temperele
templates. This is true  for several reasons: (1) the
climate differs substantially, (2) most of the lakes aire
manmade, and (3) the public perception of the value of
lakes in  Puerto Rico differs from that in the United
States.  The climatic differences primarily  involve
higher and less variable  temperature, higher rainfall,
fairly persistent,  moderately strong  winds,  and
relatively constant solar insolation.  These climatDl-
ogical factors  in turn strongly interact with the  local
geomorphological features which in general include
high relief because of the  tendency of Puerto Rican
reservoirs to be located in the  mountains.
  The lakes  in question differ  from natural lakes ir  a
number  of  features  because they  are reservoirs.
Thornton et  al. (1980), for  example, describe the in-
crease in importance in reservoirs of hydrological in-
puts in mixing and nutrient  regimes and point out that
reservoirs have generally larger  drainage areas
relative to the volume.
  Possibly because of the greater accessibility of the
ocean the  Puerto Rican  public appears  to  place
greater relative value on providing abundant quality
drinking water  rather than on suitability for recreation
and wildlife. As noted by  Reckhow (1979)  there re-
mains a  strong element of subjectivity  in  erecting
trophic state criteria and critical loading limits. For
lakes built by man primarily for power or water supply
in mountainous and rainy subtropical climate condi-
tions norms are difficult to find or to judge. In this
paper we describe the nature of the data available for
Puerto Rican lakes,  draw comparisons among  lake
systems, attempt to classify lakes using standard and
other criteria and discuss a strategy for managing and
studying lakes in this region.
Characteristics of Puerto Rican Lakes

There are 27 lakes or reservoirs in Puerto Rico (Brown
et al. 1979). We found limnological information on all
of these (Table 1), but data for annual cycles on only
two  lakes. La  Plata and Loiza are the only lakes for
which the data represent a full year of regular sam-
pling. The others are averages or in many cases single
results extracted from the literature and as such may
be less representative.
  Table 2 contrasts the average Puerto Rican system
with the temperate reservoirs and lakes from the Na-
tional Eutrophication Survey compared by Thornton et
al. (1980). In general, by comparison with the average
temperate reservoir, Puerto Rican lakes are deeper,
smaller  in area,  and   more  heavily  loaded  with
nutrients. In other regards the lakes' features appear
unremarkable  on the  basis of the  data available
whether in comparison  with average lakes or with
reservoirs.
  More details are available for La Plata and Loiza
reservoirs (Table 2) since each was the object of a full
year's  study  (Garcia  and  Tilly,  1983;  Quinones-
Marquez, 1980). Unfortunately, Loiza was  studied in
1973-74 and La Plata in  1981-82; hence, data from the
two  lakes are  not strictly comparable. Nevertheless,
an examination of the  data  from these two studies
enables some generalizations to be made.
  Both lakes are located in the northeastern sector of
Puerto  Rico (Fig. 1). Both  have  extensive develop-
ments of  water hyacinths, Eichhornia crassipes, with
percent coverage greater in Loiza. La Plata lies in a
                                                106

-------
Table 1.—Limnological characteristics for Puerto Rican reservoirs.
Lake
Adjuntas
Caonillas
Carite
Carraizo
Cidra
Coamo
Comerio 1
Comerio 2
Dos Bocas
Garzas
Guajataca
Guayabal
Guayo
Guineo
Jordan
La Plata
Las Curias
Loco
Lucchetti
Matrullas
Patillas
Pellejas
Prieto
Toa Vaca
Toro
Vivi
Mean
± SE
Temp
°C

27.98
22.88
26.30
24.35



25.55

26.53




24.7


27.10
25.27
26.32


27.32


25.85
±.45
DO
g/m3

6.69
5.31
6.19
2.76



6.28
6.75
6.58

7.55


1.4


9.25
5.35
6.74


3.63


5.73
±.58
Cond.
umhos
/cm

1O1 .67
80.17
260
95



198.33

261




256


204.2
127.5
155


268.33


190.65
± 20.65
Alka.
mgl\
CaCO3

72.5
29.0
84.5
48.5



68.5

137.5




98


97
24.5
44.5


130.5


75.9
±11.5
PH
Std.
units

7.08
7.15
7.4
6.87



7.38
7.64
7.63

7.63


6.7


7.68
6.58
7.63


7.15


7.27
±.11
Secchi
M

1.04
1.60
.49
.68



1.18
1.74
2.40

2.15


1.6



.85
1.56





1.39
±.18
Turb.
FTU

14.0
5.6
27.0
40.0



7.65
1.61
7.8

4.5


19.8*


6.9
8.4
13.2


5.5


12.54
±2,98
Total
Phos.
mg/l
.25
.04
.02
.36
.02
1.10
.27
.09
.04
.04
.01
.07
0.18

0.2
.22
.07
0.4
.025
0
.026
.01
.01
.005
.01
.04
.11
±.04
N03-N
mg/m3
.22
.25
.25
.82
.13
3.70
.50

.20
.09
.14
.12
.32
.03
.01
.27
.85
.79
.23
.024
.43
.02
.27
.039
.44
1.03
.45
±.08
NH3
mg/1

.03
.07
.04
.53





.05




.09


.70
.05
.03


.10


.17
±.08
Net P
mg1/3
m/hr

101.5
86.7
283.3
213



74.3
29.3
47.3

77


212


195.3
147
111.3





131.5
±22.6
Chl-A
mg/3m

22.03
12
15.8
7



6.07
5.0
7.4

6.98


17.0


13

7.8





10.92
±1.65
Color
STD
Units

10.9
7.8
21.7
8.6



11.3
7.9
8.6

10.5


9.2*


5.3
5.0
10.6


6.1


9.5
±1.17
Pnyto
plankton
cells/ml

100.5
87.5
19.8
28.1



13.4
17.1


16.76






35.8
29.4





38.7
±10.8
Vol.
10«M3

60.4
13.9
14.9
0.65



39.5
22.9
40.2

18.6


30.85


20.3
4.4
21.2


38.2


25.08
±4.55
Inflow Mean
107M3 Depth
yr M

10.4 20.5
34.8
28.0 6.1
21.4 17



33.8
24.2 39.3
50.5




34.2 10.0



12.5
21.2





27.1 18.6
±3.76 ±6.8
Annual
Rainfall
cm

76.8
66.44
64.13
76.04



77.7

62.64




70.9



86.99
66.45





72.01
±2.66
Drainage Surface
Area Area
108M2 103M2

130.5 3371
20.5
538 2425
22.2



440.3 2318
16.1 583
63.7

24.9


450 3070



11.4
65.3





162.1 2353
±62.1 ±484.2















H
H
m
TJ
3
O
CD
ZD
S
D
m
m
5
TJ
m
^
3]
O
                                                                                                                                 m
                                                                                                                                 co
                                                                                                                                 3
                                                                                                                                 is
                                                                                                                                 m

                                                                                                                                 9
                                                                                                                                 m
                                                                                                                                 co

-------
 LAKE AND RESERVOIR MANAGEMENT
 basin of greater relief and Loiza is situated in an area
 with about 11/z times the population density found
 around La Plata. Loiza is larger in area, shallower and
 receives phosphorus and/or nitrogen loading similar
 to La Plata. The N:P ratios of loadings suggest that
 whereas  phosphorus  may   be  limiting  in  Loiza,
 nitrogen appears to limit La Plata. Loiza exhibits less
 evidence of thermal or chemical stratification than the
 deeper and more protected La Plata.
 Typical  Lake Year

 Based on inspection of the data from Loiza (Quinones-
 Marquez, 1980) and reflecting on our observations of
 La Plata, we constructed an hypothetical lake year. As
 pointed out by Garcia and Tilly (1983) for La Plata, the
 main factors influencing the  lake trajectory in Puerto
 Rico  are  hydrological  events  and  endogenous
 responses  to  nutrient  loading  superimposed en a
 weakly seasonal  climatological pattern  interacting
 with specific  features of lake morphometry. The tra-
 jectory of an hypothesized typical year is dictated by
 the relative probability of certain critical conditions
 through time. The year may be assumed to begin in Ju-
 ly with the  onset of hurricane season, high tempera-
 ture and high average rainfall conditions. As summer
 progresses the probability of the extreme rain event in-
 creases.  From October through February, tempera-
 tures fall and stabilities decrease as does the pro-
 bability of heavy rain events. With winds increasing
 during this  period the probability of mixing caused by
 either hydrological input or  wind remains constant.
 From  February through April stability remains  low
 while the probability of heavy rain diminishes and mix-
 ing if  it occurs is  more  likely caused by increased
 winter winds.
Figure 1.—Locations of La Plata reservoir and Lake Loiza,
Puerto Rico.
  From April through  June, both temperatures and
stability increase markedly. Meanwhile, although rain
and wind  probabilities both also increase, stratifica-
tion may  be fairly conspicuous and persistent.  It  is
during  this  period  especially  that  endogenous
biological events may dominate the  lake. Nutrients
enter the lake continually as a result of runoff events,
but radical  flushing doesn't occur since incoming
water tends to  be high in temperature (and  lower  in
density).   Incoming water  tends  to  remain in the
euphotic zone  due to  this density  difference.  Light
penetration remains fairly high in a large portion of the
lake (away from tributaries) and plant productivity may
be high. During periods of low nutrient input produc-
tivity  may be  sustained by  nutrient  reserves ac-
cumulated in plant tissue and possibly by advective
injection from hydrological inputs mixing the upper
layers of the hypolimnion into the epilimnion.
  Much of this description is either conjectural or
hypothetical, based as  it is on so little direct  observa-
tion. The  meteorological  trajectories, however, are
reasonably  representative. We  hypothesize   that
although the tendency exists  for these lakes to be
warm  monomictic  as  suggested   by  Lewis (1983),
hydrological events obscure any simple pattern of
mixing in these reservoirs.
RESULTS AND DISCUSSION

We  have computed  commonly  used  indices  to
evaluate their applicability in ranking Puerto Rican
lakes by trophic status. Table  3  (reproduced  from
Reckhow, 1981, after the EPA 1974  National Eutrophi-
cation Survey) shows critical  levels for chlorophyll a,
Secchi disk,  and total phosphorus selected by dif-
ferent authors and the associated ranking of the Puer-
to Rican lakes.
  According  to  the  EPA-NES index  for chlorophyll
three of the 11 lakes having chlorophyll data rank as
oligotrophic,   four  as mesotrophic,   and four  as
eutrophic. According to other chlorophyll indices none
is oligotrophic and from 5 to 11  lakes are eutrophic.
The EPA-NES phosphorus  index places the lakes 1:
8:17  oligotrophic:mesotrophic:eutrophic and  Secchi
disk  index ranks them 0:2:9 on the  same basis.
  Multivariate indices are regarded  as  potentially
more useful because they avoid  biases due to errors
inherent in any one variable  estimate (Reckhow, 1970).
Carlson's (1977) index applied to the 10 lakes for
which data were available clustered all of these in the
                                Table 2.—Comparison of lakes and reservoirs.1
309 Natural 107 Reservoirs PR Avg
Lakes Avg Avg Reservoir
Drainage area, km2
Surface area, km2
Maximum depth, m
Mean depth, m
Volume, 106m3
Hydraulic residence time, yr
Total phosphorus, g/m3
Chlorophyll a (mg/m3)
P loading (g/m2/yr)
N loading (g/m2/yr)
N:P ratio
222
5.6
10.7
4.5
25.2
0.74
0.054
14
0.87
18
47.1
3228
34.5
19.9
6.9
238.1
0.37
0.039
8.9
1.7
28
38.1
113
1.6
41.3
11.1
19.3
.33
.107
10.9
—
—
—
La Plata Reservoir2 Loiza Reservoir3
Puerto Rico
450
3.1
40
10
30.8
.07
.13
17.0
32.2
167
12:1
538
2.4
17.2
6.1
14.9
.05
.36
15.8
29
235
17:1
 1 Modified from Thornton et a/., 1980
 1 This study.
 3 Qumones-Marquez, 1980
                                                 108

-------
                                                        STATE PROGRAM DEVELOPMENT: PRIORITIES & STRATEGIES
40 to 70 point range, with two in the 40 to 50 range, six
in the 50 to 60 range, and two falling between 60 and
70. This placement is obviously above the mid scale in
lake condition.  Neither extreme end of the scale was
represented among the  lakes ranked.
   Walker's (1977) index, similar to Carlson's in being
based on TP, Chi a and Secchi disk, resulted in the
categorization of the same 10 lakes: 0 as oligotrophic,
1  as mesotrophic,  7  as eutrophic, and 2 as  hyper-
eutrophic. The only two lakes which could be ranked
by the TSI of Shannon  and Brezonik (1972) were La
Plata at 7.78 units and Loiza at 10.94. Both lakes rank-
ed well  into the eutrophic range according to this in-
dex. Although potentially applicable to some Puerto
                               Rican lakes on the basis of available data, the index of
                               Porcella et al. (1980) was not used because of its  ex-
                               plicit dependence on temperate lake cycles.
                                 A trophic state index for Puerto Rican lakes was
                               devised by Gomez and Gonzalez (1978) based on a
                               multiple linear regression  for net  primary productivity
                               (Net P) on NO3- N, total phosphorus, and chlorophyll
                               a for seven lakes. A problem existed in that only 9 of
                               the 27 lakes had  available chlorophyll a information.
                               The authors supplied estimates  by extrapolation  for
                               the remaining 18. The regression  was then applied to
                               the remaining lakes to estimate net productivity. The
                               collective set of estimated and measured net produc-
                               tivities were used to rank the lakes from high to low.
                                      Table 3.—Trophic state indices.

                        Exhibit 10a. Trophic state vs. chlorophyll a (from EPA-NES, 1974).
Trophic Condition
Sakamoto
f*
Chlorophyll a

 Academy      f
                                            Dobson
                                          f
EPA-NES
Oligotrophic
Mesotrophic
Eutrophic

0.3 - 2.5
1 -15
5-140
Totals
0
8
11
11
0-4
4 - 10
>10

0
6
5
11

4.3 - 8.8
>8.8

6
6
5
11
<7
7-12
>12

3
4
4
11
                      Exhibit 10b. EPA-NES trophic state delineation (from EPA-NES, 1974).
Trophic State
Oligotrophic
Mesotrophic
Eutrophic
Chlorophyll a
<7
7-12
Totals
f
3
4
4
11
Total Phosphorus
10-20
>20
f
1
8
17
26
Secchi Disk
Depth (m)
>3.7
2.0 - 3.7
<2.0
f
0
2
9
11
*f = Frequency of occurrence of each class Note that one lake may appear in more than one class in Sakomoto's index because of range overlap


                 Table 4.—Comparisons of estimated and measured net productivities,


Lake
Carite
Dos Bocas
Garzas
Guajataca
Guayo
Luchetti
Patillas

La Plata
Cidra
Loiza
Matrullas
Caonillas

Adjuntas
Coamo
Comerio #1
Toa Vaca
Guayabal
Jordon
Pellejas
Las Curias
Vivi
Prieto
Toro
Loco
Yahuecas
Measured
Mean Gomez & Gonzalez
Compiled Original Data
.172 .134
.131 .228
.068 .045
.094 .076
.154 .080
.390 .171
.222 .203

.424
.426
.566
.294
.204














Estimated
from 7 lake
NO3N + TP
Regression










.174
.102
.910
.017
.225

.941
3.590
.996
.144
.355
.108
.075
.326
.239
.040
.046
.230
.220
Estimated
from 12 lake
NO3N+TP
Regression
















.506
1.173
.499
.259
.292
.242
.228
.208
.202
.200
.180
.178
.150
Comments*


r = .500
compiled x = .176 ± .028
G&G mean = .133 ± .044
Not significantly correlated
but means not significantly
different
log transformed r = .710
significantly correlated
r = .69
compiled x = .383 ± .069
estimated mean = .286 ± .179
Not significantly correlated
means not significantly different
log transformed r = .43
r = .985
7 lake regression mean = .562 ± .278
12 lake regression mean = .332 ± .080
very significantly correlated
means not significantly different








'See text for further description
                                                  109

-------
LAKE AND RESERVOIR MANAGEMENT
  Using  their data  and excluding chlorophyll, we
found  a  significant  relation  between  Net  P and
N03- N  + TP which produced the same ranking they
reported.  The equation based on  Gomez and  Gon-
zalez's (1978) data was: Net  P  =  .07  -.152  (mg/l
N03-N) + 4.399 (mg/l TP). The correlation coefficient
was .80, significant with 5 degrees of freedom at p<
.05 and a standard error of the estimate of .052.
  Thinking to improve the strength of  this TSI we
assembled data from  Brown et al. (1979), Martinez
(1979), Quinones-Marquez (1980), Navarez and Villamil
(1981), Negron (1983) and Garcia and Tilly (1983), com-
bined them with the data for the original seven lakes
of  Gomez  and  Gonzalez (1978),  and calculated
averages for net productivity, N03 - N, and total phos-
phorus for 12 lakes. We developed a significant re-
gression:  Net P (g 02/m3/hr) = .218 - .115 (NO3 - N
mg/l) +  4.399 (TP mg/l). (The  correlation coefficient
was .678; P< .02; standard error of the estimate was
.128.)
  Using this regression to estimate net productivity
from the remaining 13 lakes for which we could find
nitrate and total phosphorus data we assembled the
ranking shown in Table 4. The two  sets of data cor-
related significantly (r = .98) but the "average" lake
was 52 percent lower in estimated productivity by our
index. The largest deviation in rank for any lake was
only three units among the 13 lakes for which net P
was being estimated. In the literature we found net
productivity values for five lakes not included in the
original regression. When we compared the  regres-
sion predicted values with the observed (Table 4) we
found a change in relative ranking for three of the five,
and  no significant correlation.
  It  is possible to compare a common set of 10 lakes
by the several indices described (Table 5). Overall, the
several   indices  do  agree in ranking the   lakes.
(Kendall's Coefficient of  Concordance, W  is  .59,
significantly different from zero at P< .01.) This is. not
too surprising given that phosphorus, chlorophyll, and
Secchi transparency are recurrent elements in this; set
of indices.
  The  fact  that such indices  agree  does  not
necessarily mean that the response to these variables
is similar in intensity in Puerto Rico. Even a cursory
examination  of  the available data for  Puerto Rico
leads to the conclusion that net productivities are very
high. Such productivities elsewhere are associated
with conditions of hypereutrophy in which "pea soup"
algal blooms are seen. In fact, we could not find one
 report of blooms with that extreme quality in Puerto
 Rico.
  To further examine  this apparent  disparity,  we
decided to make some comparison inter-regionally us-
ing ratios of net productivity, total  phosphorus, and
chlorophyll. Using the data set from  Shannon and
Brezonik (1972)  for  north  central Florida  and  that
available for Puerto  Rico we produced Table 6. We
found that mean net productivities in  the 12 Puerto
Rican lakes were almost five times those from the 24
lakes in Florida (28.18 ±6.74 versus  131.5 ±23.63).
Chlorophyll  values  averaged  8.85±1.54 mg/m3 in
Florida, not  significantly different from  10.92+ .174 in
Puerto  Rico. Total phosphorus  averages   were
.150+ .03 g/m3 and .068+ .03 for Florida and Puerto
Rican lakes, respectively.
  Even if one accepts the idea that a disproportionate
fraction of the problem lakes were included in the data
set for Puerto Rico one is impressed with the fact that
productivities  are higher for essentially equivalent
phosphorus and chlorophyll concentrations. Produc-
tivity  per unit chlorophyll was  12.5±2.2 for Puerto
Rican lakes compared to 2.8±.4 for the Florida set.
Per  unit  phosphorus,  net productivity averaged
3906 + 982 in the Puerto Rico set compared to 283 + 66
for  Florida. (Both differences were significant at P<
.05). The higher  average assimilation numbers (pro-
ductivity/unit chlorophyll) were reported by Carl and
Small (1965) to  be associated  with more  balanced
ratios of nitrogen and phosphorus. Smith concluded
that lower  chlorophyll  to phosphorus ratios  were
associated in part with low N:P ratios.
  We  provisionally  suggest that  overall  the dif-
ferences result from  more balanced nutrients, higher
average  temperatures,  and  greater  uniformity in
temperature, day length, and light intensity in Puerto
Rico. Under such conditions we suggest that metabo-
lism is substantially higher,  allowing  more   rapid
recycling of nutrients. Blooms do not occur because
nutrients are consumed rapidly all year and do not ac-
cumulate to the same degree as they do in the winter
and spring in the north  and then burst into crop in-
creases. Nor is there the substantial increase in day
length promoting photosynthesis in the presence of
such  nutrient  richness  as occurs in the temperate
regions.
  Lewis (1974)  suggested  that  tropical  lakes  have
higher productivities  because of a "greater equitabili-
ty" in distribution of nutrients through time along with
an efficient recycling and reprocessing of nutrients in
                              Table 5.—Comparison of ranking by various TSI.
Lake
Caonillas
La Plata
Loiza
Carite
Patillas
Guajataca
Cidra
Guayo
Dos Bocas
Garzas
Rank by
Chi a
1
2
3
4
5
6
7
8
g
10
Rank by
Secchi
3
7.5 (t)
1
7.5 (t)
5
10
2
6
4
9
Rank by
TP
4(t)'
2
1
7.5 (t)
6
10
7.5 (t)
9
4.0 (t)
4(t)
Rank by
Net P
5
3
1
6
4
9
2
7
8
10
Rank by
Walker's
3
2
1
6
8
10
5
9
4
7
Rank by
Carlson's
3
2
1
8
6
10
5
9
4
7
Sum of
Ranks
19
18.5
8
39
34.0
55
28.5
48
33
45
Kendall's coefficient of concordance, W = .59, significantly different from zero, (p< 01)
• (t = ties)
                                                 110

-------
 the  epilimnion.  In tropical  reservoir  systems,
 atelomixis  advective additions of nutrients  from
 tributary discharges (Garcia and Tilly,  1983) may ac-
 count for further  augmentation of nutrient supplies
 that maintain high productivities.
   More data are  needed about the relationship be-
 tween nutrient crop, turnover, and productivity in dif-
 ferent  classes of  tropical  lakes. Robarts  (1982)
 reported for Lake Mcllwaine,  Africa, average annual
 net  productivity  per  unit  chlorophyll of  about 8
 mg/C/hr/mg  Chi a, slightly  lower than the averages
 found for Puerto Rico and close to three times values
 obtained in Florida (Shannon  and Brezonik, 1972).
 Lake Mcllwaine, unlike Puerto Rican Reservoirs, has a
 more conspicuous  seasonality and experiences a
 distinct cool period. The average annual temperature
 is still  relatively high, but also intermediate between
 the Puerto Rican and Florida lakes compared. Tending
 to support the idea that nutrient distribution is more
 "equitable" is the finding that the average coefficient
 of variation for monthly TP values was  110 percent of
 Florida's Anderson-Cue Lake versus 43 percent in La
 Plata and 30 to 34 percent in Loiza in Puerto Rico.
 CONCLUSIONS AND RECOMMENDATIONS
 FOR PLANNING AND MANAGEMENT

 Puerto Rican lakes share certain features with other
 tropical  systems in  differing from  temperate lakes.
 The tropical  lake model of Lewis (1974) may be the
 general  background for function but superimposed
 upon this warm monomictic pattern is the relatively
 aperiodic influence of advective mixing.  Storm runoff
  STATE PROGRAM DEVELOPMENT: PRIORITIES & STRATEGIES

interacting  with  the  specific  morphometry  and
geology of each basin may result in mixing, including
complete turnover at any season. The aperiodic and
periodic mixing events dictate the trajectory of lake
function  primarily  by  restoring  oxygen  and
regenerating  nutrients.  Biological   processing  of
nutrients and organic matter occurs rapidly and pro-
ductivities  in  general  are  much  higher  than  in
temperate lakes.
  Management implications of these features include
the following points:
  1. Lakes in Puerto Rico may tolerate larger nutrient
loadings  without  developing  unacceptable blooms.
Their assimilative capacity for nutrient loading seems
substantially higher.
  2. This higher productivity may be able to be chan-
neled into top carnivore production if mixing sufficient
to prevent major anaerobiosis can be maintained.
  3. Full annual trajectories for at least temperature,
oxygen, and chlorophyll  need to be  developed for
more lakes in different regions of Puerto Rico. If possi-
ble,  such  measures should, in addition,  include
nitrogen and phosphorus concentrations in the lake
proper  and in major tributaries.
  4. The classification of lakes into oligotrophic,
mesotrophic, and  eutrophic categories is of  little use
in these systems. Richer and poorer lakes do occur,
but  on  different scales of productivity than in other
latitudes. It is likely that natural fertility and erodibility
of substrates yield a baseline richness greater than
usually found elsewhere. A program to examine the
natural background of loading likely to occur must be
mounted across the different regions of Puerto Rico in
order to begin to set standards. Concomitant with
                Table 6.—Comparison of productivity factors for Puerto Rican and Florida lakes.*
Lake
Caonillas
Carite
Loiza
Cidra
Dos Bocas
Qarzas
Guajataca
Guayo
La Plata
Luchetti
Patillas

Mean ± S.E. =











Net P
per unit
Phosphorus
mgc/hr/g TP
2538
4335
787
10650
1858
732
4730
4278
964
7812
4281

390 ±982











Net P
per unit
Chlorophyll
mgc/hr/mg Chi
4.61
7.23
17.93
30.43
12.24
5.86
6.39
11.03
12.47
15.02
14.27

12.50 ±2.20











Chlorophyll
per unit
Phosphorus
mg chl/g TP
500.75
600.00
43.89
350.00
151.75
125.00
740.00
387.78
77.27
520.00
292.13

349 ± 74











Lake
Santa Fe
Newman's
Orange
Lochloosa
Altho
Cooler
L. Santa Fe
L. Orange
Tuscawilla
Watermelon
Wauberg
Alice
Bivens Arm
Burnt
Elizabeth
Hawthorne
Hickory
Jeggord
Kanapaha'
Long
Moss Lee
Palatka
Trout
#10
Net P
per unit
Phosphorus
mg<=/hr/g TP
123
1072
860
712
114
544
66
318
44
107
829
0
242
151
5
694
188
35
64
36
300
42
66
174
Net P
per unit
Chlorophyll
mgc/hr/mg Chi
2.43
6.22
4.44
2.85
2.12
3.97
2.01
3.09
2.12
.83
4.13
0.0
7.04
2.54
0.19
2.86
1.18
1.00
6.28
0.55
4.11
1.01
0.97
5.29
Chlorophyll
per unit
Phosphorus
mg chl/g TP
50
172
194
250
54
137
33
103
21
129
201
3
34
59
27
242
159
36
23
64
73
42
68
33
Mean ± S.E. =
             283 + 66
                                                                            2.80 ±.42
92 ±16
*See text for sources of data
                                                 111

-------
LAKE AND RESERVOIR MANAGEMENT

such determinations must be more detailed studies of
primary productivity in relation  to standing crop of
nutrients and consumers.
  5. A trophic state index based  on epilimnetic or sur-
face water chlorophyll a concentrations appears to be
the most desirable. We are considering developing an
index using integral chlorophyll a per m2 down to Sec-
chi depth and incorporating in that value estimates, of
macrophyte chlorophyll  normalized per m2 of total
lake surface.
REFERENCES

Brown, R.A., et al. 1979. Preliminary results from a survey of
  water quality in some Puerto Rican lakes. CEER-15. Center
  Energy Environ. Res.

Carl, H., and L.F. Small. 1965. Variations in photosynthesis
  assimilation rates in natural phytoplankton communit es.
  Limnol. Oceanogr. 10 (Supl.) R67-R73.
Carlson, R.E.  1977. A trophic state index for lakes. Limnol.
  Oceanogr. 22(2):361-69.
Garcia, J.R., and L.J. Tilly. 1983. Spatial and  seasonal pattern
  of nutrient  availability in  La Plata Lake, Puerto Rico. In
  Lake and Reservoir Management, Proc. Symp. N. Am. Lake
  Manage.  Soc.  U.S. Environ. Prot.  Agency, Washington,
  D.C.
Gomez, F.G.,  and A.T.  Gonzalez.  1978. Preliminary trophic
  state classification of seven reservoirs in Puerto Rico (and
  extrapolation to include other island lakes). Prep,  for La
  Junta de Calidad Ambiental, U.S. Geolog. Surv.
Lewis, W.M. Jr.  1974. Primary production in the  plankton
  community of a tropical lake. Ecol. Mono. 44: 377-409.
	.  1983.  Temperature heat  and   mixing  in  Lake
  Valencia, Venezuela. Am.  Soc. Limnol. Oceanogr.  28:
  273-86.

Martinez, R. 1979. Estudio comparativo de la limnologia de
  los embalses mayores de Puerto Rico.  Master's Thesis.
  Dep. Biology, Univ. Puerto Rico (Rio Piedras).
Negron, E. 1983. A study of eutrophication and  aquatic
  plants growth in selected lakes and rivers of Puerto Rico.
  Proj. No. A-071-PR. Univ. Puerto Rico Water Resour. Res.
  Inst.

Nevarez, R. Jr.,  and J. Villamil.  1981.  Productividad y
  contenido nutricional  del  jacinto de  agua,  Eichhornia
  crassipes Mart (Solms), en relacion a algunos aspectos
  limnologicos del Lago Carraizo, Puerto Rico. CEER-T-096.
  Center Energy Environ. Res.

Porcella, D.B., S.A. Peterson, and D.P. Larsen. 1980. An index
  to evaluate lake restoration. J. Environ. Eng. Div. Proc. Am.
  Soc. Civil Eng. 106(EE6).

Quinones-Marquez, F. 1980. Limnology of Lago Loiza, Puerto
  Rico. WRI 79-97. U.S. Geolog. Surv.

Robarts, R.D.  1982. Primary production of Lake Mcllwaine.
  Pages 110-17 in J.A. Thornton, ed. Lake Mcllwaine: the
  Eutrophication and  Recovery of a Tropical African Man-
  made  Lake.  W. Junk Publishers, The Hague-Boston-
  London.

Reckhow, K.H. 1979. Quantitative Assessment of Lake Quali-
  ty.   EPA-440/5-79-015.  U.S.  Environ.  Prot. Agency,
  Washington, D.C.

	1981.  Lake  data  analysis and nutrient  budget
  modeling. EPA-660/3-81-011. U.S.  Environ. Prot. Agency,
  Washington, D.C.

Shannon,  E.D.,   and  P.L.  Brezonik. 1972. Eutrophication
  analysis: A multivariate approach. J. San. Eng. Div. Am.
  Soc. Civil Eng. 998(1):37-57.

Thornton,  K.W.  et al. 1980.  Reservoir sedimentation and
  water quality—an heuristic model. Pages 654-61.  In Proc.
  Sym. Surface Water Impoundments. Am. Soc. Civil Eng.
  June 2-5, Minneapolis, Minn.
Walker, W.W. 1977.  Use of hypolimnetic oxygen depletion
  rate as a trophic state index for lakes. Water Resour. Res.
  15(6):1463-70.
                                                     112

-------
                           Internal  Nutrient   Cycling
ENHANCEMENT  OF INTERNAL  CYCLING OF
PHOSPHORUS BY AQUATIC  MACROPHYTES, WITH
IMPLICATIONS FOR LAKE MANAGEMENT
B. C. MOORE
H. L GIBBONS
W. H. FUNK
T. McKARNS
J. NYZNYK
M. V. GIBBONS
Washington State University
Pullman, Washington



           ABSTRACT

           Observations over a 14-year period at Liberty Lake, Wash., have indicated a close relationship bet-
           ween the seasonal decline of aquatic macrophyte populations and the onset of planktonic Cyanobacteria
           blooms. Tracer methods, using radiophosphorus, have been employed in laboratory and in in situ
           experiments to investigate the ability of Elodea canadensis, an important component of the Liberty
           Lake macrophyte community, to translocate phosphorus from sediments to the open water. Results
           of these experiments showed good agreement between release rates determined in the laboratory
           and in situ for senescing macrophytes. Experiments with actively growing Elodea plants indicate some
           release or leakage of phosphorus from healthy plants. Nutrient budgets and a phosphorus model for
           Liberty Lake indicate that internal cycling of sediment phosphorus by aquatic macrophytes is an im-
           portant source of phosphorus to planktonic primary production as well as direct sediment/water ex-
           change. Indeed, in Liberty Lake, it is possible that macrophyte influence on the dynamic cycling of
           phosphorus in the lake may  not only influence, but also control the pattern, timing, and community
           composition of planktonic production  A conceptual framework that can be applied by lake managers
           for determining the potential contribution of macrophyte phosphorus cycling in lakes is discussed.
INTRODUCTION

The significance of internal cycling of phosphorus in
the lentic environment has been recognized for some
time (Hutchinson, 1957, 1967; Wetzel, 1983). However,
the mechanisms that govern internal cycling are still
being explored. At Liberty Lake, Wash., the pathways
of phosphorus cycling within the lake are only now be-
ginning to be defined.
  An intensive limnological investigation has been
underway at Liberty Lake since the early 1970's (Funk
et  al.  1975, 1982,  1983;  Gibbons, 1976). In these
studies, it has been postulated that the phytoplankton
productivity of  the lake is highly  dependent on the
roles of the sediments and aquatic macrophytes in the
phosphorus  cycle.  The  dependency of the phyto-
plankton on these sources of internal phosphorus is
such that the pattern, timing, and community struc-
ture of the planktonic productivity is controlled by the
availability of phosphorus from the sediments and
                                            113

-------
 LAKE AND RESERVOIR MANAGEMENT
 aquatic macrophytes. This was especially evident in
 the study by Gibbons (1981) where it was observed
 that the algal highs and lows in productivity corres-
 pond to the availability of phosphorus from internal
 sources.
   There have been two large pulses in the phytoplank-
 tonic productivity during  the growth seasons over
 the past 2 years of investigation at Liberty Lake. Tie
 first pulse  in midsummer  was composed of diatoms
 and  blue-greens.  Although  the  productivity  was
 elevated throughout the water column during these
 episodes, the bulk of the biomass and photosynthel ic
 activity was near  and among  the aquatic  macro-
 phytes. That phenomenon led to the hypothesis that
 the sediment was releasing a significant quantity of
 phosphorus to the overlying  water. Measurement of
 the sediment release of phosphorus has been defined
 by Mawson et al. (1983). The second pulse of phylo-
 planktonic productivity was related to the intensity
 and timing of the  macrophyte senescence. The quanti-
 ty of phosphorus released into the water has  been
 determined in  previous investigations (Funk  et  al.
 1983).
   The objectives  of the present study were to deter-
 mine if an aquatic macrophyte, Elodea canadensis,
 was  acting as  a  phosphorus pump during active
 growth and, if so, to quantify that enhancement of the
 internal cycling of  phosphorus by the macrophytes.
 An additional objective was to assess  the implica-
 tions that  this internal cycling  of phosphorus may
 have  for lake management programs.
 METHODS

 A comprehensive nutrient budget for Liberty Lake has
 been constructed since 1977. The budget has included
 data for  external  and  internal components.  The
 measurements of surface water, 1982,  1983  ground-
 water, runoff and direct precipitation  inputs have  been
 made and presented by Funk et al. (1982,1983).  Loading
 via the sediments and aquatic macrophytes senescence
 WICOMICO
 BEACH
          DREAMWOOO BAY
   LAKE SURFACE ELEVATION = 62428 METERS
LAKE DEPTHS IN METERS

[Jj]] CERATOPHYLLUM
j=^ ELODEA
§ NUPHAR
m UTRICULARIA

^ NITELLA
^ POTAMOGETON
[ | ELEOCHARIS
p-T] ISOETES
ORK ^fejJT EAST




                COMPOSITE  I982
Figure 1.—Location of the dense aquatic macrophyte beds
within Liberty Lake showing the dominant plants; composite
of coverage over the growth season for 1982.
 has also been  made (Moore,  1981;  Mawson et al.
 1983;  Funk et al. 1983). The determinations of the
 release of phosphorus during active growth by the
 aquatic macrophyte E. canadensis, a dominant plant
 in  Liberty Lake (Fig.  1),  were  done  in  both  the
 laboratory and in situ.
   Rates of translocation and release of phosphorus
 from £. canadensis were determined during periods of
 active growth in the laboratory in the spring and early
 summer of 1983. Specimens of  E. canadensis, lake
 water, and  sediments  were collected  from  Liberty
 Lake.  The  sediment  was injected  with 32p as
 NaH232PO4 and allowed to equilibrate for 30 days. The
 sediment was then placed in 250  m  Erhlenmeyer
 flasks and partially sealed with paraffin wax, leaving a
 1-cm diameter opening. Three preweighed  segments
 of  E.  canadensis were planted  into the  sediment
 through  the  opening in the seal.  Eight flasks were
 placed in 38-liter aquaria filled  with  filtered (80-^m
 filter) Liberty  Lake water. The aquaria were exposed to
 a 16:8 hour light/dark cycle, aerated, and kept in a 20°
 Celsius bioassay room. The control aquarium was ex-
 posed to the same conditions, except that a wooden
 stake was placed in the sediment instead of the grow-
 ing plants. Hence, any release of radiophosphorus in
 the control aquarium was the result of direct diffusion
 from the sediments to the overlying water. The results
 of the tracking of phosphorus in the control  aquarium
 were subtracted  from the experimental results. During
 the course of the experiment, temperature, pH, and
 32P in  the aquaria water were monitored daily. Total
 phosphorus concentrations were determined every 10
 days for  the  60-day experiment. Periodically, the 32P
 content of the periphyton on the bottom  and sides of
 the  aquaria  was also  measured. Growth  rates of
 Elodea canadensis were made using 14C method and
 biomass  determinations.
  The  in situ experiments were conducted in Liberty
 Lake using two specially constructed 0.61 m (2 ft)
 square by 1.2 m  (4 ft) high plexiglass enclosures. The
 enclosures were fitted with a lid to prevent trapping of
 air when placing the boxes. A metal skirt was attached
 to the bottom of the plexiglass to support the chamber
 in the soft sediments and to prevent lateral  diffusion
 of the tracer.  A pair of rubber gloves was fitted to one
 wall of the enclosures so that materials inside the
 boxes  could  be  manipulated without contacting the
 material  inside.  A 2-inch diameter hole fitted with a
 stopper was provided for passing materials into and
 out of  the boxes. Plexiglass strips for sampling the
 periphyton growth were hung inside the boxes.
  The  boxes  were placed in the lake at a  depth of
 about 3.5 m (11 ft) with the lids left open for approxi-
 mately 1  month  before the  initiation of  each experi-
 ment. Macrophytes in one enclosure were completely
 removed  and  were left intact in the other enclosure.
 Elodea canadensis was  the dominant macrophyte in
the enclosure as well as in the area where the boxes
were located.  A small growth of Ceratophyllum demer-
sum was also present. The  experiments were begun
 by first sealing the boxes, then injecting a buffered
 monosodium  phosphoric acid solution containing 10
 millicuries of 32p. The injections were made by syringe
to  a depth of about 10 centimeters  at  20  different
points  in  the sediment enclosed by each  box.
  Following the tracer  injections, the  boxes were
sampled  weekly  by  scuba divers.  Samples  taken on
each trip included water samples from the top, middle,
and bottom of each box, periphyton strips from each
box, and macrophyte samples from the macrophyte
box. Macrophytes were sampled by using tweezers to
                                                 114

-------
                                                                             INTERNAL NUTRIENT CYCLING
remove leaves from the basal and apical portions of
the plant. Phosphorus determinations were made on
the samples on  a Technicon Auto-Analyser II,  and
radiophosphorus  activity was determined using  a li-
quid scintillation counter. Phosphorus measurements
made from within the box without macrophytes were
used to determine phosphorus release caused by the
sediments alone and were used to correct the results
obtained in the box with macrophytes for direct sedi-
ment inputs.
   The  experimental relationship observed between
dissolved oxygen and the release of phosphorus from
the sediment were used to estimate  Internal loading
directly  caused  by  the  sediments  into  the  lake
(Mawson et al. 1983).


STUDY AREA

Liberty  Lake,  located 21  km  (13   miles) east  of
Spokane, Wash., is a soft-water lake (288 ha) of glacial
origin enclosed on three sides by a small mountain
range 300 to 500 m above the lake surface. Most of the
watershed (3,445 ha)  lies in this  horseshoe-shaped
basin,  forested  with  Ponderosa  pine,  grand  fir,
Douglas fir, larch, white pine, and aspen. The major
tributary, Liberty Creek, originates in the higher south-
eastern slopes and passes through a soil series of
Moscow  and Springdale types  before  reaching  the
Spokane and Semihoo muck series adjacent to and in
a marsh. The stream flows along the eastern margin of
the marsh (and until recently, overflowed into it) before
entering the lake. Most of the tributary area is under-
lain by quartz-feldsparbiotite paragneiss. Residential
areas occupy 87 percent of the shoreline and overlie
relatively  shallow soils  (Spokane  series).  Gneiss
(western side and northern shore) and Columbia River
basalt (western shore) form the  bedrock. A small  un-
named creek enters the lake from the  northwestern
side. Until 1979, waste disposal had  been by septic
tank and an old sewer system built in 1910, which serv-
ed approximately 40 percent of the residents. In late
1979, a sewage collection system which now serves
about 2,000 permanent residents was completed. This
system  diverted  99  percent of  the domestic sewage
from the lake basin.
  The mean residence time of lake waters is 3 years.
Approximately 2.76 x  106 m3 yr- 1 is lost by seepage,
presumably through the bottom at the northern end of
the lake. The lake may become weakly stratified for
short periods of time during the mid- and late-summer
period.
RESULTS

The  phosphorus tracer  study  conducted  in  the
laboratory  not  only generated  a positive  rate of
release of phosphorus by actively growing plants but
also verified a contention made by Gibbons (1981) that
the exchange of phosphorus from one component to
another in Liberty Lake  was controlled and driven
dynamically by the biota.  This fact is evident by com-
paring the results of the experiments presented in
Figures 2 and 3. The difference in the magnitude of
phosphorus lost  from  the  actively  growing  Elodea
canadensis (see Fig. 2 and 3) was caused by analyzing
water alone (Fig. 2) versus analyzing periphyton  plus
the water (Fig. 3). The loss rate of 0.035 ^gP«g-i »D-1
of E. canadensis was observed when the 32P concen-
tration within  the periphyton  on the  sides of the
aquaria were ignored. However, when the periphyton
 was analyzed for 32pj it was found that they contained
 several orders of magnitude more phosphorus than
 the water did. Thus, the real loss of phosphorus by E.
 canadensis into the aquarium was 25ngP«g-1»D-1.
 The  release of phosphorus from £. canadensis could
 have been enhanced by the biotic sink made up of the
 periphyton community.
   It  must be noted that the release or  loss of phos-
 phorus from E. canadensis has been presented as a
 rate per biomass of the plants. With an increase in the
 biomass of the plants, there was a corresponding in-
                                           70
Figure 2.—Rate of phosphorus release to the overlying water
via Elodea canadensis. The regression line was y = 0.035x
+ 4.19, r  = 0.999.

     I400r
  „ 1200
  — 1000
                                                         800
   co
 31
   g
   m
600
     400
     aoo
           20      30      40     50
                      TIME (days)
                                 60
                                        70
Figure 3.—Summation of phosphorus release from Elodea
canadensis within the experimental microcosm to the water
and periphyton.
                                                115

-------
 LAKE AND RESERVOIR MANAGEMENT
 crease in the total quantity of phosphorus released
 from the plants. It appears that the rate of release
 depends on a mechanism that is independent of the
 size and surface area of the plants. The release rate
 reflects the ability of £ canadensls to translocate
 phosphorus from the sediment to the water in excess
 of what would be released by the sediments alone.
   Figure 4 presents the total phosphorus (TP) data ob-
 tained  from the in situ  experiments.  Based on  the
 data, it can be seen that the TP concentrations in the
 water from the chamber with sediments and no E.
 canadensls only increases  slightly. However,  there
 may have been a tendency for higher concentrations if
 the experiment  had  been continued. The TP concen-
 trations reflect the availability of phosphorus from the
 sediments. Since the data in Figure 4 are only for the
 last 60 days of the 95-day incubation period, the abil ity
 of the sediments to supply  phosphorus to the overly-
 ing water on a daily rate was small as compared to the
 release rate in the Elodea canadensis chamber. The
 TP concentration of the water from within the plant
 chamber increased with time. Because of the char-
 acteristics of the chamber, the  increase  in phos-
 phorus was the  result of  loss from £ canadensis. The
 plants  both inside and  outside  of the chamber  in-
 creased in biomass and size during the experiment, so
 the observed release of  phosphorus occurred during
 active growth. The release rate of £ canadensis minus
 the sediment contribution was calculated to be 1.27
 ngP»ml-i«D-i.
  The net results of the in situ radiophosphorus trace
experiment are in close agreement with the TP results.
The calculated release rates from E. canadensis dur-
ing active growth were 1.02  ngP»m|-1«D-1  and 0.86
ngP«ml-1»D-1 (Fig. 5). High measurements  made on

Table 1.—Rates of phosphorus release by Elodea canadentis
          during active growth in Liberty Lake.
Mode of Determination
                                    Release Rate,
Laboratory experiments
In situ experiments
  Total phosphorus concentration
  32P - high rate
  32p - low rate
                                        13.1


                                        11.0
                                         8.8
                                         7.5
  Table 2.—Phosphorus loading Into Liberty Lake from tho
  release of phosphorus by Elodea canadensis during the"
               growth season (105 days).

                                     Release Rate,
                                         KgP
Mode of Determination
 Laboratory experiments
 In situ experiments
  Total phosphorus concentration
  32R - high rate
  32P - low rate
                                        38.5

                                        32.3
                                        25.9
                                        22.1
                                                     Days 18 and 50 were not included in the construction of
                                                     the regression line that yielded the low rate of release
                                                     but were included in the regression  line that yielded
                                                     the high rate of release.
                                                       The availability of light was greatly reduced on Days
                                                     28,  33, 42, and 60 because of a succession of algal
                                                     blooms in  the lake. Just prior to and on day 50, the
                                                     available  light at the experimental chambers  was
                                                     higher as  a result of a break in the algal blooms. This
                                                     would possibly indicate that photosynthetic activity
                                                     may have  a  direct effect  upon the release of phos-
                                                     phorus by £ canadensis.
                                                       Table 1  summarizes the calculated release rates of
                                                     phosphorus as they apply to Liberty Lake during the
                                                     growth  of E. canadensis. Assuming the  period of
                                                     growth  for £ canadensls to be  105 days (which is
                                                     most likely an underestimation of the actual time of
                                                     growth), and given that the plant covers approximately
                                                     28 ha (Fig. 1), the loading of phosphorus during active
                                                     growth is  between 22.1 and 38.5 kg (Table 2).  The
                                                     estimate of loading varies with the method of calcula-
                                                     tion used. Significance of the relatively narrow range
                                                     of the release rates calculated by three independent
                                                     techniques should not be overlooked.
   loo



    90


    80


    70



r   60
<»<
 E
£L  50
 o>
 c

    40



    30


    20


    10
                                                                   MACROPHYTE * SEDIMENT
                                                                   SEDIMENT
                                                                   MACROPHYTE
                                                                                       y= I.27X * 6.45
            10
                                       50
                                               60
                   20     30     40
                    TIME (days)
Figure 4.—In situ total phosphorus accumulation within the
sediment and the aquatic macrophyte chambers. Regression
line of macrophyte minus the sediment phosphorus was y =
1.27x + 6.45, r = 0.77.
                 Table 3.—Total phosphorus loading to Liberty Lake detailing internal sources in kg.
Year
1980
1981
1982
External
Sources
194.4
476.5
401.5
Sediment
Release
30.9
31.1
31.1
Macrophyte
Senescence
13.0
27.5
18.0
Elodea canadensis
Release
22.1 to 38.5*
22.1 to 38.5*
22.1 to 38.5
% Internal
Loading
34 to 42
17 to 24
18 to 22
'1980 and 1981 estimates taken from 1982 measurements.
                                                  116

-------
                                                                                 INTERNAL NUTRIENT CYCLING
E
Q.
140



100'



90



80



70



60



50



40



30



20
                y= I.OI5X + 29
             10
                                      50
                                                60
                    20     30     40
                     TIME  (days)
Figure  5.—In  situ phosphorus accumulation within  the
macrophyte chamber minus sediment release as measured
by radiophosphorus tracer. Both high and low regressions
are presented.

   Summation of phosphorus loading data (Table 3)
reveals that internal loading makes up  17 to 42 per-
cent of the total loading.  The aquatic  macrophytes
were responsible for more than 50 percent of the inter-
nal loading, and the timing of that loading is extremely
important in  maintaining the character  of the phyto-
planktonic productivity. It is  realized  that  Elodea
canadensis is only one of the species of  rooted plants
in the lake. Further research is needed to determine if
the other  plant  species  are  phosphorus sinks  or
sources.

DISCUSSION

It has been shown that internal cycling of phosphorus
is an important source of phosphorus to Liberty Lake.
Because of the nature of the experimental conditions,
the macrophyte data represent a lower estimate of the
actual macrophyte contribution.
   The question of the release of phosphorus by living
aquatic  macrophytes  has  not  been  adequately
answered.  Conflicting evidence has been reported by
different investigators based on  laboratory and field
observations.  For  example,  McRoy  and  Barsdate
(1970) reported leaching of phosphorus by eelgrass in
laboratory  experiments,  and  Twilley  et al.  (1977)
reported leaching of phosphorus  from Nupharluteum
under  both field  and laboratory  conditions.  The pre-
sent   experiments  with  E.   canadensis  have
demonstrated that the release of phosphorus by ac-
tively growing plants does  take place both  in  the
laboratory  and in situ. On  the other hand, Barko and
Smart (1980) observed almost no phosphorus release
by living plants in laboratory experiments, and similar
results from field observations have been made  by
Adams and  Prentki (1982) and  Carignan and Kalff
(1982). However,  these investigators were observing
plant  species  other than Elodea canadensis.
  In past in situ experiments with £ canadensis in
Liberty  Lake, the rate of phosphorus release during
senescence was observed to be between 15 and 32
gP«g-i«D-i. The rate of phosphorus release during
senescence compared to the  rate  of phosphorus
release  during  active  growth  of  E.  canadensis
demonstrates  how important release during growth
can be to a system. This is especially true in Liberty
Lake, where the nutrients are in critical flux.
  For the lake manager faced with decisions regar-
ding the possible control strategies for limiting phyto-
plankton  production,  a  number of  options  are
available for controlling internal sources. However, in
shallow lakes,  internal cycling may be a  significant
source of nutrients, and the control of those nutrients
may be more difficult. It is extremely important  that
the manager have reliable and  realistic data on the
nutrient dynamics of a lake so that effective as well as
efficient strategies can be developed.  Additional work
is needed in this area. It is possible that the release of
phosphorus  by  aquatic  plants  is  species-  and
substrate-dependent.

 REFERENCES

Adams,  M.S., and R.T. Prentki.  1982. Biology,  metabolism,
  and functions of littoral  submersed weedbeds of  Lake
  Wingra, Wis. USA: A summary and review. Arch. Hydrobiol.
  (Suppl.) 62:333-409.
Barko, J.W, and  R.M. Smart. 1980. Mobilization of sediment
  phosphorus  by  submersed   freshwater macrophytes.
  Freshw. Biol. 10:229-38.
Carignan, R., and J. Kalff. 1982.  Phosphorus release by sub-
  mersed  macrophytes:  significance to  epiphyton  and
  phytoplankton. Limnol. Oceanogr. 27:419-27.
Funk, W.H., et al. 1975. Determination, extent, and nature of
  nonpoint source enrichment of Liberty  Lake and possible
  treatment. Wash.   Water Res.  Center Rep.  No.  23.
  Washington State Univ., Pullman.
	1982. Preliminary assessment of multiphase restor-
  ation  efforts at  Liberty Lake, Washington.  Wash. Water
  Res. Center Rep. No. 43. Washington State Univ., Pullman.
	1983. Post-treatment investigation of a multiphase
  lake restoration of Liberty Lake, Washington. Wash. Water
  Res. Center, Washington State Univ., Pullman.
Gibbons, H.L 1976. The primary productivity and related
  factors of Liberty Lake, Newman, and  Williams Lakes in
  Eastern Washington. M.S. thesis, Washington State Univ.,
  Pullman.
	1981.  Phytoplankton  production  and regulating
  factors in Liberty Lake, Washington with possible implica-
  tion of restoration activities on the primary  productivity.
  Ph.D. thesis, Washington State Univ., Pullman.
Hutchinson, G.E. 1957. A Treatise of Limnology. Vol. I. Geo-
  graphy, physics, and  chemistry. John Wiley and Sons,
  New York.
	1967. A treatise on limnology. Vol. II.  Introduction
  to lake biology  and the limnoplankton. John Wiley and
  Sons, New York.
Mawson, S.J., H.L. Gibbons, Jr., W.H.  Funk, and K.E. Hartz.
  1983. Phosphorus flux rates in lake sediments. J. Water
  Pollut. Control Fed. 55(8): 1105-10.
McRoy, C.P., and R.J. Barsdate. 1970. Phosphate absorption
  in eelgrass. Limnol. Oceanogr. 15:6-13.
Moore, B.C. 1981. Release of sediment phosphorus by Elodea
  canadensis. M.S. thesis, Washington State Univ., Pullman.
Twilley, R.R., M.M. Brison, and G.J. Davis.  1977. Phosphorus
  absorption, translocation, and secretion in Nupharluteum.
  Limnol. Oceanogr. 22:1022-32.
Wetzel, R.G. 1983. Limnology. Saunders College Publishing,
  New York.
                                                  117

-------
 REDUCING SEDIMENT PHOSPHORUS  RELEASE
 RATES IN  LONG  LAKE  THROUGH THE USE OF
 CALCIUM  NITRATE
PETER R. WILLENBRING
MARK S.  MILLER
WILLIAM  D. WEIDENBACHER
E. A. Hickok and Associates
Wayzata,  Minnesota



            ABSTRACT

            The effect of injecting different dosage rates of calcium nitrate into the bottom sediments of Long
            Lake in New Brighton, Minn., was observed in a laboratory study. The purpose of the study was to
            determine the minimum dosage necessary to reduce phosphorus release rates from the sediments
            to satisfactory levels. The study included measurement of sediment phosphorus release rates under
            both aerobic and anaerobic conditions, and evaluated the effect of adding ferric chloride to the sediments
            along with the calcium nitrate. The addition of ferric chloride was included in the study to determine
            if the iron available in the sediment was adequate to sorb the PO,f3 when oxidized conditions were
            provided. The study concluded that at least for the short term (90 days), injection of calcium nitrate
            could eliminate virtually all phosphorus releases from the sediments previously releasing phosphorus
            at a rate of 7 mg P/m2/day, and actually result in the sediments becoming a sink for phosphorus in
            the water column. The study also concluded that although the addition of iron enhanced the calcium
            nitrate treatment's effectiveness, similar results could be achieved by increasing the calcium nitrate
            dose slightly and not adding the iron, which was a more cost-effective alternative.
INTRODUCTION

Phosphorus release from the bottom sediments ha;;
been  identified as one  of  the  major factors con-
tributing to the accelerated eutrophication of Long
Lake.  Studies previously completed on the oxidation
of lake sediments with nitrate (Ripl,  1978) indicated
that  phosphorus release from  the  sediment of a
sewage-impacted lake could be virtually eliminated if
a sufficient dose of calcium nitrate was applied to the
sediment.  For this reason, the technique of oxidizing
lake sediments with  nitrate is being seriously con-
sidered by the Rice Creek Watershed District as a way
to reduce  the  internal phosphorus loading to Long
Lake.  Prior to  application, however,  it was deemed
necessary to conduct  laboratory experiments to quan-
tify the effect of the treatment on  Long Lake sediment
under conditions similar to those found in Long Lake.
Specifically, this study was developed to provide infor-
mation on:
  1. Anaerobic phosphorus release rates  of the  un-
treated sediment.
    Table 1.—Typical analysis of liquid calcium nitrate:
                    66 percent.
66.05 percent
11.20 percent
 7.80 percent
 0.003 percent
 0.005 percent
 0.01 percent
1.45-1.46 at 15°C
- 20°C/ - 4.0°F
3.0 at 15°C
         4 H2O
 Calcium, as Ca
 Nitrogen, from NO3
 Iron, as Fe
 Manganese, as Mn
 Ammonia, as NH3
 Specific gravity
 Freezing point
 pH
                    2. Anaerobic  phosphorus release  rates of  the
                  sediments  treated with differing dosage rates of
                  calcium nitrate, along with the establishment of the
                  minimum  dosage of  calcium  nitrate  necessary to
                  reduce phosphorus loading from Long Lake sediment
                  to the desired levels.
                    3. Beneficial effects that could be obtained by ad-
                  ding ferric chloride and slaked lime to the sediment in
                  conjunction with the addition of calcium nitrate.
                    4. Depth and uniformity of sediment oxidation for a
                  given calcium nitrate dosage rate.
BACKGROUND

Phosphorus can be released from the bottom sedi-
ments  through  two  primary  mechanisms.  One  is
through the decomposition of bottom sediment, a pro-
cess through which  organic matter or reduced in-
organic matter  is  stabilized  through aerobic  or
anaerobic chemical or biological reactions. The other
allows  for phosphorus to be introduced into the water
column as ferric hydroxides and complexes are reduc-
ed.
  The method investigated in this study is that of ap-
plying  liquid  calcium  nitrate (LCN)  four  hydrate
[(CaNO3)2 • 4 H2O] (see Table 1 for typical  analysis)
directly into the sediment in an effort to significantly
reduce the amount of phosphorus released  into the
water column above  the sediment during  anaerobic
periods caused by the decomposition process. The
calcium nitrate injection  deposits additional nitrate
(N03) in the sediment. This nitrate is subsequently
reduced to nitrite and nitrogen gas through denitrifica-
                                                118

-------
                                                                             INTERNAL NUTRIENT CYCLING
tion, which will oxidize the top centimeters of sedi-
ment if a sufficient amount of nitrate is added. The ox-
idation process in turn converts iron in the sediment
and water column from a ferrous state (Fe + +) to ferric
hydroxide (FeOH)3. Ferric hydroxide  readily combines
with phosphate (P04) and prevents phosphate from
being released into the water column  from the sedi-
ment.
STUDY AREA DESCRIPTION

The South Basin of Long Lake is located in Section 20,
T30N, R23W, in the city of New Brighton, a northern
suburb of Minneapolis, Minn., in Ramsey County.  In-
terstate 694 borders the lake on the south and Inter-
state 35W is approximately 1 kilometer east of Long
Lake. A 4,570 hectare drainage area contributes runoff
to the South Basin of Long Lake, with a major portion
of this watershed runoff filtered  through  lakes and
wetlands prior to its ultimate disposition into the lake.
   The South Basin of Long Lake is 48 hectares in size,
with approximately 28 hectares of that area exceeding
a  depth of 4.5 meters. Stratification typically occurs
above  the 4.5 meter depth  and therefore results in
most of the organic bottom sediments being exposed
to anaerobic conditions during periods of stratifica-
tion. Analysis of the organic  lake bottom sediment in-
dicates that the sediment has, on a dry weight basis,
an average phosphorus concentration of 270 mg/kg,
and an average iron concentration of 27,900 mg/kg.
   Soils in the upstream area are predominantly of the
Hayden  series, being  typically  well  drained  and
moderately permeable on  ground moraine. The lake
bottom is comprised  of predominantly sandy sedi-
ments  in areas of the lake with depths less than 4.5
meters. In areas of the lake with depths greater than
4.5 meters,  mucky organic  sediments predominate.
Few rooted aquatic macrophytes are present in  the
lake or along its shallow water fringes.
RESULTS

As can be observed from Figures 1 through 4 and
Table 3, the peak phosphorus release rates from the
untreated sediment ranged from 5.9 to 8.6 mg phos-
phorus/square meter/day, depending on the time inter-
val used. For sediment treated with LCN, phosphorus
release rates dropped as the LCN dosage rate increas-
ed. Results also indicate that applying  iron and lime
along  with  calcium nitrate further  reduced  phos-
                        P Concentration In Water Column
                        ADOVO untreated Sediment
            4       8      1216      20     24
                      TIME (DAYS)
Figure 1.—Phosphorus concentration in Long Lake for un-
treated sediment.
PROCEDURES AND METHODS FOLLOWED
Organic sediment deposits obtained from Long Lake
at water depths of 4.5 meters or greater were placed in
a number of aquaria, each with a volume of 75.7 liters.
The aquaria dimensions were 0.61 m wide by 0.30 m
deep by 0.41 m  high. The sediment was placed 0.25 m
deep in the aquaria; the remaining 0.16 m was filled
with water from Long Lake.
  The sediment was treated with LCN, ferric chloride,
and calcium hydroxide at the dosage rates shown in
Table 2. The chemicals were hydraulically injected in-
to the  sediment through nozzles spaced 5 cm apart
and to a depth of  12.7 cm. The aquaria were then
covered and the denitrification process allowed to pro-
ceed for  45  days.  After  this time, triplicate core
samples 15 cm deep with a total volume of 300 ml
were transferred from  each aquaria  into  550  ml
polyethylene bottles. 0.2 grams of glucose were added
to maintain anaerobic conditions inside the chamber
and 250 ml of lake water was added to bring the water
level up to the brim of the bottle. The bottles were then
sealed  and analyzed for soluble reactive phosphorus
and total iron at 1-week intervals to study the effect of
the additions of ferric chloride, calcium hydroxide,
and varying amounts  of LCN on the rate of phos-
phorus release  from the sediment under anaerobic
conditions.
              P Continuation In Water Column
              bovo Sediment Treated with LCN
              I  70 g-N/M*

               Concentration in Wattr Column
              bovo Sediment Treated with LCN
                70 g-N/M', Iron, and Limo
                  i
                         12
                      TIME (DAYS)

Figure 2.—Treatment with LCN at
                                                 119

-------
 LAKE AND RESERVOIR MANAGEMENT

 phorus release rates from the treated sediment over
 applying calcium nitrate alone. One anomalous value
 was obtained from sediment samples  treated  with
 LCN only  at  70 g  nitrogen/meter, as  phosphorus
 release rates from these sediments were higher than
 untreated sediments. This increased release rate may
 have resulted from the sediment being agitated during
 the application of LCN at a low dosage rate.
   Visual observations on the depth, consistency, and
 uniformity of the oxidized sediment as a function of
 application rate are shown in Table 4. As would be ex-
 pected, as  the application rate of calcium  nitrate in-
 creases, so does the depth and uniformity  of the ox-
 idized sediment. The oxidized sediment also became
 more  fluffy,  more finely textured, and  less  con-
 solidated than the unoxidized sediment,  making it
 more susceptible to displacement by currents and
 bottom-feeding fish.

 CONCLUSIONS

 1.  The addition of iron and lime to the sediment along
 with the calcium nitrate improves the effectiveness of
the  treatment  in  reducing  sediment  phosphorus
release rates.  This  improved  effectiveness is also
more prevalent at the lower calcium nitrate dosage
rates.
  2. For Long Lake sediment, the beneficial effects of
adding iron and lime can be obtained without adding
these chemicals if the calcium nitrate dosage rate is
increased (Fig. 3 and 4).
  3. Treating   Long   Lake  sediment  with  calcium
nitrate alone at a dosage rate of 140 g nitrogen/square
meter is the most cost-effective treatment alternative
studied that would eliminate phosphorus release from
the sediments. Although similar results could be ob-
tained by adding calcium  nitrate at a rate of  105 g
nitrogen/square meter along with iron and lime, the
cost for treating  the lake  using this  option was
estimated at $410,000 compared to a cost of $350,000
for  adding calcium nitrate alone  at a rate of  140 g
nitrogen/square meter.
  4. The nitrate treatment  should, at a minimum, ox-
idize the top  10 cm of lake sediments. Experiments
previously completed on lake muds (Hynes and Greib,
1970) indicate  that under  anoxic conditions,  phos-
                                  Table 2.—Sedimont treatment test results.
Description of
Sediment Treatment
Untreated - Control
Untreated - Control
Untreated - Control
Untreated - Control
Liquid calcium nitrate (LCN)
applied at 70 g N/m2
LCN applied at 70 g N/m2
plus iron and limed added3
LCN applied at 105 g N/m2

LCN applied at 105 g N/m2
plus iron and lime added3
LCN applied at 140 g N/m2

LCN applied at 140 g N/m2
plus iron and lime added3
Sample
No.
1
2
3
4
1
2
1
2
1
2
1
2
1
2
1
2
Day 9
SRP
.17
.63
.71
.55
.05
.10
.01
.36
.04
.16
.02
.08
.04
.02
.01
.01
Fe
(mg/l>2
17
16
13
15
24
24
14
24
19
26
32
52
24
24
35
38
PH
6.2
6.1
5.8
5.9
5.9
5.9
5.9
5.9
6.0
5.9
5.9
5.9
6.0
6.0
5.9
5.9
Day 14
SRP
1.0
.82
1.2
1.2
.55
.99
.03
.01
.04
—
.02
.01
.01
.02
.01
.01
Fe
(mg/l)2
18
20
16
16
20
20
34
30
22
16
32
40
32
26
35
35
PH
6.6
6.5
6.2
6.3
6.2
6.2
6.2
6.3
6.6
6.5
6.4
6.4
6.3
6.1
6.2
6.2
Day 22
SRP
1.1
.98
1.3
1.2
1.3
1.5
.04
.03
.06
.93
.02
.01
.02
.01
.01
.01
Fe
(mg/l)2
18
20
15
18
18
12
20
28
13
5.5
33
36
28
18
34
43
pH
6.4
6.5
6.5
6.4
6.7
6.8
7.0
7.0
6.9
6.8
6.9
6.9
6.8
6.7
7.0
6.9
1 Soluble Reactive Phosphorus concentration determined by acid and persulfate digestion and spectrophotometric analysis - EPA Method No 365 2
2lron concentration determined by acid digestion and atomic absorption - EPA Method No 236.1

3lron (FeClj) was added to the sediment at a dosage rate of 146 g Fe/m* and lime (Ca(OH),) was added at a dosage rate of 180 g Calm'.
             Table 3.—Phosphorus release rates for differing calcium nitrate treatment application rates.



Description of
Sediment Treatment
Untreated
CafNOa) at 70 g N/m2 only
Ca(NC-3) at 70 g N/m2 plus iron and lime
CafNOa) at 105 g N/m2 only
Ca(NC>3) at 105 g N/m2 plus iron and lime
Ca(NC>3) at 140 g N/m2 only
Ca(NO;j) at 140 g N/m2 plus iron and lime
Maximum
4-Day Change
in Phosphorus
Concentration
(mg/l)
0.9-0.29
0.98-0.18
0.17-0.12
0.36-0.11
0.015-0.015
0.015-0.015
0.01-0.01

Maximum 4-Day
Phosphorus
Release Rate
(mg P/m2/Day)
8.6
11.3
0.7
3.5
0.0
0.0
0.0
Maximum 8-Day
Change in
Phosphorus
Concentration
(mg/l)
1.02-0.18
1.18-0.08
0.17-0.08
0.5-0.04
0.015-0.015
0.015-0.015
0.01-0.01

Maximum 8-Day
Phosphorus
Release Rate
(mg P/m2/day)
5.9
7.7
0.6
3.2
0.0
0.0
0.0
                                                  120

-------
                                                                                    INTERNAL NUTRIENT CYCLING
         ^____ P Concontratlon In Water Column
               Aeova Sediment Treated with LCN
               01 109 g-N/M*

         	_ P Concentration in Watar Column
               Above Saolmant Treated with LCN
               at 109 a-N/M', Iron, and Lima
                           A
                       TIME (DAYS)
                                   . P Coneontratlon in Wotor Column
                                    Abova Sadlmant  Treat** with LCN
                                    at 140 o-N/M1

                                   . P Coneontratlon in Watar Column
                                    Above Sodlmont  Traotad with LCN
                                    at I4O a-N/M', Iron, on* Lime
                                             TIME (DAYS)
Figure 3.—Treatment with LCN at 105 g/N/m2.
                      Figure 4.—Treatment with LCN at 140 g/N/m2.
                         Table 4.—Visual observations of sediment 45 days after treatment.
 Description
 of Treatment
Depth of Sediment
Oxidation from Surface
          Comments
 Untreated
 sediment
 LCN applied
 at 70 g N/m2
 LCN aplied
 at 105 g N/m2
 LCN applied
 at 140 g N/m2
No oxidation observed
1-13 cm
5-15 cm
10-20 cm
Sediment compacted, firm, dark

Spotty; non-uniform oxidation
of sediment; oxidized sediment
coarse grained; slight sediment
expansion observed in areas of
oxidation
Uniform oxidation of top 4 cm of
sediment; oxidized sediment is
finer grained than that observed
at application rate of 70 g N/m2;
oxidized sediment expanded,
sediment depth increased 5 cm
Uniform oxidation of top 8-10 cm
of sediment; sediment volume
expanded to increase total
sediment depth 8-10 cm; top
layer of sediment very fine
grained
phorus can move upward  readily from a depth of at
least 10  cm. Therefore, treating sediments to this
depth or greater would be  prudent. Also, physical ex-
periments conducted on  the  oxidized  sediment  in-
dicated that the sediment became  lighter and less
consolidated than the unoxidized sediment,  allowing
it to be displaced more easily. Oxidation of sediment
to depths less than 10 m could allow the upper oxidiz-
ed layer of sediment to be disturbed by currents and
fish, thereby allowing the unoxidized sediment below
to release nutrients directly into the water column
above.
                     REFERENCES
                     Hynes, H.B.N. and B.J. Greib. 1970. Movement of phosphorus
                       and other ions from and through lake muds. J. Fish. Res
                       Board Can. 27:653-68.

                     Ripl, W. 1978. Oxidation of Lake Sediments with Nitrate—A
                       Restoration Method  for Former Recipients. Inst. Limnol.
                       Univ. Lund, Lund, Sweden.
                                                     121

-------
THE ROLE OF  INTERNAL PHOSPHORUS  LOADING
ON THE  TROPHIC STATUS  OF  NEW JERSEY'S  TWO
LARGEST LAKES
STEPHEN J. SOUZA

JOHN D.  KOPPEN

Princeton Aqua Science
New Brunswick,  New Jersey



            ABSTRACT

            Under conditions of anoxia, lake sediments will lioerate sorbed phosphorus compounds into the overlying
            water. The magnitude of the sediment-liberated phosphorus load will be a function of the concentra-
            tion of phosphorus in the sediments, the area of lake bottom overlayed by anoxic water, and the tem-
            poral duration of anoxic conditions. In some cases the resulting phosphorus load may be a substan-
            tial component of the lake's annual phosphorus budget. This may have serious implications in the
            restoration and management of such lakes. The importance of internally-generated phosphorus loads
            in the nutrient budget of New Jersey's two largest lakes, Lake Hopatcong and Greenwood Lake, was
            assessed. The formation and depth of the thermocline were established from temperature profiles
            The volume of the hypolimnion and the area of lake bottom overlayed by the hypolimnion were
            calculated. Following stratification, water column profiles of total phosphorus and dissolved oxygen
            concentrations were monitored. These data wero used to compute the internally generated phosphorus
            load. In both lakes, a 10-fold increase in the concentration of total phosphorus was observed follow-
            ing the depletion of oxygen in the hypolimnion.  The internally generated load associated with this
            phenomenon represents 5.9 percent and 29.3 percent of the annual total phosphorus budget of Lake
            Hopatcong and Greenwood Lake, respectively. The relevance of endogenous phosphorus loading
            is discussed for each lake in relation to its existing trophic status and future restoration and management
INTRODUCTION

The internal regeneration of phosphorus from lake
sediment can represent a significant source of phos-
phorus. Under aerobic conditions, lake sediment acts
as a phosphorus sink (Fillos and Biswas, 1976). The
settling  of  organic material  and  its  subsequent
decomposition lead to the remineralization  of phos-
phorus. In the presence of an oxidizing state, much of
the  remineralized  phosphorus complexes  with
hydrous ferric oxides and becomes adsorbed on the
sediments (Bannerman et al. 1974).
  When anaerobic conditions persist at the sediment-
water  interface, the   ferric-phosphate  complex
decreases.  Under these conditions,  phosphorus is
regenerated from the sediments and liberated into the
overlying water column (Fillos and Biswas, 1976). Con-
centration gradients favor  the translocation of this
liberated phosphorus into the overlying anoxic water
(Mawson et  al.  1983).  Among  other factors, the  in-
terstitial concentration of phosphorus and the sedi-
ment redox potential will determine the rate of phos-
phorus release and the  magnitude of the internally
regenerated load (Armstrong, 1979).
  The internally  regenerated phosphorus load  must
be taken into account when developing lake restora-
tion/management programs. Failure to do so could
result in an ineffective lake restoration program. It has;
been observed that the internal regeneration of phos-
phorus has delayed the recovery of some lakes even
following a sizable decrease  in the external phos-
phorus load (Freedman and Canale, 1977; Welch and
Rock, 1980).
  In this  study, the importance of  the  internally
regenerated phosphorus load is examined in relation
to the restoration of two of New Jersey's largest lakes.
The phosphorus load  generated by  sediment flux
under anaerobic conditions  is computed for Lake
Hopatcong and Greenwood Lake. These data are dis-
cussed  in relationship to developing an effective
restoration program for each lake.

METHODS

Water quality, dissolved  oxygen,  and temperature
were monitored  twice monthly  from April through
September, and  monthly throughout the remainder of
the year. For both lakes, water samples were collected
at the station of greatest  depth at 2-meter intervals.
Dissolved  oxygen  and   water  temperature  were
monitored at 1.0-meter intervals.
  Water samples were collected with a nonmetallic
Kemmerer bottle. Total phosphorus was analyzed ac-
cording to Standard Methods for the Examination of
Water and Wastewater (1975). A Rexnord Deep Water
Probe was used to measure dissolved oxygen and
temperature.
  Sediment cores were obtained using a K-B free fall
corer, equipped with 60 cm PVC liners and "egg shell"
core catchers. The sediment samples were sealed and
stored on ice in an upright  position. Upon return to the
lab, the  redox  potential  of  the upper 5 cm  was
measured. The cores were  then frozen. Once solid, the
top 5 cm strata from each core were analyzed for grain
size, total phosphorus, and  organic  content using
Standard Methods (1975).


RESULTS
Lake Hopatcong

Lake Hopatcong is  a 1,087  hectare (ha) waterbody
located  in Morris/Sussex  Counties,  N.J. (Fig. 1).  The
                                               122

-------
                                                                             INTERNAL NUTRIENT CYCLING
lake has an "oak leaf" configuration. It consists of a
central deep basin from which emanate a number of
shallow coves and embayments. Lake Hopatcong has
a total volume of 5.56 x  107m3, a mean depth of 5.49
m, and a maximum depth of 16.7 m (Table 1). In 1982,
the lake stratified on May 26 and remained stratified
until Sept. 10. The hypolimnion had a volume of 8.93 x
106m3 and overlayed a total of 2.95 x  1Q6m2 of lake
bottom.
  Anoxic  conditions persisted  in the hypolimnion at
depths in  excess of 12 meters  from early July until
lake overturn on Sept. 10, a period of about 60 days.
The anoxic layer had a volume of 2.5  x  106m3 and
overlayed  1.65 x 106rrt2 of lake bottom.
  Prior to the stratification of the lake, the concentra-
tion of total phosphorus  (TP) at depths greater than
12.0  meters averaged  0.021  g  m-3.  Following
stratification and the establishment of anoxic condi-
tions, the  concentration of TP increased dramatically
(Fig. 2 and 3). The mean TP concentration measured in
the anoxic hypolimnion was 0.145 g m-3, but con-
centrations as high as 0.50 g m-3 were recorded (Fig.
2).
  Sediment cores obtained from the deeper sections
of Lake Hopatcong (Z >5.0 m) had an average TP con-
centration of 191 mg kg. The redox potential of these
sediments ranged from -90 mV to -200 mV.

Greenwood Lake

Greenwood  Lake is located in  Passaic County, N.J.
and Orange County, N.Y. (Fig. 1). The 777 ha lake has a
volume of 5.34 x  107m3, a mean depth of 5.2 m, and a
maximum depth of 17.4 m (Table 1). Greenwood Lake
differs markedly  from  Lake Hopatcong in terms  of
morphometry and  configuration. Greenwood  Lake is
essentially composed of two basins, a relatively deep,
steep-sided  basin, and a shallow, gradually  sloping
basin. Most of  the lake's volume, and  approximately
half its surface area,  is associated with the deep
basin. The lake has a long, narrow configuration and
lacks any prominent coves or embayments.
   In 1981, temperature profiles indicated that the lake
stratified  on May 27 and remained so  until early Oc-
tober. Dissolved oxygen  became depleted at depths
     PENNSYLVANIA
                                      NEW YORK CITY
greater than 9.0 meters in  mid-June. Anoxic condi-
tions persisted until the autumnal overturn, a period of
approximately 90 days. The anoxic hypolimnion had a
volume of 1.112 x 107m3 and overlayed 5.46 x 106m2
of lake bottom.
   Prior to stratification the mean concentration of TP
measured in the profundal zone (Z > 9.0 m) of Green-
wood Lake was 0.04 g m-3. Following  stratification
and the onset of anoxic conditions, the mean concen-
tration of TP in the deep water layers  increased to
0.132 g m -3 (Fig. 4 and 5). This represents a 3.3-fold in-
crease in average concentrations. However, concen-
trations as  high as 0.48 g m-3 were measured in the
anoxic hypolimnion. The concentration of phosphorus
associated  with the sediments averaged  186 mg kg -1.
The  redox  potential of  these  sediments was not
measured.  However, judging  from  particle  size
analysis and organic composition data, the redox of
the sediments of Greenwood Lake is probably similar
to that measured for the sediments of Lake Hopa-
tcong.

Internal Phosphorus  Load

Hypsographic  data were used in  conjunction with
dissolved oxygen/temperature profile data, sediment
core data, and measured hypolimnetic TP concentra-
Figure 1.—Relative location of Lake Hopatcong and Green-
wood Lake in northern New Jersey.
 Figure 2.—Seasonal variation in the depth profile of total
 phosphorus concentrations measured at the deep station of
 Lake Hopatcong.
                                                 123

-------
 LAKE AND RESERVOIR MANAGEMENT
tions to quantify the internally generated phosphorus
load for Lake Hopatcong and Greenwood Lake.
  A flux rate was selected from the literature to com-
pute the  daily liberation  of  phosphorus  from  the
sediments   under  anaerobic   condition:!.
Sediment-phosphorus release rates  for lakes with
anoxic hypolimnia range from 0.80 to 96 mg m-2d-1,
but average about  15 mg m-2 d-1 (Nurnberg, 1982;
Holdren and Armstrong,  1980). Particular  attention
was given to phosphorus flux rates for dimictic north
temperate, non-calcareous lakes, with average depths
of about 5 meters. From the available data, it appears
that   under  anaerobic  conditions   a   sedi-
ment-phosphorus flux rate for Lake Hopatcong and
Greenwood Lake of 6 mg m-2 d~1 is realistic.
  In Lake Hopatcong, the  hypolimnion remained
anoxic for 60 days. The anoxic layer had a volume of
2.517 x 1Q6m3 and overlayed 1.65  x 106m2 of lake
bottom. Based on the (assumed) phosphorus release
rate of 6 mg m - 2 d -1, the gross internally regenerated
TP load was predicted to be 595 kg.
  The hypolimnion of Greenwood Lake is more exten-
sive and remains anoxic for a longer period of time
than that of Lake Hopatcong. A total of 3.22 x 106m2
of lake bottom was overlayed  by anoxic water for 90
days. Under these conditions,  a  gross internally
regenerated TP load of 1738.8 kg was predicted.
  The validity of the gross load was tested by compar-
ing it to the net mass of TP in the anoxic hypolimnicn
of each lake. By multiplying the average summer con-
centration of TP measured in the anoxic hypolimnicn
by the volume of the anoxic hypolimnion, a net esli-
                                                   mate of phosphorus mass was obtained.  Following
                                                   stratification, mixing of the hypolimnion is minimal
                                                   and the settling of paniculate material  from the
                                                   epilimnion is inhibited because of density differences
                                                   encountered at  the  thermocline.  Some  of  the
                                                   phosphorus measured  in the hypolimnion  is derived
                                                   from decomposed organic material contributed from
                                                   the trophogenic zone. However, a substantial portion
                                                   of the phosphorus mass of the hypolimnion should
                                                   result from the  regeneration of sediment-bound phos-
                                                   phorus. The phosphorus mass of the hypolimnion thus
                                                   represents the net internally regenerated phosphorus
                                                   load.
                                                    The mean concentration  of TP  measured in the
                                                   anoxic hypolimnion of  Lake Hopatcong  was 0.145 g
                                                   m-3. The volume of the anoxic layer totalled 2.517 x
                                                   106m3. The resulting TP mass associated with the
                                                   anoxic hypolimnion is 365 kg. This net load  compares
                                                   reasonably well with the gross load of 595 kg.
                                                    In Greenwood Lake the summer TP concentrations
                                                   measured in the anoxic hypolimnion averaged 0.132 g
                                                   m-3. The volume of the anoxic layer totalled 11.12 x
                                                   106m3. The TP mass associated with the anoxic hypo-
                                                   limnion of Greenwood Lake is 1,468 kg. This net load
                                                   compares reasonably well with the gross load of 1 739
                                                   kg.

                                                   DISCUSSION

                                                   The  liberation of phosphorus from  anaerobic sedi-
                                                   ments results in the gross contribution of 595 kg to the
                                                   phosphorus budget of Lake Hopatcong, and 1,739 kg
m
m
   450-



   400-


jj.  350-


°  300-
 w

 o
 o
   250-


   200-


   160-

   150-

   140-

   130-

   120-

   110-

   100-

    90-

    80-

    70-
    60-

    50-

    40-

    30-

    20-

    10-
n—r
 CM    (D

 -s.    •—
 00    00
o   >o
^   CM
                                         CM   M

                                         ••-   M
                                                  en   CM
                                                  CM   »-
                                                         00   CM
                                                         CM   *-
'O
CM
o   •*
T-   CM
T

 CM
                                                                                            T"
                                                                                             oo
Figure 3.—Temporal variation in the concentration of total phosphorus monitored at 0.5 m and 14.5 m in Lake Hopatcong.
                                                124

-------
                                                                              INTERNAL NUTRIENT CYCLING
to that of Greenwood Lake (Table 2). In Lake Hopat-
cong, the internally regenerated load represents 15.9
percent of the  annual load, whereas in Greenwood
Lake  it represents 29.3  percent. According to  the
trophic  status  model  of  Dillon (1975), Greenwood
Lake's internal  load is, in itself, greater than the per-
missible phosphorus loading limit for lakes of similar
morphometry and hydrology (Fig. 6).
   For both lakes,  the gross internal load is greater
than the net TP mass. The net load accounts, to an ex-
tent,  for phosphorus loss through  short-term epilim-
netic/hypolimnetic mixing. However, it includes phos-
Figure 4.—Seasonal variation in the depth profile of total
phosphorus concentrations measured at the deep station of
Greenwood Lake.
phorus contributed from the trophogenic zone as a
result of the decomposition of sinking algae, aquatic
macrophytes, tissues, etc. These variables, among
other factors, are probably responsible for the observ-
ed differences  between the net and  gross  internal
phosphorus load. The objective of this study was to
obtain an estimate of the internally regenerated phos-
phorus load which could be used in developing effec-
tive  restoration programs for each  lake. The  gross
load and the net  mass data agree fairly well. There-
fore, for the purposes of this study, the load generated
using the phosphorus release rate of 6 mg m-2d-1 is
considered to be a reasonable estimate of internally
regenerated phosphorus.
  The large difference in the magnitude of the inter-
nally regenerated phosphorus  loads  of Lake Hopat-
cong and  Greenwood  Lake  reflects  their  mor-
phometry. As mentioned previously, Greenwood Lake
is composed of two basins. The deep basin is steep-
sided and its configuration is such that most of it is
deeper than 9 meters. As a result, a large proportion of
the lake's bottom becomes overlayed by anoxic water
following   stratification.  Lake  Hopatcong  is  char-
acterized by a number of large shallow embayments.
In comparison with Greenwood Lake, Lake  Hopa-
tcong's bottom  contour  is  much   more gradually
sloped.
  The proportion of total bottom area associated with
Lake Hopatcong's deeper sections is much  less than
that observed in Greenwood Lake. As compared with
Greenwood Lake, only a relatively small area of Lake
Hopatcong's sediment becomes overlayed by anoxic
water during summer stratification. Thus,  although
the  overall surface area and volume of Lake Hopa-
tcong is greater than that of Greenwood Lake, its con-
figuration  and  morphometry are such that the area
overlayed by its anoxic hypolimnion is less.
  The internally regenerated phosphorus load affects
the productivity and trophic status of Lake Hopatcong
and Greenwood Lake. Although the internal load is an
important  component in  the phosphorus budget of
both lakes, its role in relation to in-lake productivity is
much more substantial in Greenwood Lake.  In both
lakes the boundary of the  anoxic hypolimnion is fairly
close to the thermocline. Storm events probably cause
the temporary erosion of the thermocline and the mix-
ing  of phosphorus-rich hypolimnetic water into the
lakes' trophogenic zone (Kortmann  et al. 1982). This
can contribute to the development or maintenance of
summer algal blooms. Since the anoxic  boundary of
Greenwood Lake extends up to the thermocline, this
 phenomena of epilimnetic/hypolimnetic mixing prob-
 ably occurs more frequently in  this lake than in Lake
 Hopatcong. In  late summer the complete destratifica-
 tion and overturn of both lakes occur within a fairly
            Table 1.—Pertinent hydrologic and morphometric data for Lake Hopatcong and Greenwood Lake.
                                                     Lake Hopatcong
                                                                                      Greenwood Lake
  Surface area
  Volume
  Mean depth
  Maximum depth
  Hydraulic retention time
  Depth of thermocline
  Volume of hypolimnion
  Area overlayed  by hypolimnion
  Period of time hypolimnion is anoxic
  Volume of anoxic hypolimnion
  Area overlayed  by anoxic hypolimnion
1.087 x 10?m2
5.56 x 10?m3
5.5 m
16.7m
623 days
9 m
8.93 x 1Q6m3
2.95 x 106m2
60 days
2.517 x 106m3
1.65 x 106m2
7.77 x 106m2
5.34 x 107m3
5.2m
17.4m
346 days
9 m
1.112 x 107m3
5.46 x 1Q6m2
90 days
1.112 x 107m3
5.46 x 106m2
                                                  125

-------
 LAKE AND RESERVOIR MANAGEMENT
 short period of time: 14 days. This favors the circula-
 tion of the liberated phosphorus into the trophogenic
 zone rather than its  resedimentation. As  has been
 observed in both lakes, this leads to the development
 of an autumnal algal bloom (Fig. 7 and 8).
    For both lakes, stormwater-related nonpoint source
 loads consititute the major component of the annual
 phosphorus budget. In addition, septic contributions
 represent a substantial  fraction  of the  phosphorus
 budget of Lake Hopatcong. The restoration/manage-
 ment plans that have  been developed for these lakes
 prioritize the need to decrease such external  loads
 and set forth measures by which this can be accomp-
 lished. These measures are costly and involve sewer-
 ing, stormwater control, and passive stormwater treat-
                                                  ment. However, for lakes with a substantial internally
                                                  regenerated phosphorus load, an improvement in lake
                                                  status may be delayed following restoration. This has
                                                  been particularly true in those cases where restora-
                                                  tion efforts concentrated  on decreasing the external
                                                  load but did little, if anything, to decrease the internal
                                                  load (Welch and Rock, 1980). The restoration/manage-
                                                  ment plan  developed for each lake should  reflect the
                                                  relative importance of the internal load.
                                                     It is recognized that for both Lake Hopatcong and
                                                  Greenwood Lake the  internal  load  does exert  a
                                                  eutrophying  effect  (Fig.  7 and  8).  However,  the
                                                  magnitude of the internal load and its potential role in
                                                  the eutrophication of these  two waterbodies is much
                                                  greater for Greenwood Lake. Therefore,  more em-
           Table 2.—External and internal total phosphorus loading for Lake Hopatcong and Greenwood Lake.
 Source
 External
   Point sources
   Septic tanks
   Nonpoint source
   Wet/dry fallout on lake
 Internal
   Regeneration of sediment-bound phosphorus
 Total
                                                      Lake Hopatcong
                                                          kgyr-1
                                                                     Greenwood Lake
                                                                         kgyr-1
                                                           165.3
                                                          1600.6
                                                          1616.7
                                                           271.8


                                                           595.0
                                                          4249.4
                                                                           339.2
                                                                           535.0
                                                                          3129.4
                                                                           194.0


                                                                          1738.8
                                                                          5936.4
o
z
o

I
HI
o
U
    450
    400
    350
    300-
160-

150-

140-

130-

120

110-

100-

 90-

 80-

 70-

 60-

 50-

 40-

 30-

 20-

 10-
                                                                             14.5m
                                                                              0.5m
T
 «
                01
                CM
CM   T-
-»   ».
10   CD
                                     T~
                                      o
                                  eg

                                  CO
                                      CM
                                      CM
                                           IO

                                           CO
                                O
                                CM
     00
CM    T-
O
«
                                                                 T"

                                                                  CM
                 T"
                  10
                  CM
i-    CO   0)
CO    »-   CM
                                                   CO
                                                            0>
Figure 5.-Temporal variation in the concentration of total phosphorus monitored at 0.5 m and 14.5 m in Greenwood Lake
                                                  126

-------
                                                                                  INTERNAL NUTRIENT CYCLING
phasis should  be placed on decreasing the internal
phosphorus load of Greenwood Lake than that of Lake
Hopatcong.
  The availability and amount of funding available for
the restoration of lakes has diminished, making it im-
perative that monies be spent as expediently as possi-
ble. Although it would benefit the lake to some degree,
it does not appear cost effective to spend a great deal
                        A - GREENWOOD LAKE. INTERNAL AND EXTERNAL LOAD
                        B - LAKE HOPATCONG, INTERNAL AND EXTERNAL LOAD
                        C - GREENWOOD LAKE, INTERNAL LOAD
                        D - LAKE HOPATCONG. INTERNAL LOAD
                     MEAN DEPTH z (m)
Figure 6.—Trophic state relationship of Lake Hopatcong and
Greenwood Lake.
of time and money in decreasing Lake Hopatcong's in-
ternal load. In relation to other phosphorus sources,
the regenerated load  is not that substantial a compo-
nent of Lake Hopatcong's annual phosphorus budget
(Table 2). For Greenwood Lake, however, the internal
load is a more  important component  of  the phos-
phorus  budget (Table 2). Further, data suggest that
failure  to account for the internal regeneration of
phosphorus  in  Greenwood  Lake could  delay  its
restoration  (Fig. 6).  Therefore,  allocating funds to
decrease the internal regeneration of phosphorus is
more appropriate and should lead to more construc-
tive results for Greenwood Lake  than for Lake Hopat-
cong.
REFERENCES

Armstrong, D.E. 1979. Phosphorus transport across the sedi-
  ment-water interface. Pages 169-175 in Lake Restoration.
  EPA 440/5-79-001. U.S. Environ. Prot. Agency, Washington,
  D.C.
Bannerman, R.T., D.E. Armstrong,  G.C. Holdren,  and  R.F.
  Harris. 1974.  Phosphorus  mobility  in  Lake  Ontario
  sediments (IFYGL). Pages 158-178  in Proc.  17th Conf.
  Great Lakes Res. Int. Ass. Gr. Lakes Res.
Dillon, P.J. 1975. The phosphorus budget of Cameron Lake,
  Ontario: The  importance of flushing  rate  to the degree of
  eutrophying of lakes. Limnol. Oceanogr. 20:28-39.
Fillos, J., and H. Biswas. 1976. Phosphate release and sorp-
  tion by Lake Mohegan sediments. J. Environ. Eng.  Div.,
  Am. Soc.  Civil Eng. 102 (EE2):239-49.
Freedman, P.L, and R.P. Canale. 1977.  Nutrient release from
  anaerobic sediments. J. Environ. Eng. Div. Am. Soc.  Civil
  Eng. 103 (EE2):233-44.
Figure 7.—Temporal changes in chlorophyll a concentra-
tions in Lake Hopatcong at Station LH-2, 0.5 meters.
 Figure 8.—Temporal changes in chlorophyll a concentration
 in Greenwood Lake at Station GL-2, 0.5 meters.
                                                    127

-------
LAKE AND RESERVOIR MANAGEMENT
Holdren, G.C. Jr., and D.E. Armstrong. 1980. Factors affecting
  phosphorus release from intact lake sediment cores. Am.
  Chem. Soc. 14(1):79-87.

Kortmann, R.W., D.D. Henry, A.  Kuether, and S. Kaufman.
  1982. Epilimnetic nutrient loading by metalimnetic erosion
  and resultant algae responses in Lake Waramaug, Conn.
  Hydrobiologia 92:501-10.

Mawson, S.J., H.L Gibbons, Jr., W.H. Funk, and K.E. Hart;:.
  1983. Phosphorus flux rates in lake sediments. J. Water
  Pollut. Control Fed. 55(8):1105-10.
Nurnberg, G.K. 1982. The prediction of internal phosphorus
  load in lakes with anoxic hypolimnia. Limnol. Oceanogr.
  (In press).

Standard Methods for the Examination of Water and Waste-
  water. 1975. 14th ed. Am. Pub. Health Ass., Washington,
  D.C.

Welch, E.B., and C.A. Rock. 1980. Lake Sammamish response
  to wastewater diversion  and increasing urban runoff.
  Water Res. 14:821-28.
                                                     128

-------
THE IMPORTANCE OF SEDIMENT RELEASE IN THE
ASSESSMENT  OF A SHALLOW, EUTROPHIC LAKE
FOR PHOSPHORUS CONTROL
PATRICIA MITCHELL
Water Quality Control Branch
Alberta Environment
Edmonton, Alberta,  Canada


            ABSTRACT

            Complaints of declining water quality and increasing macrophyte growth resulted in a 3-year study
            to determine the feasibility of phosphorus control on recreationally important Lake Wabamun, Alber-
            ta. All nutrient inputs including groundwater were measured or estimated to assess the importance
            of each source. A preliminary phosphorus loading calculation suggested that the sediment may supply
            a large quantity of phosphorus during the annual cycle. Sediment release was estimated using a mass
            balance approach and by phosphorus analysis of sediment cores taken at biweekly intervals. The
            mass balance suggested that release occurred in late summer, and represented a gross input that
            exceeded annual external supplies This was supported by a decline in non-apatite inorganic phosphorus
            in cores during the same time. These results led to the conclusion that major phosphorus control pro-
            jects in the watershed were not warranted.
INTRODUCTION

There is much evidence in the recent literature (Cooke
et al. 1977; Larsen et al. 1981; Jacoby et al. 1982) that
internal loading  can have a major impact on  the
trophic status of lakes. Yet an assessment of this im-
portant component often is not undertaken because
sediment release  is difficult to measure. If lake quality
improvement  through  phosphorus control  is  con-
templated, however, a detailed phosphorus budget
that includes internal loading is essential.
  Lake users have complained for many years about
nuisance macrophyte growth in Lake Wabamun, a
popular recreational lake  60 km west  of Edmonton,
Alberta. The concern that increasing macrophytes
were a symptom  of a general deterioration in water
quality led to a comprehensive 3-year study to deter-
mine whether phosphorus inputs could be controlled
sufficiently to maintain  the  lake's  present trophic
status,  as represented   by  algal  biomass,  total
phosphorus, and  water clarity.
  In spite of public complaints, Lake Wabamun is on-
ly  mildly eutrophic (open water  season averages:
chlorophyll a 12 ^g I-1, total phosphorus (TP) 31 ^g
I-1; Secchi transparency 2.0 m). The lake is shallow,
with the littoral zone (<4 m deep) comprising about 30
percent of its area. Because of its size (see Fig. 1) and
proximity to Edmonton, it  is one of the most popular
lakes in the province, with 1200 cottages, a provincial
park, and several sailing clubs. Sewage is treated by
septic fields, pump-out tanks, or pit toilets.
  A coal-fired power station  on the lakeshore  uses
lake water for cooling. Although eutrophication prob-
lems are often attributed to the heated discharge, it af-
fects less than 2 percent of the lake volume (Nursall et
al. 1972). The  surrounding watershed contains mixed
agriculture, natural poplar-spruce forest, and open pit
coal  mines.
  Initially it was  thought that since the lake is com-
pletely mixed and well oxygenated throughout the
summer,  and  only mildly  eutrophic,  internal loading
would  not be  significant. However, preliminary phos-
phorus budget calculations based on theoretical coef-
ficients (Reynoldson and Hamilton, 1982) suggested
there was more phosphorus in the lake than might be
expected from external loading. Therefore, the study
was directed toward a careful quantification of both
the internal  and  external  loads. Short-term mass
balances and results from phosphorus analysis in
sediment cores collected biweekly demonstrate that
sediment release is important even in a well mixed
lake of moderate productivity.

METHODS

Lake Wabamun was sampled  biweekly through  the
open water season (May-October) in  1980 and 1981
with weighted Tygon tubing that was lowered to 0.5 m
above the bottom. Approximately 20 volume-weighted
sample units from each half of the lake and Moonlight
Bay were combined to yield a composite sample for
each of the three areas; these were subsampled for
analysis of chemical and biological variables. Addi-
tional monthly discrete (1 m interval) samples were ob-
tained  with a Van Dorn sampler at the east and west
sites (Fig. 1) to  identify phosphorus stratification.
Since phosphorus was usually uniform from top to
bottom, phosphorus data from both discrete and com-
posite samples, collected alternate weeks, were used
to calculate the total mass in the lake for each sampl-
ing occasion. Phosphorus was analyzed by standard
methods (Alberta  Environ.,  1977) in 1980, and by a
modified Menzel and Corwin  technique (Prepas and
Rigler, 1982) in 1981. Depth profiles of dissolved ox-
ygen and  other standard  field measurements were
made biweekly with Hydrolab meters.
   Sediment cores were collected monthly in 1980 and
biweekly in 1981 with a multiple corer  (Hamilton et al.
1970) from east and west sites (Fig. 1) during the open
water season. The top 4 centimeters from each of two
cores were combined, freeze-dried  and analyzed for
phosphorus fractions (Alberta Environ., 1977).
                                               129

-------
LAKE AND RESERVOIR MANAGEMENT
   External loading from streams and precipitation
 was  measured  directly.  Approximately  25  inflow
 streams and the lake outlet were sampled for TP con-
 centration  and discharge every 2 or 3 days  during
 spring  snowmelt  runoff  and  weekly in  summer;
 streams were frozen in winter. Loading was calculated
 by integrating the  product of daily stream discharge
 and phosphorus concentration over a 2-week interval.
 Diffuse runoff loading was calculated from  export
 coefficients developed from neighboring subwatet-
 sheds by dividing  the annual load by the subwater-
 shed  area.
   Atmospheric  deposition of  phosphorus was col-
 lected in eight bulk precipitation collectors (Likens et
 al. 1977) around the lakeshore; these  were sampled bi-
 weekly or whenever rain volume warranted. In winter
 larger  collectors  accumulated  snow  until   spring.
 Loading was calculated by applying  volume-weighted
 phosphorus concentrations to meteorological station
 precipitation data  (Hydrol. Branch, pers. comm.).
   Because groundwater  was considered  an  impor-
 tant  source of water to Lake  Wabamun  (Fritz and
 Krouse 1972; Schwartz and Gallup, 1978), a watershed"
 response model was used to determine the long-term
 water and  salinity balance, and thereby to estimate
 groundwater input and outflow (Crowe and Schwartz,
 1982). Phosphorus concentrations were measured  in
 33 domestic wells around the lake (Baldwin 1981), and
 in a new set of wells installed to intercept the theoriz-
 ed aquifer from the northwest (Fritz and Krouse, 1972).
 Since the  input volumes of the two major types of
 groundwater entering the lake were not known, con-
 centrations  could not  be volume-weighted.  The
 highest concentrations (200 ^g M TP) were found  in
 two wells of 16 believed to be  in aquifers that may
 enter the lake. Water from the remaining 14 wells had
 TP concentrations ranging between  3 and 49 ^g 1-1.
 Therefore,  I applied an average TP  concentration  to
 the estimated input volume from the model to obtain a
 phosphorus  load  from groundwater. To  calculate
 phosphorus loss via groundwater I multiplied average
 lake concentration by the ground water outflow  as
 estimated by the watershed model.
   Sewage inputs of phosphorus  were estimated  by
 surveying shallow areas in front of about 250 cottages
 using  a septic leachate detector  ("Septic Snooper,"
 Endeco Inc.). Positive readings associated with high
 fecal conforms in bacteriological samples at nine cot-
 tages  were assumed to represent the proportion  of
 cottages supplying sewage influx; this proportion was
 extrapolated to the total number of cottages on the
 lake. Loading was calculated using a per capita phos-
 phorus output coefficient (0.93 kg cap-1  year-1)
 calculated from analysis of treated municipal sewage
 (Trew et al. 1978), and assuming zero soil retention.
   The  mass  balance phosphorus  budget  was
 calculated for each 2 week period  from May through
 October 1980 and  1981 according to the following
 equation:

     Net gain or loss =  ± mass of TP in lake +
     outflow TP mass  - inflow TP  mass

 A positive balance was considered to be evidence for
 internal loading of phosphorus, a negative balance for
 sedimentation.
RESULTS

The mass of phosphorus  increased in the  lake be-
tween mid-August and mid-October both study years
(Fig. 2).  In 1980 this represented a mass increase of
about 7,000  kg  TP; in 1981 this amounted to about
10,000 kg TP. In 1981, there was also a large increase
in mass between June 2 and June 17, with the concen-
tration  of total  dissolved phosphorus approximately
doubling. Hourly wind data for the Lake Wabamun
area (collected  for TransAlta  Utilities) indicate that
there were strong  northwest winds on June 16, 1981,
the day prior to sampling. It seems likely that wind
played an important role in this dramatic phosphorus
Figure 1—Bathymetric map of Lake Wabamun showing sampling sites and presenting morphometric data.


                                                130

-------
                                                                            INTERNAL NUTRIENT CYCLING
increase, as De Groot (1981) found in a shallow Dutch
lake.
  The seasonal pattern of the external load was dif-
ferent for the 2 study years. The pattern for 1980 was
typical for central Alberta with snowmelt runoff begin-
ning in early April followed by rainy periods in late May
and June. In 1981, snowmelt occurred  in March, and
was below normal, so that by the time the ice left the
lake in late April, external phosphorus loads were very
low. For both years the external load was lowest dur-
ing the late summer and fall. Use of the mass balance
requires accurate measurement of  inflow  and outflow
of phosphorus. For Lake Wabamun, the inputs that
were  not  measured  directly, i.e.,  groundwater and
sewage, comprised 2 percent or less of the total exter-
nal annual input.
  The net gain or loss of phosphorus calculated from
the mass balance  is shown in the upper section of
Figure 2 for each year. Data for the 2  years suggest
that a net gain, or internal loading,  occurs in late sum-
mer when external loads are low. There was  a net loss,
or sedimentation, through June and July 1981 after the
large wind-induced pulse in mid-June.
  I calculated a release rate for the mid-August to
mid-October period by adding net  gains and dividing
by the number of  days. This  amounted to 77.5 kg
day-1 in 1980, and  136.8 kg day-1  in  1981. When
these figures are extrapolated to profundal sediments
(those deeper  than 4 m) the release rates are 1.4 mg
m-2day-1 and 2.4 mg m-2day-1. For the June 1981
pulse, the rate is 5.0 mg m-2 day-1.
  Although  internal  loading  is  evident  in  Lake
Wabamun, the mass balance approach does not iden-
tify the source of the phosphorus increase. Landers
(1982) has  shown  that senescing macrophytes can
provide  large  quantities of  phosphorus to the sur-
rounding water. However, data from Moonlight Bay in
Lake  Wabamun (Fig. 1) suggest  that  macrophyte
senescence did not increase the phosphorus content
of the water. The shallow bay contains a  dense
population  of  macrophytes, yet phosphorus concen-
trations declined in this bay (Fig. 3) over the period of
release  in the main  basin.  Additionally, in the main
basin there was no difference in phosphorus concen-
trations between the shallow, weedy east end of the
lake and the deeper, less littoral west end. Thus, the
sediments appear to be the source of the late summer
phosphorus increase.
   This is also substantiated by phosphorus analysis
of sediment core samples collected biweekly in 1981.
The mobile phosphorus fraction, termed non-apatite
inorganic phosphorus (NAIP) (Williams  et al.  1976),
showed decreases that roughly corresponded to inter-
nal loading as estimated by the mass balance (Fig. 4).
The large decline of NAIP in June relates well to the
wind-induced  release between  June 2 and June 17
mentioned earlier.  There was also a decline in NAIP in
late August,  although concentrations leveled  off in
September and October. I calculated the quantity of
phosphorus  lost  from the  sediments during  the
          I98I
                         ••    •-
                          v/
                                ,-.
                            \
                                                              I98I
                                                     o
                                                     o:
                                                     o
                                                                            TP
                                                                            \
                                                                              CHLOROPHYLL
                                                           	1	1	1	1	1       I
                                                            MAY     JUN     JUL     AUG     SEP    OCT
                                                    Figure 3.—Concentration of total phosphorus and chloro-
                                                    phyll a in Moonlight Bay, Lake Wabamun, 1981.
 Figure 2.—The total  phosphorus mass balance  for Lake
 Wabamun. The upper graph for each year shows net gains
 and losses, the center line shows seasonal changes in the
 mass of total phosphorus and the lower line shows the total
 external load.
 Figure 4.—Concentrations of non-apatite inorganic phos-
 phorus in sediment cores from East and West sites, Lake
 Wabamun, 1981, compared with the mass balance over the
 same time period.
                                                 131

-------
 LAKE AND RESERVOIR MANAGEMENT
 decline  in NAIP in  cores  from  June and  from lale
 August;  release rates ranged between 15 and 25 mg
 m-2 day-1 for the two sites. Assuming  that release
 occurs only from profundal sediments (deeper than 4
 m), the 300 mg m-2 lost  from cores in  June would
 have increased the total mass in the water by 16,800
 kg, or four times that calculated by mass balance. The
 amount  lost in late August was similar, and again ex-
 ceeded mass balance estimates.
   The phosphorus budget  for the open water season
 1980 to 1981 in Lake Wabamun is presented in Table I.
 The  internal  supply was  calculated  by adding net
 gains from the mass balance over each season; sedi-
 mentation was calculated by adding net losses.
   During the open water season, inputs from the sedi-
 ments greatly exceeded external supplies. Even the
 external  supply for the  entire year in  1981  (6,530 kg)
 was  considerably less  than the estimated internal
 supply.
 DISCUSSION

 Results from the mass balance indicate that internal
 phosphorus loading occurs in Lake Wabamun even
 though it  is well-mixed and  mildly eutrophic.  The
 decline in NAIP in the cores in June and August 1981,
 which corresponded to net gains of phosphorus in the
 water column,  suggests that the sediments were the
 source. It is unlikely that phosphorus released from
 macrophyte decay contributed significantly to the in-
 ternal load, since total phosphorus declined in Moon-
 light Bay  during  the period  of macrophyte  sene-
 scence.
   The loss of NAIP in cores represents release rates
 of 15 to 25 mg  m-2 day-1. These rates are based on
 the  time period between sampling  dates; the actuail
 release may have occurred over a much shorter period
 of time, and  hence rates would  be higher. However,
 this rate is within the range reported by other workers
 on shallow lakes: 26 mg m-2 day-1  (DiGiano  and
 Snow, 1977), 14-38 mg m-2 day-1 (Stevens and  Gib-
 son, 1977), 9-47 mg m-2 day-1 (Ryding and Forsberci,
 1977).
   The net release rate based on the mass balance is
 considerably less than the release estimated  fron
 cores. Sedimentation may account for the difference,
  Table 1.—Total phosphorus budget for Lake Wabamun,
        Alberta, May to October 1980 and  1981.
                             1980
                                           1981
Measured External Supply
  Streams
  Diffuse runoff
  Ash lagoon effluent
  Bulk precipitation
  Sewage effluent
  Groundwater
             Total

Internal Supply

Outputs
  Lake outlet
  Groundwater
  Sedimentation
             Total


Change in Mass in Lake
1087 kg
723
719
1210
75
70
3884 kg
5856 kg
718 kg
406
157
1819
75
70
3423 kg
14,306 kg
   49kg
  326
12,898
13,273 kg
-3533kg
+ 7425 kg
                          but  this  could  not be  measured  directly  in  Lake
                          Wabamun because sediment  traps  placed in the
                          deeper areas always contained resuspended material.
                           This discrepancy also could  be explained if there
                          were a thin zone of phosphorus-rich water above the
                          sediments as a result of temporary deoxygenation and
                          subsequent   sediment  release.  Although  biweekly
                          measurements usually showed high dissolved oxygen
                          throughout the water column, deoxygenation  near the
                          bottom was noted on one or two sampling occasions
                          each summer. It is probable that this potential phos-
                          phorus  pool  was  inadequately  sampled,  since
                          samples were collected only to within 0.5 m of the
                          sediments. Additionally, macrophytes  could  have
                          taken up  a portion  of  the phosphorus in the water.
                          Thus, the total mass of phosphorus in the lake was un-
                          doubtedly underestimated.  Core release rates could
                          have been overestimated as well, since there was
                          often a layer of flocculant material overlying the core
                          samples that was easily suspended and hence often
                          discarded during slicing of the cores. This material
                          may have contained NAIP that should have been in-
                          cluded in the core samples.
                           In spite of these sampling problems, it is apparent
                          that  there is  a large reservoir of phosphorus in (or
                          near) the sediments in Lake Wabamun.  This can be
                          made available to algal populations through release
                          and subsequent mixing by wind.
                           The  phosphorus  budget  indicates that  internal
                          loading in the summer can be more important than ex-
                         ternal loading. This was especially true in 1981, when
                         the internal supply was four times the external supply.
                          In light of these  findings, the benefits of controlling
                         watershed sources  of  phosphorus seemed dubious.
                          Even if all of the land presently supporting human ac-
                         tivity in Lake Wabamun's watershed could be returned
                         to a  natural condition,  the external supply would be
                         reduced by only one  third, based  on export from
                         forested subwatersheds around the  lake. My conclu-
                         sion  was that major phosphorus control projects were
                         not warranted for this  lake,  but that lake users and
                         property owners should be educated in land use prac-
                         tices that prevent excessive phosphorus export.

                         ACKNOWLEDGEMENTS: I thank D. Allan, S. Livingstone,
                         and crew for data collection, D. Prosser for assistance with
                         ground water, and D. O. Trew and L Corkum for suggesting
                         beneficial changes to the manuscript.
REFERENCES

Alberta Environment. 1977. Methods manual for chemical
  analysis of water and wastes. Supplement on analysis of
  phosphorus fractions in sediments, by S. Ramamoorthy.
Baldwin, R. 1981. A statistical  analysis of groundwater
  samples in the Wabamun area. Earth Sci. Div., Alberta En-
  viron, (internal rep.).

Cooke, G. D., M. R. McComas, D.W. Waller, and R.H. Ken-
  nedy. 1977. The occurrence of internal phosphorus loading
  in two small, eutrophic, glacial lakes in northeastern Ohio.
  Hydrobiologia 56(2): 129-35.
Crowe, A., and F.W. Schwartz. 1982. The ground water com-
  ponent of the Wabamun Lake Eutrophication Study. Prep.
  Alberta Environ. Water Qual. Control Branch.
DeGroot, W.T. 1981. Phosphate and wind in a shallow lake.
  Arch. Hydrobiol. 91(4): 475-89.
DiGiano, F.S., and P.O. Snow. 1977. Consideration of phos-
  phorus release from sediments in a lake model. Pages
  318-23 in H.L.  Golterman, ed. Interactions Between Sedi-
  ments and Fresh Water. W. Junk Publishers, The  Hague.
                                                 132

-------
Fritz, P., and H.R. Krouse. 1973. Wabamun Lake Past and
  Present: an isotope  study  of the water budget. Symp.
  Lakes of Western  Canada. Univ. Alberta Water Resour.
  Centre 2: 244-58.

Hamilton, A.L, W. Burton, and J.F. Flannagan. 1970. A multi-
  ple corer for sampling profundal benthos. J.  Fish. Res.
  Board Can. 27:1867-69.
Hydrology Branch, Alberta  Environment. 1981, 1982. Pers.
  comm.
Jacoby, J.M., D.D. Lynch, E.B. Welch, and M.S. Perkins. 1982.
  Internal  phosphorus loading in a shallow eutrophic lake.
  Water Res. 16:911-19.
Landers, D.H.  1982. Effects of naturally senescing aquatic
  macrophytes on nutrient chemistry and chlorophyll a of
  surrounding waters. Limnol. Oceanogr. 27(3):428-29.
Larsen, D. P., D.W. Schults and K.W. Malueg.  1981. Summer
  internal phosphorus supplies in Shagawa Lake, Minn. Lim-
  nol.  Oceanogr. 26(4):740-53.
Likens, G.E., et al. 1977. Biogeochemistry of a Forested Eco-
  system. Springer-Verlag, N.Y.
Nursall, J.R., J.B. Nuttall, and P. Fritz. 1972. The effect of
  thermal effluent  in  Lake Wabamun, Alberta. Verh. Int.
  Verein Limnol. 18:269-77.
                           INTERNAL NUTRIENT CYCLING

Prepas, E.E., and F.H. Rigler. 1982. Improvements in quanti-
  fying the phosphorus concentration in lake water. Can. J.
  Fish. Aquat. Sci. 39:822-29.
Reynoldson, T.B., and  H.  Hamilton.  1982. Spatial  hetero-
  geneity in  whole  lake  sediments—towards  a  loading
  estimate. Hydrobiologia 91:235-40.
Ryding, S.O., and C. Forsberg. 1977. Sediments as a nutrient
  source in shallow polluted lakes.  Pages 227-34 in H.L
  Golterman, ed. Interactions Between Sediments and Fresh
  Water. W. Junk Publishers, The Hague.
Schwartz, F.W., and D.N. Gallup.  1978. Some factors con-
  trolling the major ion chemistry of small lakes: examples
  from the Prairie-Parkland of Canada. Hydrobiologia 58(1):
  65-81.
Stevens, R.J., and C.E. Gibson. 1977. Sediment release of
  phosphorus in Lough  Neagh,  Northern Ireland.  Pages
  343-47 in H.L.  Golterman, ed. Interactions Between Sedi-
  ments and Fresh Water. W. Junk Publishers, The Hague.
Trew, D.O., D.J. Beliveau, and E.I. Yonge. 1978. The Baptiste
  Lake Study Summary Report. Water Qual. Control Branch,
  Alberta Environ.
Williams, J.D.H., J.M. Jaquet, and R.L. Thomas. 1976. Forms
  of phosphorus in the sediments of Lake Erie. J. Fish. Res.
  Board Can. 33:413-29.
                                                       133

-------
 LONG TERM EFFECT OF HYPOLIMNETIC AERATION OF LAKES
 AND RESERVOIRS, WITH SPECIAL CONSIDERATION OF DRINKING
 WATER QUALITY AND PREPARATION  COSTS
 BO VERNER
 Atlas Copco Airpower
 Wilrijk, Belgium


             ABSTRACT
             Aeration with the commercially available LIMNO aerator has now been in use for about 10 years. Over
             40 aerator units have been installed and successfully operated in nine different projects. From the
             abundant experience gathered during this relatively long period some examples are chosen to illustrate
             the efforts on increased mineralization capac ty of allochthonous and autochthonous organic material.
             The ubiquitous eutrophication problems especially degrade water quality in lakes and reservoirs. Ex-
             cessively produced algal material in the hypclimnetic layer leads to serious anoxia, preventing higher
             life and aerobic respiration processes.  Furthermore, internal recycling processes connected with the
             metabolic and geochemical cycles of manganese and iron enable the multiple use of recycled
             phosphorus. All of these processes are possible to control by aeration On the other hand, necessary
             treatment for drinking water preparation, such as filtration, flocculation, and chlorination can be reduced
             if a sufficient oxygen balance is maintained in the hypohmnetic water body. Results are presented
             from two drinking water reservoirs and some  lakes where drastic improvements were observed.
The oxygen conditions in lakes, especially in stratified
lakes and reservoirs in temperate climate, are highly
dependent on morphometric properties as the relation
between epilimnetic and hypolimnetic volume and the
trophic status of the lake.
  The trophic status of many lakes has been altered
recently by pollution from organic matter or nutrients,
mainly phosphorus. In manmade magazines, flooded
organic material causes increased oxygen demand.
  The solubility of oxygen in water, equilibrated with
the atmosphere, is rather restricted and amounts to
about 8 to 14 mg/l, depending on temperature and
barometric pressure.  Production processes in  lakes
lead to  supersaturation of oxygen in the epilimnion
and successively to equilibration through the water
surface with the atmosphere. The sedimentation of
the produced algae to the hypolimnion, on the other
hand, reduces oxygen by respiration processes. If the
oxygen  in a limited hypolimnion  becomes depleted,
serious  anoxic  conditions arise. Fermentation pro-
cesses  reduce  both  inorganic  material  as   iron,
manganese, nitrogen compounds, and sulfate, and to
some extent suitable organic matter to form methane.
Higher organisms cannot live in anoxic environments.
The phosphorus  binding capacities of  iron  com-
pounds  in the sediments become depleted since iron
is either dissolved to a larger extent in the divalent
state  or transformed to sulfide, thus releasing large
amounts of phosphorus to the hypolimnetic water.
  To avoid the adverse effects of anoxia, vertical mix-
ing  of the water column was carried out in several
cases, but  led mostly to intensified production pro-
cesses from increased water temperature, a larger
epilimnion, and fertilization  by the direct contact Df
the  productive layer with the sediments.
  Already in the forties Muller in Lake Pfaffiker and in
the  early fifties,  Mercier in Lac du Bret, tried to aerate
the  hypolimnetic layer  without  stratification   with
variable success.
  Bernhardt (1967) reported  successful hypolimnetic
aeration in the Wahnbachtal magazine leading  to
 reduced  iron and manganese concentrations  in the
 hypolimnion.
   Since we considered this design less advantageous
 hydromechanically, in 1969 we designed a straight air-
 lift pump discharging  into a floating basin in which a
 head was created to allow for the return of the water
 down into the hypolimnion. In 1970 Fast tested an air-
 lift design comprising two concentric tubes. Then
 Berhardt made a new design  of  his aerator in 1971.
 Later on several designs were realized or proposed. All
 these aerators have in common that they can aerate
 the water to just the oxygen-saturation value at at-
 mospheric pressure.
   From our experience gained at the two research
 projects in 1969 and 1970 we now wanted to design an
 aerator that
   1. Made use of Henry's law to reach a high oxygen
 transfer efficiency,
   2. Could be operated during the winter stagnation
 period  when an ice-cover is present;
   3. Could be  operated at varying  water  levels of
 several meters as valid for drinking water reservoirs
 without too much attendance;
   4. Was easy to transport by truck and would not de-
 mand special installation equipment;
   5. Was of noncorrosive material.
   These criteria resulted in 1972 in the LIMNO design,
 a submerged aerator. Since that time  aeration of the
 hypolimnion has been carried out in numerous lakes
 and reservoirs and is  frequently used for various  pur-
 poses  in water management  and with varying suc-
 cess.
   Hypolimnetic aeration is the preferable tool, limno-
 logically,  for  increasing  breakdown  efficiency in
 stratified recipients still loaded by either organic mat-
ter or  nutrients. The  redox conditions obtained by
aeration keep the sediment surface oxidized and  pre-
vent the recycling of nutrients  from the sediments to
the water. The limited oxidized layer of the sediment
surface  seals  the  sediment  against transfer of
nutrients through the  interface and increases  phos-
                                               134

-------
                                                                            INTERNAL NUTRIENT CYCLING
phorous sorbing capacity by converting iron sulfide in
the sediments to iron hydroxides.
  The construction of artificial reservoirs, for drinking
water supply or  for the production of hydroelectric
power plants, usually implies the inundation  of top-
soils rich in organic material, thus demanding oxygen.
The lack of oxygen reduces the redox conditions in
these water bodies. Divalent iron and manganese is
dissolved, in several cases  even hydrogen sulfide is
produced. The occurrence of these corrosive and toxic
compounds degrades water quality with respect to
network  corrosion  and increases the preparation
costs for drinking water. Hypolimnetic aeration is the
method of choice in most of these cases. It has now
been in use for several years, and sufficient  experi-
ence has been gained  for evaluating the  efficiency,
benefits, and costs of the various applications.


LAKE GREBIN

The first LIMNO unit was ordered by Ohle in 1972 for a
research project in Lake Grebin. This unit was made of
steel plate  and rather heavy. It was found that the in-
stallation work itself represented a big part of the pro-
ject cost. Later,  to facilitate this work, the unit was
manufactured in  polyester. At about the same time all
accessories were changed to noncorrosive material to
withstand hydrogen sulfide if present in the hypolim-
nion.
   The first projects were in highly eutrophic lakes
which  for  a  long time had  served or still serve as
receivers of wastewater. The aim  with the aeration
was thus to improve the receiver function and to avoid
the stench of hydrogen sulfide (Ohle, 1974).
 LAKE BRUNNSVIKEN

 In 1972, four LIMNO units were installed in lake Brunn-
 sviken in the center of Stockholm. The lake volume is
 about 10 million m3, the area  1.5 km2 and the max-
 imum depth 14 m. The lake quality did not improve dur-
 ing the several years following unloading the waste-
 water. Hydrogen  sulfide concentrations of  over 20
 g/m3 were registered during both the winter and sum-
 mer stagnation periods and the offensive smell when
 more than 20 tons of hydrogen sulfide was released at
 the  circulation periods was  probably  injurious  to
 health.  During 8 years 775 kg of oxygen per day was
 supplied to the lake for 6 months per year. In this way
 a positive oxygen balance was maintained throughout
 the years.
 LAKE KOLBOTN

 Lake Kolbotn, close to Oslo, Norway, is 0.8 km2 and
 has a maximum depth of 24 m. One LIMNO unit was
 installed in 1973 and has maintained aerobic condi-
 tions in the hypolimnion. In 1981 the unit came up to
 the water surface as the anchoring line broke off. Im-
 mediately the situation rebounded to that before the
 aeration started: anaerobic conditions with high con-
 centrations of hydrogen sulfide. The aerator unit was
 put back  into  service  to  continue assisting  the
 receiver function.

 LAGO Dl CALDONAZZO

 Lago di Caldonazzo in Trento,  North Italy, has been
 aerated since 1974 by using six LIMNO units. The lake
is 7 km2 and has a maximum depth of 50 m. The ring-
canalization is unfortunately of such bad quality that
much wastewater still enters the lake. The 2 tons of
oxygen  supplied  to the  lake  by  the  aerators con-
siderably improve the oxygen balance.
  In 1973 two LIMNO units were installed in the 0.7
km2 Wacabuc, N.J., lake. This project was part of a
lake restoration study carried out by Union  Carbide.
After the study the installation was handed over to the
local homeowners association, which  operates the
units each year to avoid anoxia in  the hypolimnion
caused  by diffuse leakage of nutrients into the lake.
THE ENNEPETAL MAGAZINE

The Ennepetalsperre, close to the city of Hagen, Ger-
man Free Republic, with an area of 1.03 km2, a max-
imum depth of 25 m and a volume of 10.6 million m3
serves as a drinking water magazine. Each day, 23,000
m3 of drinking water are delivered, or 8.4 million m3 a
year. In addition to the drinking water at least 50,000
m3 water per day were necessary to ensure the reci-
pient function for the little brook below the magazine.
The runoff area (watershed) of 48  km2 provides about
45 million m3 water per year. The water renewal time is
2 to 3 months.
  The oxygen conditions  in the Ennepetalsperre turn
critical during summer stratification since water in the
tributary is of objectionable quality. The precipitation
area of the Ennepetalsperre, largely used for farming
and  a small town  (pop. about  12,000), discharges
polluted  water from a treatment plant. Until 1973 the
mean phosphorus concentration  entering from the
treatment plant was about 10-15 mg P/l or about 6 ton-
nes of P/year. P-elimination measures in the treatment
plant were able to reduce the amount of P discharged
to about 2.0-2.5 tonnes  P/year. Further  phosphorus
reductions are  planned for the near future and the
goal is to achieve below 0.4 mg P/l.
   To meet  the oxygen  problems, the  first LIMNO
aerator was installed in November 1976. The capacity
of one aerator is about 350 kg O2/day. From the data
supplied from the AVU water supply company one in-
cident of oxygen shortage was reconstructed. Hypo-
limnetic  aeration usually is carried  out between the
middle of June until  the end of September. In 1981 a
short stop of the compressor from the 15 to 18  of
August  led to  oxygen deficiency and a sudden in-
crease in manganese from about 0.07 to 0.3 mg/l. The
oxygen deficiency could not be terminated by turning
on the repaired aeration device. Total circulation was
immediately induced Aug. 27 by bubbling close to the
outlet; this stopped the critical situation and revealed
one of  the  weak points of the installations.  The
capacity of one aerator was found to be insufficient. A
second one was ordered  and installed in November.
   At the same time the old piston compressor for the
first LIMNO unit was replaced by three rotary screw
compressors to be able to operate with  different  ox-
ygenation  capacities for different  storage volumes
 and oxygen demands. The oxygenation capacity can
 be varied between 360 and 1,330  kg oxygen per day.
The effect absorbed by the compressors depends on
 the  water level in the reservoir as the air pressure  re-
 quired is equal to that of the hydrostatic pressure of
 the  LIMNO air diffusor plus some very small pressure
 losses in the air supply lines.
   Aeration was  started  in June 1982,  and oxygen
 records  were taken in the beginning of July. The ox-
 ygen demand in the beginning of about 1.2 tonnes ox-
                                                 135

-------
 LAKE AND RESERVOIR MANAGEMENT
 ygen a day was calculated from successively taken
 oxygen profiles. Since the water intake for the drink-
 ing water works and for the discharge to the brook is
 from about the same strata  as the  aerating device
 close to the bottom of the reservoir, aerated water is
 drawn from the magazine. The water drawn from the
 magazine amounted to about  70,000 m3/day and  con-
 tained between 500 and 550 kg/O2 a day. This means
 that oxygen corresponding to  about 75 percent of the
 total aeration  capacity was transported  from the
 magazine. With both aerators, however, it is possible
 to maintain an oxygen concentration of between 6 and
 8 mg O2/l, with one exception when 2.5 mg O2/l was ob-
 tained.
   The metalimnion had  been  lowered during  this
 period from 6 to 14 m depth, indicating a hypolimnetic
 loss of 2.3 million m3  from July 1 to Sept. 1; this is a
 loss of about 72,000 m3/day. In 1983 the water supply
 to the brook was limited to 30,000 m3/day. This seems
 to maintain a  sufficient  oxygen level  in the water
 around the outlet; however, it proved insufficient  to
 keep oxygen saturation above 20 percent in the whole
 hypolimnion.
   To solve the problem, an  intake tower  with  two
 outlets, one for the drinking water supply at a highar
 level and one close to the bottom for maintaining the
 recipient function in the brook, are needed, or another
 hypolimnetic aerator has to be installed. The alterna-
 tives to aeration are dramatically increased treatment
 costs, since an additional filtration  step (active car-
 bon) would be required to eliminate the fermentation
 products in cases of  anoxic conditions in the hypo-
 limnion. The trophic level of the Ennepetalsperre is so
 high that sulfate reduction would occur immediately
 under anoxia. Destratification  for a short period does
 not solve the problem since it increases the produc-
 tive  layer and  results in increased biomasses  and
 temperatures hardly acceptable  by the users.
   A difficulty believed to be a consequence of hypo-
 limnetic aeration occurred  sometimes  in raw water
 treatment. It is the degassing of the water causing
 partial flotation of  flocks after flocculation  of  raw
 water  in the filtration step. It was  reported  that  in
 cases of total circulation the problem ceased. The pro-
 blem will be met in the future  by  bubbling and de-
 gassing before the filtration step.
   Hypolimnetic aeration  is not used during  winter,
 since at occasions of ice cover, induced circulation
 close to the outlet is needed to  protect the dam.
THE BREITENBACHTAL MAGAZINE
The  Breitenbachtal magazine close to the town of
Siegen, GFR, has served as a drinking water magazine
since 1956; however, increased demands for drinking
water called for an extension of magazine capacity
which was carried out in 1975-79. The magazine today
provides  drinking  water for about 300,000  people. A
drainage area  of  about 11.6 km2 delivers  about 96
million m3 of runoff  water to the magazine. The
volume of the magazine is now 7.8 million m3, the max-
imum depth 37 m, and the maximum intake for drink-
ing water preparation 26,000 m3/day. Before enlarge-
ment, the magazine's  capacity was only 2.6 million
m3 with only one outlet for both drinking water and the
water for the Breitenbach provided 7 m above the bot-
tom of the  magazine. This resulted in  very unstable
water quality with  almost anoxic water on some occa-
sions (Klingebiel and Weinhold, 1980).
   When the extension of the magazine capacity was
 planned, an intake tower with variable outlets and the
 installation of hypolimnetic aeration was considered.
 A LIMNO aerator with a capacity of about 400 kg  O2
 was installed in August 1979. A compressor is also us-
 ed for keeping the dam, the tower, and other installa-
 tions free from ice.
   The records from the operation during 1981 show
 that 166,000 kw hours were needed for aeration, 46,000
 kw hours during winter and 120,000 kw hours during
 the stratification period between June 22 and Nov. 10.
 The energy costs were $8750 (U.S.). The maintenance
 costs amounted to $800 (U.S.) for oil and compressor
 filters and about 130  hours of labor, which can  be
 estimated  at about $1000 (U.S.).  Four million  m3  of
 drinking water  were  produced  from  the Breiten-
 bachsperre magazine during  1981.  The  additional
 treatment cost for the aeration was (considering the
 operation period June 22 to Nov. 11) about $0.005/m3.
   A laboratory report for 1981 stated the onset of the
 summer stagnation at the end of April. Aeration was
 started at June 22 when the oxygen saturation  in the
 bottom layer had decreased to 41 percent. During
 aeration, mean saturation in the bottom layer was  60
 percent.  The   importance  of  aeration   was
 demonstrated during a short maintenance stop of the
 compressor at the end of September, when the oxygen
 saturation decreased within a couple of days from  70
 percent to 20 percent saturation.
   During the whole period of aeration,  manganese
 was present in concentrations of 0.1 mg/l. The inci-
 dent in September, however, increased manganese  to
 0.78 mg Mn/l.  The oxygen demand in the  magazine
 resulted mainly from intensive production processes
 caused by the mass development of the green alga
 Cosmarium during July and August. Up to 88 million
 cells /I were obtained with a chlorophyll content of 75
 mg/m3,  causing oxygen supersaturation  of 185 per-
 cent  and  pH  values  above  10. The  temperature
 stratification was not  disturbed by the hypolimnetic
 aeration.
  Another result of aeration is the nitrification of the
 ammonia in the hypolimnion. About 1 mg NO3 - N was
 present in the  hypolimnion while ammonia  was lack-
 ing.
TEGELER SEE, W. BERLIN.

Tegeler See with an  area of 4 km2 and a maximum
depth of 16 m, is a part of the Havel lake system and is
heavily polluted by the Nordgraben, the recipient for a
large  part  of  the  municipal wastewater  from the
eastern part  of  Berlin. Fishkills  and hypolimnetic
anoxia characterized the conditions until 1980. A large
phosphorus  elimination plant  at  Nordgraben is al-
ready begun. Installation  will  be finished  in 1985;
however, as a first measure for preventing large fish-
kills and infiltration of anoxic hypolimnetic water dur-
ing drinking water preparation, 15  LIMNO units were
installed with an aeration capacity of 4.5 tonnes of ox-
ygen a day.

  The airflow of 1.07 m3/s at 1.8 bar is supplied by two
Atlas Copco oilfree screw compressors mounted in a
container placed at the shoreline. The compressors
absorb  an effect  of  185  kw. The  installation  is
operated about 270 days per year. The installation was
completed in 1980 and the total  project cost was
about $1 million (U.S.).
  The aeration measures in Tegeler See have  been
successful, as far as  oxygenation of the hypolimnion
                                                136

-------
                                                                             INTERNAL NUTRIENT CYCLING
is concerned. The intensive use of the lake for recrea-
tional boat sports required a shortening of the LIMNO
exhaust pipes. The pipes were cut off at a water depth
of about 2 meters. The induced jet effect causes a par-
tial destratification and shortens the stagnation time
considerably. As  a  consequence of this special in-
stallation the productive zone is increased, leading to
increased temperatures in the  lake and intensified
production processes.  One  expected advantage of
aeration treatment, the reduction of nutrients because
of the fixation of phosphorus at the sediment surface,
has not been achieved so far. Measurements revealed
that phosphate occurs  in concentrations of several
mg/l in the hypolimnion and between 0.5 and 1 mg P/l
in the epilimnion. This depends on an underestimation
of the oxygen consumption of the sediment on which
the installation was dimensioned.  Extending the in-
stallation with further units is under discussion. The
Authority of Fisheries, however, reports the return of
salmonid fishes and crayfish to the lake and the re-
colonization of the  bottom sediment by chironomids
assisting in  the decomposition  and mineralization of
the sediment.
SODRA HORKEN (GRANGESBERG
SWEDEN)

Sodra Horken, an oligotrophic lake with an area of 9.1
km2,  is polluted by industrial  and municipal waste-
water entering  the  lake through a  chain of two
stratified basins with 19 m and 14 m depth, respective-
ly, separated by a sill reaching  into the epilimnetic
zone  at about 3 meters. The stratification was stabiliz-
ed by effluents rich in electrolytes, discharged by in-
dustry.  The anoxic  hypolimnetic zone was highly
enriched in phosphorus (0.4 mg P/l), iron (0.2 mg Fe/l),
and ammonia (2.5 mg/l).
  A transport  of concentrations into the main basin
would have had detrimental  effects on  the  oligo-
trophic main basin of the lake. The installation of the
LIMNO aerator implied largely reduced nutrient levels
as 0.05 mg P/l, 0.01 mg Fe/l, and 0.3 mg inorganic N/l.
The  capacity of the aerator could maintain oxygen
concentrations of 15 to 20 mg 02/l. In the case of this
part of Sodra Horken  it seems that even the develop-
ment of filamentous algae was reduced. This was in-
dicated by increased transparency, decreased chloro-
phyll  content, and decreased total phosphorus In the
epilimnion.
   In the case of Sodra Horken, the aim with aeration
was to prevent the passage of nutrient  hypolimnetic
water to the large main basin of the still oligotrophic
lake.  Since the inflow of salt-rich industrial waste-
waster from an ore flotation plant is directly entering
the hypolimnetic zone because of the increased densi-
ty, hypolimnetic water enriched with nutrients would
pass the thresholds to the main basin. Hypolimnetic
aeration enables the  decoupling of electrolyte  and
nutrient transport, thus  saving the main part of the
lake.
 CONCLUSIONS

 We were well aware that the design criteria mentioned
 earlier naturally had to result in a compromise. The
 performance data clearly show this fact. The oxygen
 transfer efficiency  drops  drastically with  the  in-
 creased waterflow pumped for an increased airflow
 discharge.
  Thus a very low airflow discharge should be used if
a high efficiency is wanted, but then the oxygenation
capacity will be low. If not only the operating cost ex-
pressed as kg oxygen per kwh but also the investment
cost for the installation is considered, frequently a
higher oxygenation capacity  at a lower transfer effi-
ciency  is motivated.  Simply, the  minimum of  the
capital and operation  costs for a certain  installation
projected has to be calculated.
  If a LIMNO installation including compressor(s) is
written off over a 10-year period with an interest of 15
percent, for 240 24-hour days per year at an energy
cost of $0.07/kwh, then the total  cost for transferred
amount of oxygen in kg/day varies with the size of the
project. For a small amount (300 kg/day) the cost is
$0.25 (U.S.) and decreases with increasing use so for
the supply of 4t/day the cost is $0.15 (U.S.).
   It should of course also be mentioned regarding the
LIMNO design  that in the late sixties the considera-
tion of the energy cost was not as great as today, and
once our investment for the relatively expensive form
for the manufacturing of the plastic units was made
the performance range was fixed.
   Although we  had  long had  ideas of an improved
design, it was not until 1982 that a prototype, the Flex-
ible LIMNO, was built and tested.  It comprises two
concentric tubes interconnected  with radial walls, all
in  plastic reinforced  fabric,  and  a top cone and cir-
cular ring bottom in polyeten. The unit has two air dif-
fusors, one below the  innertube from which the airflow
generates a waterflow through the unit and a second
diffusor just over the outlets in the outer tube  for in-
creased oxygen transfer efficiency. By a  small throttl-
ing of the water outflow a sufficient pumphead is
created by means of the interconnected walls to pre-
vent the inner tube from collapsing.
   The unit can be built in series up to an oxygenation
capacity of 2 tons/day.
   The investment cost for a Flexible LIMNO is lower
than that of the standard unit. With a higher oxygen
transfer efficiency the operating  cost is also reduced
by more than half.
   With the same calculation of  the cost for the ox-
ygen supplied  as used  for the Standard LIMNO, the
Flexible  unit  costs  $0.06-0.15 (U.S.)/kg  to operate,
depending on the project size.
   The consequences of the lowered investment and
operating  costs for  the  Flexible  Limno can  be il-
 lustrated by a project proposal  made in 1981  for a
drinking water magazine, storage volume 11.2 million
 m3, production 61,000 m3/day. The oxygen required to
 suppress the manganese  concentrations  was
 calculated to 1,800 kg/day.
   The investment cost for six Standard  LIMNOs was
$240,000 (U.S.) and the  yearly operation cost $11,000
(U.S.).  Recalculated for two Flexible LIMNO units the
 corresponding  figures  are $150,000  and $5,000,
 respectively.
   The customer had calculated his investment  cost
 for additional  chemical treatment and  filtration to
 $100,000 (U.S.) and the yearly operation cost for this
 alternative to $70,000 (U.S.). Thus  for Standard units
 break-even for the two alternative methods was reach-
 ed in 3 years. For the Flexible LIMNO this time  is less
 than 1 year.
   The experiences with hypolimnetic aeration  show
 that this method can be used for different purposes,
 as long  as  the preconditions  are known and the
 mechanisms  clear in detail.  General recommenda-
 tions  should always be  complemented by special
 studies of the lake or magazine in question. The costs
                                                 137

-------
LAKE AND RESERVOIR MANAGEMENT

for treatment can be classified as moderate in com-
parison to other measures, as for example the use of
pure oxygen or ozone.
   The argument of lethal effects to fish from enrich-
ment of dissolved molecular nitrogen in the water at
the hydrostatic pressure in the level of the installed
equipment seems to be of a hypothetical character. In
the present study no effect on  fish was observed,
which also seems plausible when the movement pat-
tern of fish is considered. Dissolved gas accumulation
always occurs at the sediment water interface from in-
tensified  respiration and  denitrification processes
during stagnation. Fish, however, were observed in
echograms in the hypolimnion which would not have
been the case under prevailing anoxic conditions.
REFERENCES

Bernhardt, H. 1967. Aeration of Wahnbach Reservoir without
  changing the temperature profile. J. Am. Water Works Ass
  59 (8): 943-64.
Fast, A.W. 1971. The effect of artificial aeration on lake
  ecology. Ph.D. Thesis. Michigan State Univ., Ann Arbor.
Klingebiel, G., and R. Weinhold. 1980. Die aufstockung der
  Breitenbachtal Speere. Wasserwirtschaft Heft 3.
Ohle, W. 1974. Typical steps in the changes of a limnetic eco-
  system by treatment with therapeutica. Verh. Int. Ver. Lim-
  nol.  19: 1250.
                                                 138

-------
                                                 Biomanipulation
THE INTERACTIONS AMONG DISSOLVED ORGANIC MATTER,

BACTERIA, SUSPENDED SEDIMENTS, AND ZOOPLANKTON


JOSEPH A. ARRUDA
Division of Environment
Kansas Department of Health and Environment
Topeka, Kansas


G.R. MARZOLF
Division of Biology
Kansas State University
Manhattan, Kansas


           ABSTRACT

           Lakes and reservoirs in agricultural watersheds are often turbid with suspended sediments. The
           interactions among suspended sediments, dissolved organic matter, bacteria, algae  and
           zooplankton will  continually modify water quality. Our object was to evaluate the effects of
           dissolved organic matter (DOM) source, bacteria, and suspended sediments on the survival and
           growth in body length of Daphnia pulex. We grew Daphnia in vitro with different DOM sources
           (bur oak and hackberry leaf leachate), with and without suspended sediments. Daphnids grown
           in hackberry leaf leachate, with or without sediments, grew more than those in the correspon-
           ding bur oak leaf leachate treatments (P < 0.01). Bacterial  density also was higher in the
           hackberry treatments compared to the bur oak treatments. Daphnids grew more in treatments
           with DOM plus suspended sediments than in the corresponding treatments without suspended
           sediments (P <  0.01), although bacterial density was lower in treatments with suspended
           sediments. Suspensions without daphnids had lower bacterial densities  than those with
           daphnids. This experiment reveals some of the complexity of the interactions among the
           sediments, DOM, bacteria, and zooplankton that influence water quality. The results may have
           some applications to lake and reservoir management, particularly the biomanipulation of water
           and wastewater.
INTRODUCTION

Many lakes and reservoirs in agricultural watersheds
receive large loadings of suspended sediments. It is
likely that suspended sediments influence  the in-
tricate relationships among dissolved organic matter,
algae, bacteria, and filter-feeding zooplankton. These
relationships are fundamental and have implications
for the management of  water quality in lakes and
reservoirs as well as for the treatment of water and
wastewater.
   Suspended sediments reduce available light and so
reduce primary production (Stern and Stickle, 1978;
Iwamoto et al. 1978; Kimmel and Lind, 1972; Marzolf
and Osborne, 1972). Production is diminished despite
the elevated nutrient levels often associated with tur-
bid inflows  (Stern  and Stickle,  1978; Paerl  and
Goldman, 1972).  Flocculation of phytoplankton and
sediments also may occur, leading to higher sedimen-
tation rates of the phytoplankton-sediment floe (Av-
nimelech et al. 1982).
  The interactions between suspended sediments
and bacteria involve ambient nutrient and suspended
sediment  concentrations,  the  nutrient  adsorbing
capacity of  the  suspended sediments,  bacterial
metabolic rates, and extra-cellular enzymatic activity
                                             139

-------
LAKE AND RESERVOIR MANAGEMENT
 (Ensminger and Gieseking, 1942; Jannasch and Prit-
 chard, 1972; Haska, 1975, 1981; Cammen, 1982).
   The  effects of  suspended  sediments  on  filter
 feeding zooplankton are not well understood (Marzolf
 and Arruda,  1981). The physical  presence of the
 sediments reduces the ingestion rates of  algae by
 zooplankton, probably to starvation levels (Sherk et al.
 1976; Arruda et al. in press). However, it has been sug-
 gested  that zooplankton in turbid waters may adjust
 their ingestion rates to maximize their energy intake at
 ambient suspended sediment concentrations (Sherk
 et al. 1976).
   Suspended  sediments adsorb dissolved materials
 from the aquatic environment (see reviews cited; Car-
 ritt and Goodgal, 1954; Bader et al. 1960; Morris and
 Calvert, 1975; Tanoue and Handa, 1979; Logan, 1982).
 Adsorbed organic matter of suitable nutritional value
 can be used  for growth  by  suspension-feeding
 zooplankton (Marzolf, 1981; Arruda et al. in  press).
   Dissolved organic matter in lakes and reservoirs is
 derived from many sources. The major inputs include
 the watershed (riparian vegetation) and the  lake  itself
 (algal and microbial extracelluar releases; Cole,  1982,
 and faunal excretion; Scavia and Gardner,  1982). We
 expect that dissolved organic matter quality is altered
 continuously  within lakes and that the  diversity of
 molecular species  is probably very high.
   We report the results of an experiment testing the
 interactions among these four components: suspend-
 ed sediments, dissolved organic matter, bacteria, and
 zooplankton. Specifically, our objective was to deter-
 mine the effects of the interactions among dissolved
 organic  matter source,  bacteria, and  suspended
 sediments  on  the  early growth  of Daphnia pulex, a
 common suspension-feeding zooplankter.
MATERIALS AND METHODS

Experimental Design and Analysis

We chose  dissolved organic matter (DOM) leached
from leaves of hackberry (Celtis occidentalism and bur
oak  (Quercus macrocarpa) trees  because they are
common riparian trees  in  Kansas watersheds and
readily  available. DOM  treatments with both leaf
leachate sources  were  used,  with  and  without
suspended  sediments. Three control treatments were
used: the  growth medium  without food  (starved),
suspended  sediments without food (raw sediments),
and the alga Chlorella vulgaris as a known high quali-
ty food  source. Daphnia pulex < 24 hours old were
pooled  and   assigned   randomly  to  the   seven
treatments. Each treatment  included six replicates. A
replicate consisted of one daphnid in 100  ml  of
feeding  suspension  in a 150 ml beaker. Each day the
feeding  suspensions were renewed and the daphnids'
body lengths measured. The experiment continued for
9 days. Data were analyzed with multiway analysis of
variance and covariance and least squares  means
separation procedures (Helwig and Council, 1979).

Preparation of Feeding  Suspensions

Dried and milled reservoir sediments (Arruda et al. in
press) were resuspended (100 mg to 1  I of distilled
water) and  allowed to settle for 90 minutes in a 1 I
graduated cylinder. The top  500 ml was siphoned out
and stored  as a stock suspension (2.65 mg/ml) at 3°
Celsius.  The  final  concentration  used  in the
treatments with suspended sediments was 44 mg/l.
  Dissolved organic matter was  leached  from air-
dried pre-abcission leaves of hackberry and bur oak.
 Five grams of hackberry leaves were leached in 250 ml
 of distilled water and 10 grams of bur oak leaves were
 leached  in 500 ml, both for 24 hours. The DOM was
 filtered through a series of filters (qualitative glass
 fiber, 0.45 and 0.2 /^m membrane filters) and stored at
 3° Celsius. Hackberry DOM contained  about 1.2 mg
 C/ml and the bur oak about 0.6 mg C/ml (McArthur, un-
 publ.) Each DOM source was used at a final concen-
 tration of 25  mg C/l.
  The alga,  Chlorella vulgaris, was taken  from  our
 stock  cultures grown in Chlorella medium  (Starr,
 1978). A  previously determined relationship between
 cell number and optical  density was used to produce
 the final  concentration of 1 x 106 cells/ml. After adding
 the appropriate stock  suspensions to the diluent
 medium  (Buchanan et al. 1982), the pH  was adjusted
 to pH 8.0.
 Bacterial Enumeration

 Bacteria were observed with  epifluorescence micro-
 scopy after staining with DAPI (Coleman, 1980; Porter
 and Feig, 1980). One beaker was chosen randomly
 from each treatment and 10 ml of the suspension was
 preserved with 0.1 ml of formalin. Later, 1.0 ml of the
 preserved suspension was placed in a filterholder with
 1.0 ml of filtered distilled water and 0.2 ml of filtered
 DAPI  over a 0.2 ^m pore size black stained membrane
 filter (Hobbie et al. 1977). The stain and sample were
 incubated in the filter holder 5 minutes before filtering.
 One filtered sample and  10 random fields per filter
 were counted for each treatment.
 RESULTS

 Although  suspensions  were  not  inoculated with
 bacteria, there was some  bacterial  growth in  all
 treatments. The bacteria were large rods, except for
 small coccoid bacteria (or small rods) in the bur oak
 treatments. The lowest bacterial densities were found
 in the bur oak/sediment treatment and raw sediment
 treatment, with slightly more in the starved control
 (Table 1). Bacterial density was higher in the hackberry
 DOM treatments than in the bur oak treatments, and
 lower in the corresponding treatments with suspend-
 ed sediments. Bacterial density usually increased dur-
 ing each 24-hour feeding period (Table 1).
  Suspended sediments interacted with the presence
 of Daphnia to modify bacterial density in the DOM
 treatments. In beakers  without daphnids, bacterial
 density was higher if suspended sediments were pre-
 sent (Table  1).  With  daphnids   present,  however,
 bacterial density  was  higher  without suspended
 sediments.
  Bacterial  density  was  determined  in  beakers
 without daphnids to see whether density was affected
 by daphnid grazing. Except for the  starved treatment,
 bacterial densities were higher in the beakers with
 daphnids (Table 1),  regardless of any  assumed loss to
 grazing. The increase was greatest in both hackberry
(901-fold) and bur oak  (168-fold) treatments without
 suspended sediments.
  Survivorship curves (Fig. 1) indicate early and rapid
 mortality of daphnids  in the  starved and  bur  oak
treatments. Survival of daphnids in the Chlorella treat-
 ment was intermediate to the other  four treatments,
 where the daphnids were longer-lived. When the  ex-
 periment was terminated at the end of the ninth day,
there were daphnids  surviving  in all but the  starved
treatment, with fewest in the bur oak treatment.
                                                140

-------
                                                                                        BIOMANIPULATION
  Table 1.—Numbers of bacterial cells (10s cells/ml) in feeding suspensions at the start and end of the feeding period, and
                    without daphnids at day 5. Bacterial density < 104 cells/ml is not detectable.

Treatment Time
Starved s
e
Raw sediment s
e
Chlorella s
e
Hackberry s
e
Hackberry/sediment s
e
Bur oak s
e
Bur oak/sediment s
e
s = start
e = end
5n = day 5 without daphnids
nd = not detectable

100 «(- — • - — •<— • 	 • — iffil— fJB —
\ \\ ^\ \ \
c \cs — eg]— H— '[B — Rf^l— f
S~ \ \ \
_ B-X- B c 	 c 	 c 	
^ ' \\
> 50 \\
oc 50 s\\
D \ \
- \\
- ; \\-B-
\
_ 	 s
Day
1 2 3 5 5n
1.35 0.90 0.32 nd —
6.16 0.60 1.18 nd 0.51
nd 0.60 nd nd —
nd nd nd 0.46 nd
4.06 1.30 6.59 4.72 -
8.97 3.46 10.8 14.7 11.4
1.30 2.68 2.53 0.16 —
19.9 13.3 132. 90.1 0.10
nd nd 2.68 nd —
nd 3.78 16.9 17.1 7.52
1.15 2.58 1.48 0.16 -
15.3 9.82 1.08 43.7 0.26
nd nd nd nd —
nd nd nd 4.52 0.31





1
\
?T=.f
\
C \R
\\
Pfffl
V|rl|


B 	 B


Initial daphnid growth was greatest in the Chlorella
treatment (P < 0.01). Daphnids in the raw sediment
treatment grew less than those in the starved treat-
ment, but they survived longer (Figs. 1 and 2). Initial
growth was mostly similar for the other treatments,
but by day 5, daphnids grown in hackberry DOM were
longer than in the corresponding bur oak DOM
treatments, with or without suspended sediments (P <
0.01). Daphnids were longer in treatments with DOM
and suspended sediments than in the corresponding
treatments without suspended sediments (P < 0.01).

DISCUSSION
          1    23456789

                      TIME  (days)
Figure 1.—Survivorship  curves of  Daphnia pulex  in these
treatments: starved  (S), Chlorella  (C), raw sediments  (Ft),
hackberry leaf leachate without (H) and with sediments  [Hj,
and bur oak leaf leachate without (B) and with sediments Tgl.
    20




 1
 J


 I 15

 O

 UJ
Q
O
CO
    to
    05
       f
            1   23456789
                    TIME (days)
Figure 2.—Growth in body length of Daphnia pulex. Under-
lined symbols join treatments with significance levels P >
0.05. Asterisks denote a significance level of P <0.05 for the
difference between adjacent treatments on the same day.
Legend as in Figure 1.
The growth and survivorship of Daphnia pulex depend-
ed on the interactions among dissolved organic mat-
ter  source,  bacterial  density,  and  suspended
sediments.  Indeed, under these experimental condi-
tions, suspended sediments enhanced the growth and
survival of daphnids in the dissolved organic  matter
treatments. Bacterial density apparently was increas-
ed  by daphnid grazing and influenced by dissolved
organic matter quality and the presence of suspended
sediments.
  These heterotrophic food chains may be significant
factors in the determination of water quality in certain
lakes, reservoirs or treatment ponds. However, the im-
plications of our results for water quality management
are not explicit. This is especially true for biomanipu-
lation (Shapiro et al. 1982; Tourbier and Pierson, 1976),
the conscious alteration of  biotic ecosystem com-
ponents  to achieve  a  particular result.  We  can,
however, speculate after summarizing the water quali-
ty implications of  this research into two categories.
The first category  involves wastewater effluents and
artificial food  chains  and the second involves lakes
and reservoirs that are used  as sources for drinking
supplies or are in need of management.
  Daphnia  are  able  to grow  in  sewage  or  waste
stabilization ponds (Daborn et al. 1978; Dinges, 1974;
Mitchell, 1980; Mitchell and Williams, 1982; Myrand
and de la Noue,  1983).  Data from the field are needed
on dissolved and particulate organic matter (bacteria
and algae)  and their  relationships to  subsequent
zooplankton growth or harvesting. These data will
allow more accurate  predictions of the  outcome  of
biomanipulations based on the interactions seen  in
the data presented here.
                                                   141

-------
 LAKE AND RESERVOIR MANAGEMENT
   Seeding of treatment ponds with bacteria, alg&e,
 and zooplankton (to achieve a desired effect) may con-
 trol the production of zooplankton and so enhance the
 possibilities for the biomanipulation of these artificial
 ecosystems  (Gordon  et al.  1982;  Mitchell,  1960;
 Goldman and Ryther, 1976; McShan et al. 1974; Pavoni
 et al. 1974; Uhlmann,  1971;  and Kryutchkova, 1963).
 Further, it may be possible that new genetic strains of
 bacteria capable of metabolizing particular substrates
 could be developed.
   The goals  of the biomanipulation of treated e;f-
 fluents  or small lakes  and reservoirs need to be
 recognized. Biomanipulations for short-term and long-
 term resource uses will differ. Further, the manipula-
 tion of some algal or  bacterial populations may be
 more  effective by using chemical or physical means
 than by biomanipulation of higher trophic levels. In the
 context of nutrient cycling, it may be feasible to direct
 the fate of dissolved nutrients into components of the
 biota  that:  (1) turn over nutrients more slowly (longer
 than seasonal time scales) and so might operate as
 nutrient sinks (for example, encouraging phosphorus
 into fish rather than macrophytes or algae) (Kitchell et
 al. 1979), or (2) place nutrients into biota that  are more
 useful as an immediate harvestable yield.
   Lakes  and  reservoirs  used  as water storage  for
 drinking water supplies may produce the precursors
 for  trihalomethanes (THM)  that  are formed  upon
 chlorination. It is clear that dissolved and particulale
 organic matter quality can affect and be affected by
 bacterial   growth,  zooplankton,  and  suspended
 sediments. It  is  possible that  these three factors
 might be used alone, or in some combination, to alter
 organic matter. The result could be the modification of
 THM precursors that will reduce THM formation dur-
 ing subsequent chlorination. Although  treatment to
 remove the potential for THM formation is available,
 the possibility remains that  biomanipulation may be
 able to remove or modify the undesirable components.
 If performed  as  part of  a broader scheme of lake
 management,  such  treatment  might   make good
 economic sense.
   Dissolved  organic  carbon  and   suspended
 sediments are key components  in the transformations
 described  here.  The  concentrations of suspended
 sediments  and  dissolved organic matter,  and  the
 quality of  the  dissolved organic  matter  will va'y
 through time  due to changes  in land   use, riparian
 vegetation, hydrologic events in the watershed, or the
 extent of wastewater  treatment. There is  little  ex-
 perimental  or descriptive work on the trophic  relations
 encompassing the dissolved organic-phytoplankton-
 bacteria-zooplankton food chain (Graham and Canale,
 1982), and  none, especiajly, that considers the in-
 fluence of suspended sediments. Our understanding
 of these  synergisms will lead to more accurate lake
 models for application to the goals  of water  manage-
 ment through biomanipulation.

 ACKNOWLEDGEMENTS: Financial support was provided by
 National  Science Foundation Grant DEB-8207214 to G.R.
 Marzolf and  J.A. Arruda. This is  contribution 84-56-J of the
 Kansas Agricultural  Experiment  Station,  Kansas  State
 University, Manhattan. We thank J. V.  McArthur and D.L.
 Smith for helpful discussions and R.T. Faulk for a supply of
 healthy Daphnia. J.V. McArthur,  D.L Smith, and S.M. Arruda
 reviewed the manuscript.
REFERENCES

Arruda, J.A., G.R. Marzolf, and R.T. Faulk. In press. The role of
  suspended sediments in the nutrition of zooplankton in turbid
  reservoirs. Ecology.
 Avnimelech, Y.,  B.W. Troeger, and L.W. Reed.  1982. Mutual
   flocculation of algae  and  clay: evidence and implication.
   Science 216:63.

 Bader, R.G, D.W. Hood, and J.B. Smith. 1960. Recovery of dis-
   solved organic matter  in sea water and organic sorption by
   particulate material. Geochim. Cosmochim. Acta 19:236.
 Buchanan, C., B. Goldberg, and R. McCartney. 1982. A labor-
   atory method  for studying zooplankton swimming behavior
   Hydrobiologia 94:77.

 Cammen, LM. 1982. Effect of particle size on organic content
   and microbial abundance within four marine sediments. Mar
   Ecol. Prog. Ser. 9:273.
 Carrit, D.E., and S. Goodgal. 1954. Sorption reactions and some
   ecological implications. Deep Sea Res. 1:224.
 Cole, J.J. 1982.  Interactions between bacteria and algae in
   aquatic ecosystems. Ann. Rev. Ecol. Syst. 13:291.
 Coleman, A.W. 1980. Enhanced detection of bacteria in natural
   environments  by  fluorochrome  staining  of DNA. Limnol.
   Oceanogr. 25:948.

 Daborn, G.R., J.A. Hayward, and T.E. Quinney. 1978. Studies on
   Daphnia pulex in sewage oxidation ponds. Can  J Zool
   56:1392.

 Dinges, R. 1974. The availability of Daphnia for water quality im-
   provement and as an animal food source. Pages 142-161 in
   Proc. Conference on Wastewater Use in the Production of
   Food and Fibre. EPA 600/2-74041. U.S. Government Print. Off.,
   Washington, D.C.

 Ensminger, L.E., and J.E. Gieseking. 1942. Resistance of clay
   adsorbed proteins to proteolytic hydrolysis. Soil Sci. 53:205.
 Goldman, J.C., and J.H. Ryther. 1976. Waste reclamation in an
   integrated food chain  system. Pages 197-214 in J. Tourbier
   and R.W. Pierson eds. Biological Control of Water Pollution.
   Univ. Pennsylvania Press, Philadelphia.
 Gordon, M.S., et al. 1982. Aquacultural approaches to recycling
   of dissolved nutrients in secondarily treated domestic waste-
   waters.  IV. Conclusions, design and operational considera-
   tions for artificial food chains. Water Res. 16:67.
 Graham, J.M., and P.P. Canale. 1982. Experimental and model-
   ing studies of a four  trophic level predator-prey system.
   Microb.  Ecol. 8:217.

 Haska, G. 1975.  Influence of clay minerals on sorption of
   bacteriolytic enzymes.  Microb. Ecol. 1:234.

       .. 1981. Activity of bacteriolytic enzymes  adsorbed to
  clays. Microb. Ecol. 7:331.
Helwig, J.T., and K.A.  Council. 1979. SAS User's Guide. 1979
  ed. SAS Institute Inc., Raleigh, N.C.
Hobbie, J.E., R.J. Daley, and S. Jasper. 1977. Use of nuclepore
  filters for counting  bacteria by fluorescence microscopy.
  Appl. Environ. Microbiol. 33:1225.
Iwamoto, R.N., E.O. Salo, M.A. Madej, and R.L McComas. 1978.
  Sediment and water quality: A review of the literature in-
  cluding a suggested approach for water quality criteria with
  summary and conclusions and recommendations. EPA
  910/9-78-048. U.S. Environ. Prot. Agency Washington, D.C.
Jannasch,  H.W., and P.M.  Pritchard.  1972. The role of inert
  particulate matter in the activity of aquatic microorganisms.
  Mem. 1st. Ital.  Idrobiol. 29 (Suppl.):289.
Kimmel, B.L, and O.T. Lind. 1972. Factors effecting phytoplank-
  ton production in a eutrophic  reservoir. Arch. Hydrobiol.
  71:124.

Kitchell,  J.F., et al. 1979.  Consumer  regulation of nutrient
  cycling. Bioscience 29:28.
Kryutchkova, N.M. 1968. The role of zooplankton on the self-
  purification of water  bodies. Hydrobiologia 31:585.
Logan, T.J. 1982. Mechanisms for release of sediment-bound
  phosphate to water and the effects  of agricultural land
  management on fluvial transport of particulate and dissolved
  phosphate. Hydrobiologia 92:519.

Marzolf, G.R. 1981. Some aspects of zooplankton existence in
  surface  water  impoundments.  Pages  1392-1399  in  H.G.
  Stefan, ed. Surface Water Impoundments. Am. Soc. Civil Ena
  New York.
                                                     142

-------
                                                                                                      BIOMANIPULATION
Marzolf, G.R., and J.A. Arruda. 1981. Roles of material exported
  by rivers into reservoirs. Pages 53-55 in Restoration of Lakes
  and Inland Waters. Proc. Sym. EPA 440/561-010. U.S. Environ.
  Prot. Agency, Washington, D.C.
Marzolf, G.R., and J.A. Osborne. 1972. Primary production in a
  Great Plains reservoir. Ver. Int. Vereing. Limnol. 18:126.
McArthur, J.V. 1982. Pers. Comm.  Kansas State Univ., Manhat-
   tan.
McShan, M., N.M. Trieff, and D. Grajcer. 1974. Biological treat-
  ment of wastewater using algae and Anemia. J. Water Pollut.
  Control Fed. 46:1742.
Mitchell, B.D. 1980. Waste stabilization ponds. Pages 360-375 in
  W.D. Williams, ed. An Ecological Basis for Water Resource
  Management. A.N.U. Press, Canberra, Australia.
Mitchell, B.D., and W.D. Williams. 1982. Factors influencing the
  seasonal occurence and abundance of the zooplankton  in
  two waste stabilization ponds. Aust. J. Mar.  Freshw.  Res.
  33:989.
Morris, R.J., and S.E. Calvert. 1975. Fatty acid uptake by marine
  sediment particles. Geochim. Cosmochim. Acta 39:377.
Myrand, B., and J. de la Noue. 1983. Ingestion and assimilation
  rates of Oocystis sp. by Daphnia magna  in treated waste-
  waters. Environ. Pollut. (Series A) 31:77.
Paerl, H.W., and C.R. Goldman. 1972. Stimulation of hetero-
  trophic and autotrophic activities of a planktonic microbial
  community by siltation at Lake Tahoe, Calif. Mem. 1st. Ital.
  Idrobiol. 29(Suppl.):129.
Pavoni, J.L, S.W. Keiber, and G.T.  Boblitt. 1974. The harvesting
  of algae as a food source from wastewater using natural and
  induced flocculation techniques. Pages 435-496 in Proc. Con-
  ference on Wastewater Use in the Production of  Food and
  Fibre.  EPA  600/2-74-041. U.S.  Government  Print.   Off.,
  Washington, DC.
Porter,  K.G., and Y.S.  Feig.  1980.  The  use of  DAPI  for
  identifying  and   counting  aquatic  microflora.  Limnol.
  Oceanogr. 25:943.
Scavia,  D., and W.S. Gardner. 1982.  Kinetics of nitrogen and
  phosphorus release in varying food supplies  by Daphnia
  magna. Hydrobiologia 96:105.

Shapiro, J., et al. 1982. Experiments  and experiences in bio-
  manipulation—studies of  biological ways to reduce algal
  abundance and eliminate blue-greens. EPA-600/3-82-096.  In-
  terim Rep. No. 19.  Limnol. Res. Center, Univ. Minnesota, Min-
  neapolis.

Sherk, J.A., Jr., J.M. O'Conner, and D.A. Neumann. 1976. Effects
  of suspended solids on  selected estuarine plankton. Misc.
  Rep.  No. 76-1. U.S. Army Coastal Eng. Res. Center, Fort
  Belvoir, Va.
Starr, R.C. 1978. The culture collection of algae at the University
  of Texas at Austin. J. Phycol. 14(Suppl.):47.

Stern, E.M., and  W.B. Stickle.  1978. Effects of turbidity and
  suspended  material in   aquatic  environments,   literature
  review. Tech. Rep.  D-78-21, June 1978. Army Eng. Waterways
  Exp. Sta. Vicksburg, Miss.
Tanoue, E., and N. Handa. 1979. Differential sorption of organic
  matter by various sized sediment particles in recent sediment
  from the Bering Sea. J. Oceanogr. Soc. Japan 35:199.

Tourbier, J., and R.W. Pierson, Jr. 1976. Biological Control of
  Water Pollution. Univ. Pennsylvania Press, Philadelphia.
Uhlmann, D. 1971. Influence of dilution, sinking, and grazing rate
  on  phytoplankton  populations of hyper-fertilized ponds and
  micro-ecosystems. Mitt. Int. Ver. Limnol. 19:100.
                                                          143

-------
 LONG TERM GRAZING CONTROL  OF ALGAL
 ABUNDANCE:  A CASE HISTORY
 RICHARD A. OSGOOD
 Metropolitan Council
 St.  Paul, Minnesota
             ABSTRACT
             Square Lake, Minn., mimics an ideal of trie biomanipulation approach. The lake's zooplankton
             community is dominated by large (body si;:e) herbivorous cladocerans (Daphnia pulicaria) whose
             collective grazing abilities have maintained a reduced standing crop of algae, apparently since
             1926. The physical and chemical environmental factors that define Daphnia's niche in Square
             Lake are discussed and include stable stratification, a large hypolimnetic volume, and
             phosphorus concentrations within certain threshold limits. These environmental  limitations
             seem to be generally important for providing a metalimnetic refuge for Dephnia. Future biotic
             reconstructions that aim to encourage large-bodied Daphnia should consider these environmen-
             tal limitations.
 INTRODUCTION

 Reduced algal biomass is the principal goal of lake
 restoration.  This is commonly  achieved through
 nutrient reductions (Chapra and Robertson, 1977; Eid-
 mondson and Lehman, 1981). Sometimes, however,
 this  approach is unsuccessful because of ecologic
 and  environmental  complexities  in  lakes (Shapf-o,
 1979),  and  attempts at  controlling algal  biomass
 through nutrient management fail (Larsen et al. 1981).
 It may be possible  in these cases to control  algal
 biomass by manipulations or reconstructions of par-
 ticular trophic associations in lakes; perhaps without
 prior nutrient reduction (Shapiro et al. 1982; Carlson
 and Schoenberg, 1983), a process termed biomanipu-
 lation.
  Square Lake, Minn.,  mimics an ideal of the bo-
 manipulation approach (Shapiro  et al. 1982).  Tie
 lake's zooplankton community is dominated by large
 (body  size)  herbivorous  cladocerans  (Daphnia
pulicaria) whose collective grazing abilities maintain a
 reduced standing crop of algae. As  well as can be
 determined, Square Lake has existed in this condition
 for many years. It is instructive to examine this  la
-------
                                                                                         BIOMANIPULATION
                       Table 1.—Hydrologic and phosphorus inputs to Square Lake, 1982'.
Source
Hydrologic Input
    (103m3)
Phosphorus Input
      (kg)
Atmospheric
Surface inflow
Ground water
Septic field infiltration
    593 (50)2
     20(6)
   2,014(265)
     9.2 (2.0)
     3.7(1.2)
   125.2 (33.5)
    51.3(10.6)
 'From Osgood (1983a)
 "Parameter Value (+ / - associated error)
ygen/temperature profiles were taken on site. In 1982,
the lake was studied more intensively. Samples were
collected in the lake from three sites during 16 visits
(once in January  and 15  times  during the ice-free
period).  Additional  analyses  in  1982  included
phosphorus  profiles  and   plankton  community
analysis.
  Table 1 summarizes the hydrologic and phosphorus
loading information gathered in  1982 (Nelson  and
Brown, 1983). Precipitation, surface outflow, and lake
level (storage) were all measured directly for volume
and phosphorus concentration. Estimates of evapora-
tion (pan data x 0.7 for the lake's surface) and surface
inflow (Oberts, 1982)  completed  the  water   and
phosphorus budgets.  Net groundwater  inflow was
computed as a residual of the water balance equation
and its phosphorus concentration was obtained from
local well data.
  Square  Lake has apparently been clear since 1926
(Frg. 2). Mean summertime Secchi disk transparency
has normally been > 7 m. Other, more recent physical/
chemical historical data show no significant change
from 1980 through 1982 (Osgood,  1981, 1982a, 1983a).
Despite the  relatively intense fish  stocking record
(Table 2), Square Lake's lower trophic ecology has ap-
parently remained  unchanged.
  The nutrient  hydrodynamics of  Square Lake in-
dicate a meso- to eutrophic lake.  Ground water is the
largest  component   of   the annual  water   and
phosphorus inputs (77 and 67 percent, respectively).
Direct precipitation and surface runoff provide the re-
mainder of the annual inputs of water and phosphorus
to the lake (Table  1). There is a single, continuously
flowing surface outlet  from the lake. Considering the
annual (1982)  phosphorus hydrodynamics and  the
lake's basin morphology, the in-lake  phosphorus con-
centration is predicted to be 19 ^g/l; 90 percent con-
fidence interval: 6-22 ng/l (Dillion and Rigler, 1974; Kir-
  12-
          1930
                        1960
                              1860
                                     1970
                                           1(80
Figure  2.—Summertime Secchi  disk  transparencies  for
Square Lake (1926-1982). From Minn.  Dep.  Nat. Resour.,
Minn.  Pollut. Control Agency; Lie (unpubl.j; Osgood (1981,
1982a, 1983a).
       chner and  Dillon, 1975;  Reckhow et al.  1980). This
       agrees quite closely with the observed phosphorus
       concentration in the lake; annual whole-lake average
       (1982) is  18 /ug/l.  The areal phosphorus loading rate
       (0.223 gm/m2/year, 1982) and the lake's observed sur-
       face phosphorus concentrations (24,  15,  18 ^g/l;
       1980-82 seasonal averages) also indicate that Square
       Lake is meso- or eutrophic (Dillon, 1975; Carlson, 1977,
       1979).
         Hypolimnetic oxygen depletion rate (ODR) is related
       to primary  production at the lake's  surface (Walker,
              Table 2.—Minnesota Department of Natural
                   Resources stocking since 1908.
                           Square Lake
Year
1908-1944









1945
1947
1947
1948
1949
1950
1951
1952
1952
1952
1952
1952
1953
1954
1954
1954
1954
1955
1956
1959
1962
1968
1972
1974
1975
1977
1978
1980
1981 (Aug.)
1981 (Oct.)
1982 (May)
1982 (June)
1982 (Oct.)
Number
10 cans3
13 Cans
250,000
5,500
44,050
950
6,795
25,000
124,450
18,400
7,200
76,000
2,000
12,250
7,105
40
504
312
1,500
78
12
3
70
1,890
2
400
100
5,130
3,400
3,300
2,500
675
114
1,290
348
420
64
783
1,740
523
200
2,100
467
Size1
—
—
Fr
—
F
—
F
—
F
Fr
F
F
F
F
F
A
Y
Y
F
A
A
A
A
F
A
A
A
F
F
F
F
Y
Y
F
Y
A
A
A
Y
Y
A
Y
Y
Species2
B&C
W
W
C
C
B
B
L
S
S
W
M
S
W
W
N
W
W
C
C
S
N
SM
SM
N
C
S
SM
SM
SM
N
N
N
R
R
BR
BR
BR
R
R
R
R
R
       'Fr = fry; F = Fingerling; Y = Yearlmg; A = Adult.
       *B = Bass; BR = Brown Trout; C = Crappie; L = Lake Trout; M = Minnow; N = Nor-
       thern  Pike; R = Rainbow  Trout, S = Sunfish; SM = Smallmouth Bass;
       W = Walleye.
       'Milk cans were used, number was probably small.
                                                  145

-------
 LAKE AND RESERVOIR MANAGEMENT
 1979; Charlton, 1980). In turn, primary productivity is
 related to ambient nutrient conditions (Smith, 1979).
 Hypolimnetic  ODR  then,  is related  to  a  lake's
 phosphorus hydrodynamics (Welch and Perkins, 1979)
 and is  predicted to be 418 mg 02/m2/day in Square
 Lake.  This agrees with the  observed  hypolimnetic
 ODR (490 and 470 mg 02/m2/day; 1981 and 1982 respec-
 tively), indicating that Square Lake is as productive as
 the phosphorus conditions  would  dictate. Again,
 hypolimnetic ODR and the associated rate of primary
 productivity indicate a  meso-  or  eutrophic lake
 (Wetzel, 1975; Welch and Perkins, 1979).
   Algal  standing  crop  (estimated  as   surface
 chlorophyll a) and Secchi disk transparency both in-
 dicate pligotrophy (Fig. 3) (Carlson, 1977). Chlorophyll
 is significantly lower than expected based on surface
 total phosphorus concentrations (Fig. 4). In  fact, ex-
 cept for one sampling date (Oct. 13,1982), chlorophyll
 is independent  of the lake's phosphorus concentra-
 tion (<  1 percent of the variance in chlorophyll is ex-
 plained by phosphorus based on the remaining sampl-
 ing dates; Fig. 4). Something other than phosphorus is
 controlling  algal abundance (but not productivity) in
 Square  Lake. Nitrogen is not likely to be limiting since
 the mean (1982) TN:NP ratio was 37 (Smith, 1982). This
 apparent trophic dysfunction is related to the grazing
 activities of Daphnia.
   Square Lake  lacks a substantial standing crop of
 algae (Fig.  3). Within modest limits,  the epilimnetic
 phytoplankton community is not responsive to the am-
 bient nutrient concentration. Phytoplankton communi-
 ty  dynamics  are illustrated  in  Figures 5 and  ID.
 Flagellates (Chlamydomonas, Chlorochromonas,
 Cryptomonas) predominate in cell  numbers  most of
 the year. Chlorochromonas is the dominant flagellate,
 except later in the season when Cryptomonas blooms
 (Fig. 6). Stephanodiscus bloomed  in  mid-June, com-
 prising a substantial portion of algal biomass at that
 time (Fig. 3 and 6). Another diatom, Cyclotella, increas-
 ed slightly later in the season. A mid-summer bloom of
       Jan Feb Mar Apr  May' June7 July ' Auo ' Sept' Oct ' Nov ' Dec '
                        1982
Figure 3.—Surface chlorophyll and  Secchi disk—Squars
Lake. TSI from Carlson (1977).
 the unicellular blue-green, Chroococcus, apparently
 did not comprise a  significant fraction  of the com-
 munity biomass (Fig. 3). Intense grazing pressure may
 have initiated the Chroococcus bloom (Porter, 1975
 1976).
   The  grazing activities of  the  large  cladoceran
 Daphnia pulicaria are  responsible for reduced algal
 biomass. This has been suggested earlier by Lie (un-
 publ. data, see Fig. 7)  and Osgood (1982b). Daphnia
 are abundant in Square Lake (Fig. 8). They are able to
 avoid predation by taking refuge in  or just below the
 metalimnion  (Lie, unpubl.)  during  the day, then
 migrate to the epilimnion at night. In this way, large
 Daphnia are able to avoid predation  (Brooks and Dod-
  I
     0     5    10    16    20    26    30    36    40
                   Total Phosphorus (ug/l)

 Figure 4.—Phosphorus/chlorophyll relationship—Square
 Lake, 1982. Lines represent a generalized regional relation-
 ship (model + / - 90 percent confidence interval of predicted
 chlorophyll; model not defined at TP < 10 ^g/l) from a past
 study (Osgood,  1981).  Points are  individual summertime
 sampling dates and circled point is 13 October. + indicates
 1981, 1982 and  1980 summertime  average values (left to
 right).
                                                       10.000-1
                                                             an ' Feb  Mar Apr  May 'June July ' Aug ' Sept Oct ' Nov : Dec
       I   I Other  j  ; ;.i Diatoms  t=> Greens  I   I Blue-greens


Figure 5.— Phytoplankton community composition— Square
Lake.
                                                  146

-------
                                                                                       BIOMANIPULATION
son, 1965; Dodson, 1970,  1974; Dodson et al. 1976;
Lynch, 1979; Lynch and Shapiro, 1981; Shapiro et al.
1982) and subsist on the productive phytoplankton.
   Daphnia's (pulicaria/pulex) grazing abilities seem to
be an important factor in maintaining the relatively
algae-free condition  in Square Lake. The epilimnetic
clearance rate of Daphnia is estimated according to
Osgood (1982c): about 1 ml/animal/hour for animals
1.8-2.1 mm. This rate is appropriate for feeding on the
smaller flagellates that are common at the lake's sur-
face (Fig. 6).  Applying this  rate for Daphnia during
nighttime hours in the epilimnion indicates that  15
percent of the epilimnion is cleared per day (1982
   10-1
         flagellates
                            Chlorochromonas
                                                    average excluding dates in September and October,
                                                    range: 6-34 percent). Accounting for larger Daphnia
                                                    (Burns, 1969), their increased filtration rates at night
                                                    (Starkweather, 1975), their  increased filtration rates
                                                    (efficiencies) on  larger phytoplankton cells (Osgood,
                                                    1982c), or the grazing of other zooplankters would in-
                                                    crease this  overall epilimnetic clearance rate. Com-
                                                    paring this rate to the  epilimnetic clearance rate on
                                                    Oct. 13,1 percent per day, demonstrates that an order
                                                    of magnitude reduction in Daphnia's grazing pressure
                                                    occurred at the end of the season.
                                                      The trophic conditions on Oct.  13, 1982 further sup-
                                                    port the hypothesis that Daphnia  controls algal abun-
                                                    dance. On this date, Daphnia was at its lowest annual
                                                    density (Fig. 8) and chlorophyll increased to a  level
                                                    within the predictive model (Fig. 4), its greatest annual
                                                    density (Fig. 3). The reduction of Daphnia allowed the
                                                    phytoplankton to increase to  levels dictated  by
                                                    phosphorus concentration (Figs. 4 and 7).
                                                      What happened to  Daphnia  on Oct.  13,  1982?
                                                    Daphnia  lost its  refuge. Daphnia began  its decline
                                                    later in August (Fig. 8). This decline coincided with a
                                                    thermocline erosion  (Fig. 9), that finally excluded
                                                    Daphnia from its refuge (the hypolimnion was anoxic
                                                    by  this time). Increased predation success,  then,
                                                    presumably  reduced  Daphnia's numbers.  Male
                                                    Daphnia or females with ephippia were not noted at
                                                    this time, so the initiation of an overwintering popula-
                                                    tion was not  indicated. Whether this late  season
                                                    decline in Daphnia is a normal annual occurrence is
                                                    not  known since historical data from this late in the
                                                    year do not exist.
7  10-,
 E
o

o   0-
        Chroococcus
        Cyclotella
  .15-,
                              Stephanodlscus
        J'F'M'A'M'J'J'A'S'O'N'D'
                         1982
Figure 6.—Seasonal occurrence of important phytoplank-
ters—Square Lake. Flagellates include:  Chlamydomonas,
Chlorochromonas and Cryptomonas.
                                                       • .100-

                                                       5
                                                           0123
                                                                                       10  11 12  13 14
                                                                               Days
                                                     Figure 7.—Phytoplankton  community  bioassy—Square
                                                     Lake, Aug. 1,  1975. Laboratory incubation of the epilimnetic
                                                     phytoplankton community with zooplankton removed under
                                                     continuous light at 22° Celsius (after Lie, unpubl.).
                                                                                 Daphnia pullearla

                                                                                 Daphnia galeata mendotae
                                                            Jan ' Feb ' Mar ' Apr ' May 'June1 July ' Aug ' Sept' Oct ' Nov ' Doc
                                                                             1982
                                                    Figure 8.—Seasonal occurrence of D. pulicaria and  D.
                                                    galeata mendotae—Square Lake.
                                                 147

-------
 LAKE AND RESERVOIR MANAGEMENT
 DISCUSSION

 What environmental conditions are required to main-
 tain the niche of large-bodied Daphnia in Square
 Lake? Square Lake is deep and its  surface is pro-
 tected from the wind (Fig. 1), thus it stratifies early in
 the season and  remains so throughout the summer
 season (April  through  October) with  a summertime
 thermocline depth from 7-8 m. The large hypolimne"ic
 volume relative to the rate of hypolimnetic oxygsn
 depletion (Charlton, 1980) also seems  important. Beth
 a  stable  stratified period and  large  hypolimnetic
 volume (relative to the lake's ODR)  seem  to be
 beneficial  for maintaining  Daphnia's  metalimnetic
 refuge. The volumetric hypolimnetic  ODR seems to
 quantitatively  describe  a suitable  refuge.  Tie
 volumetric hypolimnetic ODR for Square Lake is 64-73
 mg 02/m3/day (1981 and 1982). Presumably then, a re.te
 within this limit would indicate a safe refuge.
   The lake's  phosphorus  hydrodynamics then, are
 also  important.  Square Lake receives most of ts
 hydraulic and  nutrient input from ground water enter-
 ing more or less uniformly throughout the year. Ttie
 lake has no discrete surface inflows and a continuous-
 ly flowing outlet. This consistent (seasonally) nutrient
 input  has led to a  mildly enriched lake whose
 phosphorus  concentration  is  within  a  threshold
 described  by  Shapiro  (1982). Within this threshold
 phosphorus concentration, algal  production is low
 enough to yield a rate of oxygen depletion inadequate
 to  deplete hypolimnetic  oxygen, thus  preserving
 Daphnia's metalimnetic   refuge. This threshold
 phosphorus  concentration  then,   has  a clear
 dependence  on hypolimnetic  volume  and stable
 stratification  (Smith, 1979; Welch and Perkins,  1979;
 Charlton, 1980). When this  threshold was exceeded in
 Lake Harriet, Minn., in  1974 and 1975, large Daphnia
 disappeared and algal abundance increased (Shapiro,
 1982).
   Blue-green algae as well  as  cladocerans become
 more predominant at higher phosphorus concentra-
 tions (McNaught, 1975). These algae may be a poor
 food source for planktonic herbivores because they
 may be inefficiently filtered or toxic (Arnold, 1971). In-
 deed, large herbivores may encourage the presence of
 some blue-greens or other undesirable algae (Porter,
 1976; Lynch, 1980). This suggests another mechanism
 for a  (different)  phosphorus threshold  where, at
 elevated phosphorus concentrations, Daphnia's abili-
 ty to control algal abundance or species composition
 is reduced. Unlike Shapiro's (1982) phosphorus thres-
 hold,  this threshold would  not necessarily require
 stable stratification and its value would be indepen-
 dent of hypolimnetic volume.
   Trophic stability in Square Lake seems to be related
 to  a  combination  of  physical and chemical  en-
 vironmental factors. The lake's basin morphology and
 its protection  from wind  mixing allow it to stably
 stratify for the  entire summer season. This physical
 condition, combined with the lake's  mild enrichment
 appears  to  provide  Daphnia  with  a  reliable
 metalimnetic refuge. Also, the modest nutrient condi-
 tions in the lake appear to be important for protecting
 the quality of Daphnia's metalimnetic refuge and for
 providing an edible (thus controllable) food source.


 CONCLUSIONS AND RESEARCH NEEDS

 There is no question that Daphnia currently controls
 algal abundance in Square  Lake. Square Lake has
 been  clear  for a  long time,  presumably by  this
 mechanism. The long-term stability  of this trophic
 association seems to depend on a particular combina-
tion of physical and chemical environmental factors
                                            Dissolved Oxygen (mg/l)
               26 Augutt
                                   10 September
                                                           23 Septembe
                                                                                  13 October
Figure 9.—Sequential dissolved oxygen/temperature profilas for the last four sampling dates in 1982—Square Lake. Hatched
areas indicate an assumed refuge with boundaries defined as follows: upper, the depth following a temperature decrease of >
1°C per meter; lower, dissolved oxygen = 1 mg/l.
                                                148

-------
                                                                                              BIOMANIPULATION
 that define Daphnia's niche. These factors have been
 generally addressed and include stable stratification,
 a relatively  large (volume)  hypolimnion compared to
 its ODR arid phosphorus concentrations within cer-
 tain threshold limits. Other biological factors are also
 important for preserving Daphnia's niche (Lynch, 1979;
 Lynch and Shapiro, 1981), but appear to be unimpor-
 tant in Square  Lake. These environmental conditions
 may  be generally  important  for  successfully
 reconstructing  biotic communities in lakes for the pur-
 poses of water quality improvements.
   Nutrients are still important. Biomanipulations aim-
 ed at increasing the grazing efficiency of  a lake's
 zooplankton  community cannot strictly be an alter-
 native to nutrient management; nutrients need to  be
 considered. The degree of success of such biomanipu-
 lations  seems  to  depend on nutrient and  other en-
 vironmental factors as well.  However, the particular
 importance of these factors needs further resolution.
   The following questions need to be addressed: How
 do the threshold phosphorus concentrations depend
 on lake  morphology? Are certain lake types physically
 unable to provide a niche for large-bodied Daphnia?
 When will large Daphnia be beneficial and when will
 the presence of Daphnia do harm (such as causing
 Aphanizomenon flake blooms)?
   Biomanipulation  is  a  valuable  lake  restoration
 technique, but it will not  replace nutrient manage-
 ment. Rather,  biomanipulation  should  complement
 other lake management practices and, pending defini-
 tion of  some general environmental  limitations, can
 serve as a primary  lake management technique.

 ACKNOWLEDGEMENTS: Funding for the studies on Square
 Lake were provided at various times by the U.S. Environmen-
 tal Protection Agency, the State  of  Minnesota, and  the
 Metropolitan  Council with  cooperative funding  from  the
 Metropolitan  Waste  Control  Commission and  the U.S.
 Geological Survey. I  thank G. Lie for  unpublished data. D.
 Wright  and  T.  Nponan  reviewed  early drafts  of this
 manuscript and their  comments led to great improvements.
 REFERENCES

 Arnold,  D.E.  1971.  Ingestion,  assimilation, survival and
   reproduction by Daphnia pulex fed seven species of blue-
   green algae. Limnol. Oceanogr. 16:906-20.
 Brooks, J.L, and S.I. Dodson. 1965. Predation, body size and
   composition of plankton. Science  150:28-35.
 Burns,  C.W.  1969.  Relation  between   filtering  rate,
   temperature, and  body size in four species of Daphnia.
   Limnol. Oceanogr. 14:693-700.
 Carlson, R.E. 1977. A trophic state index for lakes. Limnol.
   Oceanogr. 22:361-69.
	1979. A review of the philosophy and construction
   of trophic state indices. Pages 1-52 in. T.E. Maloney,  ed.
   Lake   and   Reservoir Classification  System.  EPA
   600/3-79-074. U.S. Environ. Prot. Agency, Washington, D.C.
Carlson, R.E., and S.A. Schoenberg.  1983. Controlling blue-
   green algae by zooplankton grazing. Pages 228-33 in Lake
   Restoration,  Protection and Management. Second Annu.
   Conf.  N.  Am. Lake Manage. Soc.  Oct. 26-29, 1982, Van-
   couver, B.C. EPA 440/5-83-001. U.S. Environ. Prot. Agency,
  Washington, D.C.
Chapra,   S.C.,  and   A.  Robertson.   1977.  Great  Lakes
  eutrophication: The effect of point  source control of total
   phosphorus. Science 196:1448-50.
Charlton, M.N.  1980.  Hypolimnion oxygen  consumption  in
  lakes: Discussion  of productivity  and morphometry  ef-
  fects. Can. J. Fish. Aquat. Sci. 37:1531-39.
Dillon, P.J. 1975. The phosphorus budget of Cameron Lake,
  Ontario: The importance of flushing rate to the degree of
  eutrophy  in lakes. Limnol. Oceanogr. 20:28-39.
 Dillion, P.J., and F.H. Rigler. 1974. A test of a simple nutrient
   budget model predicting the phosphorus concentration in
   lake water. J. Fish. Res. Board Can. 31:1771-8.
 Dodson, S.I. 1970. Complementary feeding niches sustained
   by size-selective predation. Limnol. Oceanogr. 15:131-7.
 	. 1974. Zooplankton competition and predation: An
   experimental  test  of  the  size-efficiency  hypothesis.
   Ecology 55:605-13.
 Dodson, S.I.,  C. Edwards,  F. Wiman,  and J.C.  Normandin.
   1976. Zooplankton: specific distribution and food abun-
   dance. Limnol. Oceangr. 24:253-72.
 Edmondson,  W.T., and J.T.  Lehman. 1981. The effect  of
   changes in the  nutrient income on the condition of Lake
   Washington. Limnol. Oceanogr. 26:1-29.
 Kirchner, W.B., and  P.J.  Dillon. 1975. An empirical method
   of estimating the retention of phosphorus in lakes. Water
   Resour. Res. 11:182-3.
 Larsen, D.P., D.W. Schultz, and K.W. Malueg. 1981. Summer
   internal phosphorus supplies in Shagawa Lake, Minn. Lim-
   nol. Oceanogr. 26:740-53.
 Lie, G.B. 1978. Unpubl. data. Univ. Minnesota.

 Lynch, M.  1979.  Predation, competition and  zooplankton
   community structure:  An  experimental  study. Limnol.
   Oceangr. 24:253-72.
 Lynch, M. 1980. Aphanizomenon blooms: Alternate control
   and  cultivation by Daphnia pulex. Am. Soc. Limnol.
   Oceanogr.  Spec.  Symp. 3:299-304. Univ.  Press,  New
   England.
 Lynch, M., and J. Shapiro. 1981. Predation, enrichment, and
   phytoplankton community structure. Limnol. Oceanogr.
   26:86-102.
 McNaught, D.C. 1975. A  hypothesis to explain  the succes-
   sion from calanoids cladocerans during  eutrophication.
   Verh. Int. Verein. Limnol. 19:724-31.
 Nelson, L, and R.G. Brown. 1983.  Streamflow and  water
   quality data for lake and wetland inflows and outflows in
   the Twin Cities Metropolitan  Area, Minn., 1981-82, Draft.
   Open File Rep. 83-543.  U.S. Geolog. Surv.
 Oberts, G.L 1982. Nonpoint source pollution in the Metro-
   politan Area: Tech. Comple. Rep.  Metropolitan Counc.
   Publ. No. 10-82-016.

 Osgood, R. 1981. A study of the water quality of 60 lakes in
   the seven county Metropolitan Area. Metropolitan Counc.
   Publ. No. 01-81-047.
 	1982a. A 1981 study of the water quality of 30 lakes
   in the  seven-county  Metropolitan Area. Metropolitan
   Counc. No. 10-82-005.
 	1982b. Using differences among Carlson's trophic
   state index values in regional water quality assessment.
   Water Res.  Bull. 18:67-74.
       _. 1982c. Differential filtration efficiencies of Daphnia
  pulex. Can. J. Zool. 60:2129-33.
	1983a. Diagnostic-feasibility study of seven Metro-
  politan Area lakes. Part Two: Square Lake. Metropolitan
  Counc. Publ. No. 10-83-093G.
	1983b. Diagnostic-feasibility study of seven Metro-
  politan  Area  lakes.  Part  One:  General   Overview.
  Metropolitan Counc. Publ. No. 10-83-092.
Porter, K.G. 1975. Viable gut passage of gelatinous green
  algae by Daphnia. Verh. Int. Verein. Limnol. 19:2840-50.
       _. 1976. Enhancement of algal growth and productivi-
  ty by grazing zooplankton. Nature 244:174-180.

Reckhow, K.H., M.N. Beaulac, and J.T. Simpson. 1980. Model-
  ling  Phosphorus  and  Loading  Response  and  Lake
  Response Under Uncertainty: A Manual and Compilation
  of Export Coefficients.  EPA 440/5-80-011.  U.S. Environ.
  Prot. Agency, Washington, DC.
Shapiro, J. 1979. The need for more biology in lake restora-
  tion. Pages 161-7 In Lake Restoration, Proc.  Natl. Conf.
  August 22-24, 1978. EPA 440/5-79-001. U.S.  Environ. Prot.
  Agency, Washington, DC.
                                                     149

-------
LAKE AND RESERVOIR MANAGEMENT

	1982. The role of physical-chemical conditions in
  affecting algal abundance—Lake Harriet. In Shapiro et al.
  Experiments and experiences in biomanipulation: Studios
  of  biological ways to  reduce algal abundance  and
  eliminate  blue-greens.  Interim Rep.  No. 19. Limnol.  Res.
  Center. EPA 600/3-82-096.  U.S.  Environ.  Prot. Agency,
  Washington, DC.
Shapiro,  J.,  et al. 1982.  Experiments and  experiences in
  biomanipulation: Studies of  biological ways to reduce
  algal abundance and eliminate blue-greens. Interim  Rep.
  No.  19. Limnol.  Res. Center. EPA 600/3-82-096. U.S. En-
  viron. Prot. Agency, Washington, DC.
Smith, V.H.  1979. Nutrient  dependence of primary produc-
  tivity in lakes. Limnol. Oceanogr. 24:1051-64.
	1982. The nitrogen and phosphorus dependence of
  algal biomass in lakes:  An  empirical and  theoretical
  analysis. Limnol. Oceanogr. 27:1101-12.
Starkweather, P.L. 1975. Diel patterns of grazing in Daphnia
  pulex Leydig. Verh. Int. Verein. Limnol. 19:2851-7.
Walker, W.W., Jr. 1979. Use of hypolimnetic oxygen depletion
  rate as a trophic state index for lakes. Water Resour. Res.
  15:1463-70.

Welch,  E.B., and  M.A.  Perkins.  1979.  Oxygen  deficit-
  phosphorus loading in lakes. J. Water Pollut. Control Fed.
  51:2823-8.
Wetzel, R.G. 1975.  Limnology.  W.B. Saunders Co.  Phila-
  delphia.
                                                      150

-------
BIOLOGICAL CONTROL OF NUISANCE ALGAE  BY
DAPHNIA  PULEX:  EXPERIMENTAL STUDIES
MICHAEL  J.  VANNI
Department of Ecology, Ethology and Evolution
University  of  Illinois
Champaign,  Illinois
            ABSTRACT

            The feasibility of using the zooplankton species Daphnia pulex as a biological control agent of nuisance
            algal blooms was investigated experimentally in two Illinois lakes. Two questions were posed: (1) Can
            grazing by Daphnia pulex buffer the effects of increased nutrient loading to an initially low-nutrient
            lake? and (2) Can D. pulex significantly reduce algal biomass when introduced into a lake that already
            exhibits excessive phytoplankton growth? These questions were answered by introducing D, pulex
            individuals into large enclosures (1,000 liters) suspended in the lakes, which do not naturally contain
            D. pulex . The first question was addressed in Dynamite Lake, an oligo-mesotrophic quarry lake that
            does not normally exhibit algal blooms. Four treatments were employed within the enclosures:  (1)
            a control, (2) D. pulex added, (3) nutrients (nitrogen and phosphorus) added, and (4) D. pulex and
            nutrients added. Nutrients were not added  until the D. pulex populations became  established and
            were added at weekly intervals after initial addition. In an experiment performed in 1982 D. pulex
            displayed the ability to substantially buffer the effects of increased nutrient loading. Although
            addition of nutrients increased phytoplankton biomass in all enclosures to which they were add-
            ed,  by the end of the experiment enclosures without D.  pulex had  phytoplankton densities
            (measured as chlorophyll a concentration) greater than those with D. pulex. The second quest-
            ion  was addressed in Larimore Pond, a highly eutrophic farm pond with a dense summer surface
            bloom of phytoplankton. One experiment in the summer of 1982 was performed with a control and
            with D. pulex added. The results were quite  striking: by the end of the experiment (roughly 6 weeks)
            enclosures without D. pulex exhibited phytoplankton densities an order of magnitude greater than
            those with D. pulex . These results demonstrate that D. pulex can effectively control phytoplankton
            biomass even in lakes in which D. pulex is not a natural inhabitant. Therefore, management strategies
            designed to facilitate introduction and survival of D. pulex or other large grazers should be encour-
            aged as a viable within-lake technique for mitigating the symptoms of eutrophication.
 INTRODUCTION

 Cultural eutrophication, the artificial nutrient enrich-
 ment of lakes  and ponds from human activity, con-
 tinues to be a major water resources problem. One of
 the most conspicuous and undesirable symptoms of
 nutrient enrichment is an increased  phytoplankton
 abundance ("blooms") and consequent reduced  water
 transparency. In  addition to reducing water clarity,
 phytoplankton blooms can result in offensive odors, a
 depletion of hypolimnetic oxygen as the phytoplankton
 decompose and a general decline in water quality. Ex-
 treme depletion of dissolved oxygen can result in fish-
 kills.  Because phytoplankton blooms  exhibit  such
 undesirable qualities,  considerable effort  has  been
 put into developing methods to control  phytoplankton
 growth and restore eutrophic lakes to less productive
 states (e.g. U.S. Environ. Prot. Agency,  1979).
   In many lakes  the cause of algal blooms is an ex-
 cessively high  input of nutrients  that ordinarily limit
 phytoplankton growth, especially phosphorus and to a
 lesser  extent  nitrogen (Schindler  and Fee,  1974;
 Schindler, 1977). Hence, reduction of nutrient loading
 rates to lakes often functions as an effective restora-
 tive technique, the recovery of Lake Washington  being
 the classic example (Edmondson  and Lehman, 1981).
 However, in areas where nutrients  are deposited in
 lakes  from nonpoint sources,  such as agricultural
 watersheds, diversion  or reduction of the amount of
 nutrients entering a lake is very difficult. Consequent-
ly, the  symptoms  of  eutrophication may  best  be
treated with in-lake measures.
  Recently  the  possibility  of  manipulating higher
trophic  levels  (zooplankton and  fish)  to  control
nuisance phytoplankton  biologically has  been con-
sidered  (Shapiro et. al. 1975, 1982). There is ample
evidence that  certain relatively large herbivorous
zooplankton species, especially  the cladoceran
Daphnia pulex, can hold phytoplankton abundance at
low levels. The presence of D. pulex  or closely related
species at  natural densities is  generally associated
with low phytoplankton biomass and high water trans-
parency (Losos and Hetesa,  1973; Hurlbert  and Mulla,
1981; Lynch and Shapiro,  1981), and  it is clear that the
feeding  activities of this herbivore  can cause these
conditions (Lynch and Shapiro, 1981).
  Nevertheless, whether it is feasible to use large zoo-
plankton species  as a  means  of  reducing phyto-
plankton in  lake management practices remains  un-
certain.  No attempt has been made  to deliberately in-
troduce a large zooplankton species into a eutrophic
lake or enclosures within such a lake to determine if
these grazers are capable of alleviating the symptoms
of nutrient  enrichment. Often eutrophic lakes do not
naturally contain large grazers, and it is not clear what
conditions  in lakes are prohibitive to the  survival of
these species. Although in many lakes the absence of
large species is probably due to size-selective fish
                                                  151

-------
LAKE AND RESERVOIR MANAGEMENT
predation (Brooks  and Dodson, 1965; Lynch,  1979),
other factors may  exclude them from certain  lakes.
For  example,  temperatures  in  shallow  Midwestern
lakes often exceed 25° Celsius. Temperatures this
high have been shown to inhibit the feeding efficiency
of Daphnia pulex  (Burns,  1969; Lynch, 1977). Addi-
tionally,  certain  chemical  "water quality"
characteristics can impair the survival of D. pulex and
prevent this species from inhabiting lakes displaying
these conditions (Sprules, 1972; Neill, 1978). Before 0.
pulex or other  large grazers can be successfully used
in biological control programs, the range of conditions
under which they can survive must be determined.
  The purpose of this study was to determine whether
D. pulex can  survive and  control  phytoplankton
biomass in small, shallow Midwestern  lakes. Two
related questions were posed: (1) Can D. pulex reduce
phytoplankton  in a lake displaying  excessive algal
blooms? and (2) Can  D. pulex buffer the effects of
nutrient  loading to  a relatively  low-nutrient, low-
phytoplankton  lake? These questions were addressed
by experimentally  adding D. pulex to large in situ
enclosures that replicated the plankton communitie'S
of two small lakes. In this paper I deal only with the ef-
fects on total phytoplankton  abundance.  The effeels
of the D. pulex introductions on phytoplankton com-
munity structure and on other zooplankton  species
will be presented elsewhere.

STUDY SITE AND METHODS

The lakes chosen for this study are small (<0.6 ha),
located in Vermilion Co., III., and lie within watersheds
dominated  by  intensive  agricultural  practice!;.
Larimore Pond is a highly eutrophic  farm pond ex-
hibiting  high  nutrient levels  (summer  1982  total
phosphorus 58.4-206.2 /^g/l at surface) and a dense
surface bloom of phytoplankton (summer 1982 sur-
face chlorophyll a 23.5-169.0 ^g/l). Below the surface
bloom  of  phytoplankton,  water  rapidly becomes
depleted of dissolved oxygen. During summer of 1982
the concentration of dissolved oxygen directly above
sediments (1.5  m) was often < 1.0 mg/l. Winterkills of
fish are common (R.W. Larimore, pers. comm.).
  Dynamite Lake lies in an abandoned  limestone
quarry and  is somewhat buffered from the nutrients
deriving  from  farmland.  Nutrient levels  and phyto-
plankton abundance are low compared to other lakes
in the area (summer 1982 total phosphorus 11.2-24.4
/4I/I;  chlorophyll a 0.98-6.28 ^g/l). Dynamite  Lake is
shallow  (maximum  depth 2.0  m)  and  is  virtually
isothermal throughout summer.
  Both  lakes contain zooplanktivorous fish; in Lari-
more Pond  these  include fathead  minnow and a
sparse population  of  paddlefish while in Dynamite
Lake bluegill sunfish are the dominant planktivoreis.
The zooplankton communities of both lakes include
only  small species (maximum length <1.0 mm). Prin-
cipal  grazers  include the  cladocerans  Bosmina,
Ceriodaphnia and Diaphanosoma, the copepod D/ap-
tomus, and the rotifer Keratella in Dynamite Lake, and
Diaptomus, cyclopoid copepods, and several rotifer
species in Larimore Pond. Daphnia pulex does not in-
habit either lake.
  In  summer 1982 an enclosure experiment was con-
ducted in each lake. Enclosures were made  of clear
polyethylene  tubing  and  were  suspended  from
wooden and styrofoam frames floating at the  surface.
Enclosures extended to near the bottom of the lakes
and were sealed  at the bottom to isolate a water col-
umn from the lake, but open at the top to allow con-
tact  with the atmosphere. Enclosures were filled on
 June 29, 1982 with water from a depth of 1 m using a
 gasoline-powered  pump. Volume of the  enclosures
 was MOOO I. In Larimore Pond, two treatments, with
 two enclosures per treatment, were used: (1) a control
 in which the natural plankton community was added
 to the enclosures, and (2) the natural plankton com-
 munity with Daphnia pulex added to the enclosures.
 The Dynamite Lake  experiment was begun with the
 same two treatments (four enclosures per treatment),
 but upon establishment  of the D. pulex populations
 (see discussion of results), each treatment was split in
 two,   one  receiving   nutrients  (nitrogen   and
 phosphorus) and the other maintained as it had  been.
 The result after this split was four treatments (two
 enclosures per treatment) in Dynamite Lake: (1) con-
 trol, (2) D. pulex added, (3) nutrients added, and (4)  D.
 pulex and  nutrients added. Nutrients  were added  to
 treatments 3 and 4 weekly beginning July  16.  Each
 time, 300 \IQ N/l (as NH4NO3) and 10 ^g P/l (as KH2PO4)
 were  added,  roughly the same  N:P ratio as that  in
 Dynamite Lake. Previous experiments showed that ad-
 dition of nutrients  into Dynamite Lake enclosures  at
 these concentrations with only the natural zooplank-
 ton community present greatly increased phytoplank-
 ton abundance (M. Vanni, in prep.). The present experi-
 ment  allows comparison of the effectiveness of the
 natural Dynamite Lake zooplankton community and
 the community plus D, pulex at buffering the effects  of
 nutrient enrichment.
  D. pulex stocks were obtained from several sources,
 including temporary ponds and permanent lakes, and
 were  represented  by seven  genetically  (electro-
 phoretically) distinct clonal groups (Lynch, 1983). Each
 clonal group was cultured separately in gallon jars  in
 the laboratory using a medium modified from Murphy
 (1970), with the  algae Chlamydomonas reinhardii and
 Scenedesmus dimorphus as food.  Prior to introduc-
 tion to  enclosures, D. pulex from  the seven clonal
 groups were combined and approximately the same
 number of individuals of each clonal group added to
 each  enclosure  designated  to  receive Daphnia.  A
 mean of 303 (SE = 7.3) D. pulex were added to  each
 enclosure on July 2. This  is a low density (~0.3/l) and
 allows D. pulex to increase naturally.
  To determine if D. pulex became established in the
 enclosures,  two vertical hauls  with a  Wisconsin
 plankton net were taken from each enclosure on  each
 sampling date. Samples were preserved in Formalin
 and counted under a compound microscope to obtain
 D. pulex population  densities. Phytoplankton abun-
 dance  was measured as chlorophyll  a,  determined
 spectrophotometrically after extraction with acetone
(Strickland and Parsons, 1968). Total phosphorus was
 measured  spectrophotometrically using the ascorbic
acid method after digestion with potassium persulfate
(Am. Pub.  Health Ass., 1971). Midday temperature in
each enclosure was determined with a YSI  model 54A
meter.
RESULTS

Temperatures of both lakes and enclosures remained
above 20° Celsius throughout the experiment (Fig. 1);
in-Dynamite Lake, temperatures exceeded 25° Celsius
on all but one date. Within a lake temperatures did not
differ among  treatments and enclosure  and lake
temperatures were similar (Fig. 1).
  Total  phosphorus concentrations (TP) are given in
Fig. 2.1 tested for differences in TP among treatments
in Dynamite Lake with a two-way ANOVA, pooling for
each treatment all dates after the initial nutrient addi-
                                                152

-------
                                                                                     BIOMANIPULATION
tions (that is, after July 16). In Dynamite Lake, as ex-
pected, TP was significantly elevated in enclosures
receiving nutrients (P = 0.006), while the introduction
of D. pulex had no effect on TP (P = 0.504). No inter-
action of nutrient and D. pulex addition was observed
(P = 0.454). TP was lower in the enclosures than in the
lake on all but one date (Fig. 2). The relatively low TP in
the enclosures was probably the result of periphyton
growth on the enclosure walls (periphyton can remove
phosphorus from the water column), and perhaps to
an  elimination of nutrient regeneration  from  the
sediments.
  In Larimore Pond no significant difference in TP bet-
ween treatments was found (P = 0.232, t-test pooling
all dates after July 16). TP was much higher in the lake
(x = 178 j^g/l, SE = 15.3) than in  the  enclosures.  The
reason for the discrepancy involves the manner in
which the enclosures were filled. To do so, water was
pumped  from a depth of  1 m. Most of the TP in
Larimore Pond is in the form of phytoplankton, which
is most abundant in the top 0.5 m. Phytoplankton (and
TP) declines sharply below this level; thus enclosures
were filled with water containing less TP than the lake
surface,  from which lake samples were taken.
   Daphnla pulex populations became established in
all enclosures into which they were introduced (Fig. 3).
In Dynamite Lake D. pulex densities were greater in
enriched enclosures, probably because of  increased
food (phytoplankton) availability in these enclosures
(see chlorophyll a results). D. pulex densities were
much greater in Larimore Pond enclosures, even when
compared   to  nutrient-enriched  Dynamite  Lake
         -_o— CONTROL

         —.— DAPHNIA
       LARIMORE
       	o	NUTRIENTS

       	•	NIITRIFNTS + DAPHNIA

         » LAKE
       DYNAMITE
       DYNAMITE
      30

      JUNE
         20

         JUL
             9

           AUQ
Figure 1.—Midday
enclosures.
temperature at 1 m  in the  lakes and
      25-i
  O)
 3   i

 CO
 z>
 DC
 O

 0.     (T_|
 CO     b i
 o

 °-    20-|
                  DYNAMITE
        ..o--CONTROL  -o-NUTRIENTS
        — •--DAPHNIA   -^-NUTRIENTS & DAPHNIA
           «•••• LAKE              N
                                     ..-*-*	<•
                                                                 l_/
                  LARIMORE
                       	•— DAPHNIA
                       — o—CONTROL
       10-
                                                         60-7
     45-
                                                         30-
                                                          15-
30
JUNE
5
10
15 20
JUL
25
3'o
                                                              30
                                                              JUNE
                  10
         20

       JUL
30
                                       AUG
 Figure 2.—Total phosphorus concentrations (mean and range) in the enclosures and Dynamite Lake. Arrow with D denotes
 data of Daphniapulex introductions and arrow with N denotes date nutrient enrichment began; nutrients were added weekly
 after this date. Total phosphorus concentrations in Larimore Pond were much higher than those in the enclosures (x = 178.0,
 SE = 15.3 ngl\) and are not presented. Refer to text for further explanation.
                                                 153

-------
 LAKE AND RESERVOIR MANAGEMENT
 enclosures.  In  Larimore  Pond, D. pulex population
 density fluctuated considerably (Fig. 3).
   Daphnia pulex had substantial  impacts on  the
 amount of phytoplankton in the enclosures (Fig. 4). In
 Dynamite Lake  chlorophyll a concentrations were
 similar between treatments until nutrient additions
 were begun. Upon initiation of enrichment, chlorophyll
 a increased, both in the presence and absence of D.
 pulex, but the increase was much more pronounced in
 the absence of Daphnia (Fig. 4). To test for differences
 between  treatments  in chlorophyll a  concentration,
 two-way ANOVA with repeated measures was used,
 using each date as a repeated measurement (Winef,
 1971).  Nutrient addition  significantly  increased
 chlorophyll a (P = 0.002), while overall the presence of
 D. pulex  had  a  (barely)  insignificant effect  cm
 chlorophyll a (P = 0.096). Significant interaction  ex-
 isted between D. pulex and  nutrient addition in deter-
 mining chlorophyll a levels (P = 0.007), indicating that
 nutrients have  less of an  effect on  phytoplankton
 abundance when D. pulex is present, and that D. pulex
 may  lower phytoplankton  abundance,  but only  in
 enriched  enclosures.  Indeed,   if only enriched
 enclosures are  considered, D.  pulex significantly
 lowered chlorophyll  a concentration (P = 0.037, one-
 way ANOVA with repeated measures). In  addition, no
 D. pu/ex-sample  date  interaction  was  found
 (P = 0.256), indicating that D. pulex depressed phyto-
 plankton  abundance  a constant  amount  throughout
 the experiment, despite continued nutrient inputs.
   Daphnia pulex had a  dramatic effect on  phyto-
 plankton  abundance  in  Larimore Pond (Fig. 4). Dif-
 ferences in chlorophyll a between treatments became
 apparent  less -than 2 weeks after D.  pulex was in-
 troduced  and persisted until the experiment was ter-
 minated (Fig. 4). During this time chlorophyll a concen-
 trations were much lower in enclosures containing O.
 pulex  (P = 0.037,  one-way  ANOVA  with repeated
              10
                                   AUG
 measures using all dates  after July 16). As in the
 enriched Dynamite Lake enclosures,  no D. pulex-
 sampling date  interaction was found (P = 0.522), in-
 dicating that the effect of D. pulex on phytoplankton
 abundance was constant throughout the  experiment,
 despite considerable temporal  and within-treatment
 variation in D. pulex density (Fig. 3).

 DISCUSSION

 The results of the enclosure experiments  support the
 hypothesis  that Daphnia pulex may be able to effec-
 tively  reduce phytoplankton  biomass in eutrophic
 lakes  and  buffer the  effects of  increased  nutrient
 loading to initially low-nutrient lakes. D. pulex popula-
 tions were more effective at reducing phytoplankton
 abundance  in Larimore Pond, where phytoplankton is
 very  dense, than  they  were  in  preventing  phyto-
 plankton from increasing  when nutrients were added
 to  Dynamite Lake  communities. D. pulex increased
 more rapidly and attained higher densities in Larimore
 Pond than Dynamite Lake.  Upon reaching high  den-
 sities, D. pulex substantially reduced the abundance
 of phytoplankton. Thus substantial feedback between
 these two trophic levels is apparent.
   In general, greater phytoplankton abundance  will
 lead to a higher density of D. pulex, which will then
 result  in lowered phytoplankton abundance. If phyto-
 plankton is reduced to the point at which D. pulex can-
 not reproduce,  the D. pulex populations  may crash,
 thereby resulting  in  elevated  phytoplankton abun-
                    DYNAMITE
                                                                          PI
                              - » LAKE
                              —o-CONTROL
                              --•»--OAPHNIA

                              —o-NUTRIENTS

                              —"-NUTRIENTS & DAPHNIA
                                                                         LARIMORE
                                                     o
                                                     DC  20-
                                                     O
                           --o-CONTROL

                           —*—DAPHNIA
                                                            3'0
                                                           JUNE
            5
                         20
                         JUL
                                                                                             AUG
Figure 3.—Daphnia pulex densities (mean and range) in ths
enclosures receiving introductions. Arrows as in Figure 2.
Figure 4.—Chlorophyll a concentrations  in the enclosures
and Dynamite Lake. Arrows  as in Figure 2. For Larimore
Pond enclosures, mean and range are given. One Dynamite
Lake enclosure intended to have nutrients but not D. pulex
was colonized by D. pulex. This enclosure was excluded from
analysis, leaving only  one enclosure for  this  treatment.
Because there was no replication for this treatment, error
bars are not included.
                                                 154

-------
                                                                                           BIOMANIPULATION
 dance. Under these circumstances the  lake  would
 then have reverted to its original, undesired condition.
 The entire process may be repeated, and  if this cycle
 repeats frequently, D. pulex would be of limited use in
 management programs designed to keep  phytoplank-
 ton at consistently low  levels.  However,  D. pulex
 populations in  Larimore Pond enclosures never  ap-
 proached extinction, while phytoplankton abundance
 remained low and  relatively constant  temporally in
 enclosures   with  D. pulex (Fig. 4).  Apparently  an
 equilibrium was reached between D. pulex grazing and
 phytoplankton abundance.
   The success of D. pulex in ameliorating nutrient ad-
 ditions to Dynamite Lake enclosures  demonstrates
 that the warm summer temperatures  comrrionly  en-
 countered in shallow Midwestern lakes will not inhibit
 the effectiveness of this species as a biological con-
 trol of nuisance phytoplankton. However, a potentially
 more severe barrier to using D. pulex or other large
 grazers to control phytoplankton is that it may be in
 direct  conflict  with the  maintenance of  a  viable
 fishery. Many sportfish in lakes are zooplanktivorous
 during  at least part of their life cycle, and in many
 cases planktivorous fish  cause  local  extinction of
 large grazers (Brooks  and Dodson,  1965;  Hrbacek,
 1962; Lynch, 1979).
   In  shallow  lakes such as Dynamite  Lake  and
 Larimore Pond, where  there is no deep-water  refuge
 available where  the large zooplankton species  can
 avoid visual fish  predation (Zaret and Suffern, 1976;
 Wright  et al. 1980), the problem of fish  predation will
 be especially severe. Nevertheless, D. pulex may be
 used to control phytoplankton in shallow lakes if fish
 communities can be  manipulated so  that they are
 dominated by piscivorous fish, provided the piscivores
 can hold planktivorous fish at low levels, and that the
 piscivores themselves, many of  which  are   plank-
 tivorous as fry, do not  have a detrimental impact on
 the large zooplankton  species.  In this manner, the
 goals  of  using D. pulex  as  a biological control of
 phytoplankton and  providing  a  viable  sport  fishery
 may be attained simultaneously.
   Although   the   introductions   of   D.  pulex  to
 enclosures in Dynamite Lake and Larimore Pond were
 successful, this does not necessarily imply that the
 same effect would  be  observed if D. pulex were in-
 troduced  into the lakes themselves, even if the pro-
 blem of planktivorous fish predation  is  overcome. To
 produce populations large enough to control  phyto-
 plankton  within the span  of one summer in  even a
 small lake would require that an enormous number of
 animals be introduced into the lake. Culture or collec-
 tion of  this number of D.  pulex would  be a difficult
 task. However, D. pulex may naturally colonize lakes
 provided fish predation is relaxed and chemical and
 physical  conditions  are  suitable.  For  example,
 Daphnia pulex often  appears in lakes in  which  it
 previously was not found following winterkills of fish
 (Shapiro et al. 1982). In such lakes, if D. pulex becomes
 abundant, phytoplankton abundance is reduced and
 transparency increased. However, large grazers do not
 always  appear in lakes that have undergone fishkills;
 they appeared in only four out of eight hard winterkills
 studied  by  Shapiro  et   al.  (1982).  Enclosure  ex-
 periments of the kind  used  in  Dynamite Lake and
 Larimore  Pond may be valuable as a "bioassay" to
 determine how D. pulex will fare in a particular lake. If
 D. pulex survives and controls phytoplankton biomass
 in the enclosures, It would indicate  that  a manage-
 ment plan geared toward  ensuring the  survival of D.
pulex in the lake  (such as through alteration  of fish
 communities) may be feasible.
ACKNOWLEDGEMENTS: I thank M. Lynch for support and
advice during this study; S. Bennett, L. Crossett, K. Fausch,
and C.A. Toline for assistance in the  field; C.A.  Smyth for
statistical advice; E. Schmidt for assisting in the processing
of zooplankton samples;  and C. Turkot for processing
zooplankton samles and preparing the figures. Comments by
M.  Berenbaum  and  J.R.  Karr  greatly  improved   the
manuscript. I am especially grateful to P. Bukaveckas and L.
Coutant of the Illinois Natural History Survey for sharing
their knowledge of chlorophyll extraction procedures and to
R.W. Larimore for allowing access to his land and pond. This
research was generously supported by the  Illinois Water
Resources Center, through grant S-093-ILL to  M. Lynch.
 REFERENCES

 Brooks, J.L, and S.I. Dodson. 1965. Predation, body size and
  composition of the plankton. Science 150: 28-35.
 Burns, C.W. 1969. Relation between filtering rate, tempera-
  ture, and body size in four species of Daphnia. Limnol.
  Oceanogr. 14: 693-700.
 Edmondson, W.T.,  and  J.T.  Lehman. 1981. The  effect of
  changes  in  the nutrient income on the condition of Lake
  Washington. Limnol. Oceanogr. 26: 1-29.
 Hrbacek, J. 1962. Species composition and the amount of
  zooplankton in relation to the fish stock. Rozpr. Cesk.
  Akad. Ved Rada Matemat. Prir. Ved 72(10): 1-116.
 Hurlbert, S.H., and M.S. Mulla. 1981. Impacts of mosquitofish
  (Gambusia  affinis) predation on plankton  communities.
  Hydrobiologia 83: 125-51.
 Larimore, R.W. 1982. Pers. comm. III. Nat. Hist. Surv.
 Losos, B., and J.  Hetesa. 1973. The effect of mineral fertili-
  zation and of carp fry on the composition and dynamics of
  plankton. Hydrobiol. Stud. 3: 173-217.
 Lynch, M.  1977.  Fitness and  optimal body size in zoo-
  plankton  populations.  Ecology 58: 763-74.
	1979.  Predation, competition, and zooplankton
  community  structure:  An  experimental study. Limnol.
  Oceanogr. 24: 253-72.
       _.  1983.  Ecological   genetics of Daphnia  pulex.
  Evolution 37: 358-74.

Lynch, M., and J. Shapiro. 1981. Predation, enrichment, and
  phytoplankton community structure. Limnol. Oceanogr.
  26: 86-102.

Murphy, J.S. 1970.  A general method  for the  monoxenic
  cultivation of the Daphnidae. Biol. Bull. 139: 321-32.

Neill, W.E. 1978. Experimental  studies  on  factors  limiting
  colonization by Daphnia pulex Leydig of coastal montane
  lakes in British Columbia. Can. J. Zool. 56: 2498-2507.

Schindler, D.W. 1977. Evolution of phosphorus limitation in
  lakes. Science 195: 260-2.
Schindler, D.W., and E.J.  Fee. 1974.  Experimental Lakes
  Area: Whole-lake  experiments in eutrophication.  J. Fish.
  Res. Board Can. 31: 937-53.

Shapiro, J., V. Lamarra, and M. Lynch.  1975. Biomanipula-
  tion: an ecosystem approach to lake restoration. In P.L.
  Brezonik, and J.L. Fox, eds. Water Quality Management
  Through Biological Control. Univ. Florida, Gainesville.
Shapiro, J., et al.  1982. Experiments  and experiences in
  biomanipulation. Interim Rep. No. 19. Limnol. Res. Center,
  Univ. Minnesota.

Sprules, G. 1972. Effects of size-selective predation and food
  competition on high  altitude zooplankton  communities.
  Ecology 53: 375-86.

Standard  Methods for the Examination of Water and Waste-
  water. 1971. 13th  ed. Am. Pub.  Health Ass. Washington,
  D.C.

Strickland, J.D.H., and T.R. Parsons. 1968. A practical hand-
  book of seawater analysis. Bull. Fish. Res. Board Can. 167:
  1-311.
                                                   155

-------
LAKE AND RESERVOIR MANAGEMENT


U.S. Environmental Protection Agency. 1979. Lake Restora-     Wright, D., W.J. O'Brien, and G.L Vinyard. 1980. Adaptive
  tion. EPA 440/5-79-001. Washington, D.C.                      value of vertical migration: a simulation model argument
Vanni, M.J. In prep. The influence of nutrient enrichment and       f?r the P^dation hypothesis. In W.C. Kerfoot, ed. Evolu-
  fish predation on phytoplankton community structure in       l'on and  Ecology of Zooplankton Communities.  United
  an oligo-mesotrophic lake.                                   Press of New England, Hanover, N.H.

Winer, B.J. 1971. Statistical  Principles  in Experimental     Zaret'  T'M.;i and J-S- Suffern.  1976. Vertical migration in
  Design  McGraw-Hill  New York                               zooplankton as a predator avoidance mechanism. Limnol.
                                                             Oceanogr. 21:804-13.
                                                     156

-------
SPRING DAPHNIA RESPONSE  IN AN  URBAN LAKE
TERRY A. NOONAN
Ramsey County Department of Public Works
St. Paul, Minnesota
             ABSTRACT

             Water quality has been monitored in Lake Phalen (St. Paul, Minn.) from 1981 to 1983. Consistent
             increases in epilimnetic Daphnia abundance have been observed each spring, corresponding to
             measured Secchi depth maxima. In 1982 and 1983 the increase in Daphnia abundance followed
             an increase in the density of small algae, principally Cryptomonas sp. The increase in Daphnia
             density in 1981, when small algal forms were not abundant, suggests another factor, such as
             predation, may control spring Daphnia abundance in Phalen. The impact of the spring increase
             in Daphnia abundance on summer water quality appears slight. A significant difference in sum-
             mer mean chlorophyll a concentration between years is not related to a change in epilimnetic
             Daphnia abundance. Summer algal standing crop in Lake Phalen is related to both nutrient and
             biological factors, including partial nitrogen limitation and changes in phytoplankton communi-
             ty composition.
 INTRODUCTION

 A reduction  in  phosphorus  loading  from external
 and/or  internal  sources  with  a  corresponding
 decrease  in  nutrients available  for  summer  algal
 growth is often the primary objective of a lake restora-
 tion program. The positive relationship between sum-
 mer mean concentrations of phosphorus  and  algal
 standing crop has been demonstrated for a broad
 geographic  and morphometric  range of lakes and
 reservoirs (Canfield and Bachman, 1981).
   Seasonally, phosphorus and chlorophyll may not be
 positively related within a lake. Physical factors such
 as temperature and light may limit algal response to
 available  nutrients (Nicholls  and   Dillon,  1978).
 Nitrogen may limit algal growth at certain times of the
 year  (Lehman and Sandgrov, 1978).  Phytoplankton
 community composition may shift to species with dif-
 ferent nutrient uptake requirements and different ex-
 pressions of chlorophyll a per unit phosphorus (Smith,
 1982). Grazing pressure exerted by zooplankton on the
 phytoplankton community may also vary seasonally.
   Spring increases in Daphnia growth rate (Hall, 1964)
 and abundance (Wright, 1965)  have been related to in-
 creased food abundance and rising  water tempera-
 ture. Midsummer declines in Daphnia abundance have
 been attributed to reduced food supply or a shift in the
 phytoplankton community to  blue-greens,  increased
 water temperature, and  invertebrate and vertebrate
 predation  (Hall,  1964; Wright, 1965; Threlkeld,  1979;
 Lynch and Shapiro, 1981).
   Little data exist indicating  limitation of the  algal
 standing  crop  throughout  the  growing season by
 zooplankton grazing. Osgood (1983)  has  postulated
 that Daphnia are limiting the  summer algal standing
 crop  in Square Lake,  Minn.  Hrbacek, et  al. (1978)
 measured relatively low chlorophyll a concentrations
 in a reservoir with a low standing  stock of fish and
 abundant  large  Daphnia. Shapiro (1982) concluded
 that the loss of a spatial refuge from fish predation in
 Lake Harriet, Minn., reduced Daphnia abundance and
 permitted a larger algal standing crop than appeared
 in successive years. Edmondson and Litt (1982) have
 measured increased summer Secchi  depths in Lake
Washington since  1976 corresponding to increased
Daphnia abundance.
  Ramsey County, Minn., was awarded a Clean Lakes
grant  to  implement a restoration  program for the
Phalen chain of lakes in  1977. Monitoring began in
1980 to develop hydrologic and phosphorus budgets
for each  lake in the chain, and to evaluate current
water quality and predict the future benefits of the
restoration program. While the emphasis on water and
phosphorus loading to Lake Phalen is certainly justi-
fied, a consistent spring  increase in Daphnia abun-
dance in the lake has necessitated further considera-
tion of the biological interactions that  may affect
Phalen water quality. The  purpose of this paper is to
describe the seasonal importance of Daphnia in  Lake
Phalen and assess its potential impact on summer
water quality.

STUDY SITE

Lake Phalen is  the final lake of a five lake chain of
lakes located  in St. Paul,  Minn.,  and its northern
suburbs (Fig. 1). Lake Phalen, the deepest and second-
largest lake in the chain (Table 1), is recognized by the
public for its relatively good water quality and is inten-
sively used  in  non-motorized activities,  particularly
fishing, canoeing, and swimming.
  Drainage from the upper watershed is primarily via
two major open ditch systems, Nos. 18 and 16, enter-
ing  lakes Kohlman  and Gervais, respectively (Fig. 1).
Two interstate highways are located within the upper
watershed and two State highways intersect near the
midpoint  of the chain. The contribution of water via
direct runoff and storm sewers increases as you pro-
ceed down the chain paralleling the increase in urban
land use. Lake Phalen receives stormwater via five ma-
jor storm  sewers. Surface  outflow from  Phalen enters
the St. Paul combined sewer system. Lake Phalen is
an important source of groundwater recharge in the
region because of its highly permeable substrate of
glacial outwash sands and  gravels overlying  a deep
glacial valley (Hickok and  Assoc., 1978).
                                                 157

-------
  LAKE AND RESERVOIR MANAGEMENT
                               Table 1.—Physical data—Phalen Chain of Lakes.

Surface area (ha)
Mean depth (m)
Maximum depth (m)
Lake volume (106m3)
Flushing rate(yr~1)
Direct watershed area (ha)
Direct & storm sewer
Wakefield subwatershed
Phalen
85
7.7
27.4
6.6
1.2
747
342
405
Round
11
1 7
6.1
.18
40
343

Gervais
100
5 3
12.1
5.3
1.1
1039

Keller
30
1 **
3.0
.38
18
567

Kohlman
30
1.2
4.3
.36
12.3
2615

  METHODS

  Hydrologic and Phosphorus Budgets

  Flows in the major upstream open ditches and con-
  necting  channels  have  been monitored   since
  September  1980. Storm  sewer  water  loads were
  calculated from precipitation data obtained from two
  locations in the  Phalen watershed and runoff coeffi-
  cients taken from  a previous  study of the Phalen
  watershed   (Hickok  and  Assoc.,  1978).  Monthly
  snowmelt water loading to Phalen was estimated from
  the difference in initial and final snowcover for the
  month (U.S.  Weather Service) assuming evaporation
  loss to be 0.3 inch/day (Baker, 1972). Monthly Phalen
  outflow volume was obtained from the Metropolitan
  Waste Control Commission. Net discharge to ground
  water was determined monthly as the solution of the
  water balance equation.
   Phalen  storm  sewer  phosphorus  loads  were
  estimated from seasonal mean concentration data for
  each  storm  sewer  obtained  in a previous  study
 (Hickok and Assoc.,  1978). Phosphorus  loading to
  Phalen  from Round  and  phosphorus  export  from
  Phalen  were calculated  from   the monthly  mean
 epilimnetic   total  phosphorus  concentrations  and
 outflow volumes.

  Lake Sampling

 Two sites (z  = 27.4 and 13 m) were sampled in Laka
 Phalen monthly from September through  April and
Figure 1.—Watershed map—Phalen chain of lakes.
  semimonthly from  May through August beginning in
  fall, 1980. Vertical dissolved oxygen and temperature
  profiles were obtained at each location using a YSI
  field meter. Three discrete depths (up to 6 depths dur-
  ing overturn) in the mixed layer were sampled at each
  site using a 4-liter Van Dorn bottle. Water was also
  discretely sampled from the metalimnion and near the
  bottom of the water column during periods of thermal
  stratification  (May  through mid-October). Composite
  samples were obtained from the  upper 2 meters for
  phytoplankton analysis. Vertical net tows through the
  epilimnion were made at each site using a 80 ^m mesh
  width, 20 cm mouth diameter net forzooplankton iden-
  tification and enumeration. Epilimnetic mean  values
  were calculated after pooling  the data from the two
  sampling  sites  for each  date.  Whole lake  total
  phosphorus content was calculated by summing the
  volume-weighted  phosphorus concentration through-
  out the water column.

  Laboratory Methods

  Phytoplankton samples were preserved with Lugol's
 solution (1% v/v), settled, and enumerated using the
 clump counting procedure (Standard Methods, 1981).
   Zooplankton samples were preserved in a 5 percent
 formalin solution  and a subsample volume sufficient
 to count  100 individuals  of the  most  abundant
 Cladocera genera or Copepoda suborder was analyz-
 ed.  Daphnia  body length was measured (to a max-
 imum of 50 individuals per sample) and organized into
 seven size classes.
   Chlorophyll a corrected for phaeophytin was deter-
 mined spectrophotometrically  following acetone ex-
 traction (Standard Methods, 1981).
   Total phosphorus concentrations were determined
 by the method of Murphy and Riley (1962) after per-
 sulfate  oxidation  preparation  (Standard  Methods,
 1981). Soluble reactive phosphorus was measured as
 total phosphorus without  digestion  after filtration
 through a 0.45 ^m pore size membrane filter.
   Nitrogen samples were preserved with 0.8 ml con-
 centrated H2 SO4 /I and  frozen until analysis. Total
 Kjeldahl  and  ammonia  nitrogen  were  determined
 using an Orion ammonia electrode (U.S. Environ. Prot.
 Agency,  1979). Nitrate-nitrite nitrogen  was  deter-
 mined colorimetrically following cadmium  reduction
 (U.S. Environ. Prot. Agency, 1979).


 RESULTS

 Hydrologic and Nutrient Budgets

Annually, most (75 percent) of the water to Phalen
comes from the upstream chain of lakes (Table 2).
Water loading during the  ice-covered period  con-
tributes about 34 percent of the annual  load and con-
sists primarily of  snowmelt but does  include some
early spring rains. Surface outflow was important in
the years monitored (27 percent of total  inflow). Water
loss to ground water occurs consistently throughout
                                               158

-------
                                                                                       BIOMANIPULATION
the year but the monthly  loss rates (estimated  by
residual) are variable.
  Upstream lakes, adjacent storm sewers, and direct
drainage  contribute   similar  annual  phosphorus
loading to Phalen (Table 3). Phosphorus export to the
ground water is difficult to quantify but potentially im-
portant. Seasonally, phosphorus export  via surface
outflow may be very important in Phalen. During the
ice-covered period in 1982, phosphorus export via the
outlet was approximately  equal to the  phosphorus
load experienced during the period.
  The ice-covered periods of study were very distinct.
The  winter  of  1981 was  extremely  dry until mid-
February when two large rain storms occurred. Ice-out
on  Phalen was complete  by mid-March 1981. The
winters of 1982 and 1983 (the latter is not included in
loading summaries) included  total  snowfall amounts
of about 100 inches/year. During February 1982 deep
snow caused decreased  (< 1  mg/l) dissolved oxygen
concentrations in Round  Lake, as well as the shallow
upstream  lakes Keller and Kohlman. Phosphorus was
mobilized  from the sediments in Round Lake and  ac-
cumulated under the  ice.  Ice-out  was complete by
April 23, 1982. Despite heavy  snow in 1983, sufficient
dissolved  oxygen (>3 mg/l) was maintained in Round
Lake and the other shallow lakes in the chain at least
through February. No phosphorus buildup was observ-
ed under the ice in Round Lake in 1983 and ice-out oc-
curred by  April 7.

Seasonal  Lake Phalen Water Quality

Following ice-out, epilimnetic total phosphorus con-
centrations were elevated in Phalen in 1982 and 1983,
but not 1981 (Fig. 2). Extremely high phosphorus con-
centrations were measured throughout the water col-
umn after ice-out 1982, apparently reflecting the com-
bined  influence of heavy snowmelt loading via the
storm  sewers  as well as inflow of water  high in
phosphorus  from  Round  Lake.  The  1983  total
phosphorus concentration  peak in Phalen following
Table 2.— Annual hydrologic budget
                (ice-covered period)
                                     — Lake Phalen
INFLOWS
    From Round Lake             5,950
    Direct & storm sewers         1,480
    Precipitation                  570
                   Total        7,990

OUTFLOWS
    Surface outflow               2,190
    Ground water                5,210
    Evaporation                   530
    Change in storage            	60
                   Total        7,990
                                            (2,160)
                                             (400)
                                             (160)
                                            (2,720)
                                            (1,180)
                                            (1,490)

                                              (50)
                                            (2,720)
'Data from average of water years 81 and 82, normalized to average annual
precipitation


  Table 3.—Annual phosphorus budget1 (kg)—Lake Phalen
                 (ice-covered period)

 INPUTS
     From Round Lake              550         (300)
     Direct & storm sewers          630         (150)
     Precipitation & dryfall         	50_        	
                    Total        1,230         (450)

 OUTPUT
     Surface outflow                420         (400)
                                                    ice-out was much less because of lower loading from
                                                    Round Lake and probably lower storm sewer loading
                                                    (loading data for 1983 not summarized as yet). About
                                                    75 percent of the observed spring phosphorus peak in
                                                    1982 was soluble reactive phosphorus (SRP), with
                                                    about 33 percent SRP in 1983 (Fig. 2). SRP concentra-
                                                    tions were low in the epilimnion throughout each sum-
                                                    mer but did increase during autumnal mixing.
                                                      Whole lake phosphorus content followed the dif-
                                                    ferences in Phalen epilimnetic total phosphorus con-
                                                    centrations following ice-out  between the  years of
                                                    study (Fig. 3). In both 1982 and 1983 lake phosphorus
                                                    content  dropped sharply by the next sampling date,
                                                    largely because of phosphorus export via surface
                                                    outflow.
                                                      Secchi  disk  transparency  following  ice-out was
                                                    similarly low between years (Fig. 4). Large increases in
                                                    water transparency were measured  each year in  a
                                                    time sequence that paralleled relative ice-out dates.
                                                      Large increases in Daphnia abundance (Fig. 5) have
                                                    been measured in the mixed water layer each spring.
                                                    Each year the Daphnia increase has corresponded to
                                                    the  measured Secchi depth maximums. Small body
                                                    Daphnia forms,  principally D. galeata, dominated dur-
                                                    ing  the  density  peaks each  spring. Mean Daphnia
                                                    body  length during  the abundance peak  was  not
                                                    significantly  different  between  years  (1981-1983
                                                    means  were  0.8,  1.0,  and  0.9 mm,  respectively).
                                                    Daphnia mixed layer abundance was reduced by June
                                                    each year, and remained low for the remainder  of the
                                                    summer in both 1981 and 1982 (Fig. 5). Daphnia densi-
                                                    ty maxima and phytoplankton density minima (Fig. 6)
                                                    were related in summer 1982 but not  1981.
                                                      Water  samples taken from  discrete  depths for
                                                    suspended solids analysis in 1982 indicate that large
                                                       300|
                                                       200-
                                                        IOO-
                                                         0

                                                        400
                                                       300
                                                        200
                                                        100
                                                        300
                                                        200
                                                     E  100
                                                           1981
                                                           MAR   APR
                                                                      MAY
                                                                            JUN
                                                                                   JUL
                                                                                          AUG    SEP
                                                           I982
                                                                                              SRP n
                                                           MAR  APR     MAY    JUN    JUL
                                                                                          AUG    SEP
                                                            I983
                                                                                              SRP A
                                                           MAR  APR
                                                                       MAY
                                                                             JUN
                                                                                    JUL
                                                                                           AUG
                                                                                                 SEP
 'Data from average of water years 81 and 82.
                                                     Figure 2.—Epilimnetic total phosphorus and soluble reactive
                                                     phosphorus (SRP) concentrations for Lake Phalen. Vertical
                                                     lines  are  95  percent confidence  intervals  for  total
                                                     phosphorus for each sampling date SRP concentration less
                                                     than 10 mg/m3 unless shown.

-------
 LAKE AND RESERVOIR MANAGEMENT

 Daphnia may have been abundant in the metalimnicn
 in midsummer. On each of three dates (June 28, Ju y
 15 and 28) large Daphnia were observed at a depth of
 6.4 m. This depth was associated with dissolved  ox-
 ygen concentrations < 1 mg/l, with higher concentra-
 tions (2-3 mg/l) immediately above it. Representative
 summer oxygen  profiles for each  year  are given in
   2400
   2000
 8 1600
 I 1200
    800
    400
I98I
I982
I983
       MAR  APR
                          JUN
                                JUL
                                       AUG
                                              SEP
 Figure 3.—Whole-lake  total  phosphorus content—Laks
 Phalen.
E 3
o
                                            I98I  •
                                            1982  •
                                            I983  A
                        I
                                     I
     MAR    APR    MAY    JUN    JUL     AUG    SEP

Figure 4.—Secchi depth—Lake Phalen.
     JAN  FEB MAR APR  MAY  JUN  JUL AUG  SEP  OCT  NOV  DEC
Figures.—Epilimnetic Daphnia abundance—Lake Phalen.
Figure 7. No large increase in Daphnia abundance in
the mixed layer was measured in the late fall of 1980
or 1981. No zooplankton data are available in fall 1982
beyond September.
  Differences in spring phytoplankton  composition
are evident. After ice-out in 1982, blue-greens (primari-
ly Oscillatoria)  and flagellates (Cryptomonas) were
dominant. Phytoplankton abundance was low by May
19, 1982, corresponding to the increased water trans-
parency measured on that date. Phytoplankton com-
munity composition following ice-out in 1983 included
flagellates, filamentous blue-greens, as well as green
                   JAN FEB  MAR  APR MAY  JUN  JUL  AUG  SEP OCT NOV DEC
                                                            JAN  FEB MAR «PR  MAY  JUN  JUL AUG SEP OCT  NOV  DEC
                                                        3,800


                                                        3,400


                                                        3,000


                                                        2,600
                                                        1,800
                                                      1  1,400
                                                        1,000
                                               BLUE-GREEN  D
                                               GREEN      D
                                               CRYPTOMONAS  •
                                               OTHER      D
                                                            JAN FEB  MAR  APR  MAY  JUN JUL  AUG  SEP  OCT  NOV DEC
            Figure 6.—Phytoplankton abundance and community com-
            position—Lake Phalen.
                                                  160

-------
                                                                                     BIOMANIPULATION
algae (mostly Ankistrodesmus) in similar proportions.
Maximum spring  Secchi depth in  1983 also cor-
responded  to  a low  density  of  phytoplankton  in
Phalen.  Blue-greens  dominated the  phytoplankton
community after ice-out 1981. Phytoplankton abun-
dance increased through the spring  1981 sampling
and the  moderate Secchi depth maximum (4.6 m) did
not  correspond  to a  decrease  in  algal  density.
Phytoplankton community composition through the
summer  1981  was dominated  by filamentous  blue-
greens.  Phytoplankton  abundance in  summer  1982
was  low and community composition was dominated
by blue-greens to a lesser extent than 1981 except for
August.  A fairly large proportion of green algae per-
sisted  into   the  fall  of  1982.  Summer  1983
phytoplankton data are incomplete but early summer
dominance by blue-green algae is apparent.
  Summer (May-August)  mean mixed  layer  total
phosphorus concentration was significantly higher in
1982 relative to 1981 (Table 4). Sampling date mean
phosphorus concentration was consistently higher in
1982 than  1981, but greater  variation in measured
phosphorus concentration in the mixed layer was also
observed (Fig. 2). This variation was caused primarily
by occasional unidentified  layers of higher  total
            DISSOLVED   OXYGEN   (mg/l)

        04         8         12       16
      0
      5 -,
      10 -
      15 -
      0
 CL
 LU
 Q
      10
      15
     20
                        MAY  28
                                       I98I
      I
   AUGUST 26'

JULY  I5V
                              MAY  19
                                      I982
                                       phosphorus concentration in the mixed layer (at both
                                       sampling sites) during 1982.
                                         Summer mean chlorophyll a concentration was
                                       significantly less, and mean Secchi depth significant-
                                       ly greater, in 1982 than 1981 (Table 4, Fig. 8 and 4). The
                                       summer 1981 mean relationship between chlorophyll
                                       a and total phosphorus falls within the 90 percent con-
                                       fidence interval calculated  from  a study (Osgood,
                                       1981)  of  60 Twin  Cities area  lakes.  The mean
                                       chlorophyll yield for the 1982 mean total phosphorus
                                       concentration in  Phalen is significantly lower than
                                       predicted by the regional phosphorus-chlorophyll rela-
                                       tionship, however.
                                         Summer TN:TP ratios differed significantly between
                                       years (Table 4, Fig. 9). Mixed  layer mean TKN concen-
                                       trations did not differ significantly between summers.
                                       The early summer decline in TN:TP ratios observed
                                       each year was related to a decrease in TKN concentra-
                                       tion.The summer 1982 TN:TP minimum (10.6) occurred
                                       at this time. The TN:TP ratio  in 1981 was highest dur-
                                       ing  late  summer (July-August) corresponding to a
                                       decrease in epilimnetic total phosphorus  concentra-
                                       tions. Phalen was clearly phosphorus limited in 1981
                                       with maximum  chlorophyll concentrations occurring
                                       in late June, prior to the decrease in total phosphorus
                                       concentrations. TN:TP data indicate that nitrogen may
                                       have been limiting to algal  growth for much of the
                                       period June-August, 1982.
                                        DISCUSSION
                                        Spring increases in soluble reactive phosphorus con-
                                        centration in Phalen, primarily from snowmelt loading,
Figure 8.—Chlorophyll a concentrations—Lake Phalen. Ver-
tical lines are 95 percent confidence intervals for each date.


Table 4.—May-August  mixed layer  mean  data—Lake
                     Phalen.
                                                    Parameter
                                                               Year
                               Mean
SO
Total Phosphorus (mg/l)

Chlorophyll a (mg/m3)

Secchi depth (m)

TKN (mg/l)

TN:TP

1981**
1982
1981*
1982
1981*
1982
1981" s
1982
1981*
1982
.028
.051
10.7
2.6
2.2
3.2
.88
.74
37
17
.011
.027
3.9
2.2
.73
1.6
.22
.23
16.8
6.7
46
47
43
43
16
16
31
30
16
16
 Figure 7.—Representative summer dissolved oxygen profiles
 for 1981 and 1982—Lake Phalen.
                                        *  p < .01 that summer means are equal using Student's t-test
                                        *' p< 05
                                        n.s. p > .05
                                                 161

-------
 LAKE AND RESERVOIR MANAGEMENT
 observed in 1982 and 1983 were  used in growth by
 small flagellates and green algae (under suitable light
 and temperature conditions). The  observed  peaks in
 algal density were followed by large spring increases
 in water transparency that correspond to increases in
 Daphnia abundance. The precise mechanism and ex-
 tent (i.e. does Daphnia abundance limit algal  standing
 crop?) of this seasonal trophic interaction is unclear
 from this study.
   Rapid  decreases  in spring  phytoplankton abun-
 dance have  been  attributed  to  the  depletion of
 available phosphorus and a reduction  in growth rate
 (Lamport,  1978). In  Lake Phalen,  the reduction in
 spring epilimnetic phosphorus concentration observ-
 ed in  1982  and 1983 was also enhanced by surface
 outflow.
   The occurrence of a  Daphnia maximum  in 1981
 without an increase in spring phosphorus concentra.-
 tion or a corresponding increase in the abundance of
 small  algae, indicates that another factor,  perhaps
 predation, may control spring  Daphnia  abundance'.
 Lamport (1978) observed a similar consistent spring in-
 crease in  water transparency in  Lake  Constance.
 Following an increase in  flagellate biomass in May,
 the  standing  crop of Daphnia  increased.  Daphnia
 growth rate  did  not  vary  with  the  increase in
 phytoplankton  standing crop, however. Lamport con-
 cluded that predation by Cyclops vicinus on  Daphnia
 in the early spring allowed the phytoplankton peak to
 develop once physical conditions  (particularly light)
 for growth were met.
   No direct evidence suggests that Daphnia grazing
 is controlling summer algal standing crops in Lake
 Phalen or that the difference in algal standing crop
 between summers is related to a change in  the hei-
 bivore population.  Epilimnetic Daphnia concentra-
 tions are reduced in  Phalen in the summer.  Shapiro
 (1982) in Lake Harriet and Threlkeld (1979) in Winter-
 green  Lake found high concentrations  of Daphnia in
 the metalimnion, while Edmondson and Litt  (1982)
 observed that Daphnia were not abundant below the
 epilimnion in  Lake Washington. The observation of
 large Daphnia in the metalimnion only in summer 1982
 offers the potential that summer Daphnia concentra-
 tions in Phalen were underestimated in this study and
 that differences in the herbivore population between
 summers were  not identified. There were no apparent
 differences  in  the structure of the  Daphnia refuge
 (Shapiro, 1982)  between years.
   Differences in phytoplankton community composi-
 tion may also  have  contributed to  differential  her-
 bivore grazing pressure between years. A significant
  50
  40
T 30
z
t-
  20
    -   I98I
       1982
      MAR   APR    MAY    JUN    JUL    AUG

Figure 9.—Epilimnetic TN:TP—Lake Phalen.
         portion of the phytoplankton community through July
         in 1982  consisted  of  flagellates  and green algae,
         presumably providing a more suitable food source for
         herbivorous zooplankton  than  blue-greens  (Arnold,
         1971). In  1982 an increase in chlorophyll a concentra-
         tion  occurred  in  August  only  when blue-greens
         became clearly dominant.
           Since  the summer standing crop may have been
         underestimated, relationships that might indirectly
         reflect intensive zooplankton grazing must  be con-
         sidered.  Carlson and Schoenberg (1983) postulated
         that  zooplankton grazing  may suppress blue-greens
         indirectly by changing environmental conditions, such
         as pH. A high  percentage of phaeophytin (Shapiro,
         1982) and a  large  ratio of soluble reactive  to total
         phosphorus (Carlson and Schoenberg, 1983) have
         been  related  to intensive zooplankton grazing.  No
         significant difference in  such indirect measures of
         zooplankton grazing was observed in Phalen between
         summers.
           The significant difference in TN:TP ratios between
         summers indicates that nitrogen and phosphorus are
         both important  in regulating summer algal standing
         crops in Lake Phalen. The summer chlorophyll yield in
         this lake is often not directly related to the concentra-
         tion of either nutrient but  is rather a function of  the
         phytoplankton community composition. The equation
         of Smith  (1982)  incorporating mean total phosphorus
         and nitrogen  concentrations and the variable yield
         model of  Smith and Shapiro (1981) predict a slightly
         greater mean  chlorophyll  concentration in  1982 than
         in 1981. In fact, the relationship of higher mean TN:TP
         and  lower mean  total  phosphorus  concentration
         observed  in 1981 is analogous to the changes in these
         parameters expected following a restoration program
         (Smith, 1981),  but the mean chlorophyll concentration
         was increased (not  expected). The effect of a  change
         in TN:TP  on algal standing crop has been  related to
         the specific nutrient  uptake  physiology of  individual
         algal species (Smith,  1982), differences in "optimum"
         TN:TP  for different species (Rhee, 1978), and dif-
         ferences  in phytoplankton community composition
         (Sakamoto, 1966).
        SUMMARY

        In 1982 Lake Phalen responded to a large change in
        nutrient  conditions by  reducing the algal standing
        crop. Nitrogen was apparently the limiting nutrient in
        the  lake  for much  of  the summer. The observed
        decrease in mean chlorophyll concentration would not
        be predicted strictly on the basis of a different limiting
        nutrient, however, because nitrogen concentrations
        were similar between summers.  Phytoplankton com-
        munity composition is clearly important and may ex-
        plain a significant portion of the measured difference
        in algal  standing crop between years.  Seasonal
        trophic  interactions are also important  in Phalen.
        Spring Daphnia abundance peaks  are consistently
        associated with maximum water transparency. There
        is no direct evidence  that  herbivore grazing  was
        related to the reduced algal standing crop in 1982. It is
        hoped that 1983  data and  future monitoring  will in-
        crease understanding of the biological and nutrient
        relationships in this lake.
———'   ACKNOWLEDGEMENTS: The efforts of Lewis Soukup in the
        field and lab were substantial  and much appreciated. Dick
        Osgood and Hal Runke provided many helpful comments
        that have been included in this paper.
                                                 162

-------
                                                                                               BIOMANIPULATION
REFERENCES

Arnold, E, 1971. Ingestion, assimilation, survival, and repro-
  duction by Daphnia pulex fed seven species of blue-green
  algae. Limnol. Oceanogr. 16: 906-20.
Baker, D.  1972. On the prediction of spring runoff.  Water
  Resour. Res. 8: 966-72.
Canfield, D.E., and R.W. Bachmann. 1981. Prediction of total
  phosphorus  concentrations, chlorophyll a  and Secchi
  depths in natural and artificial lakes. Can. J. Fish. Aquat.
  Sci. 38: 414-23.
Carlson, R.E., and S.A. Schoenberg. 1983. Controlling blue-
  green algae by zooplankton grazing. Pages 228-233 in Lake
  Restoration, Protection  and Management. Proc. Second
  Annu. Conf. N. Am.  Lake Manage. Soc. EPA  440/5-83-001.
  U.S. Environ. Prot. Agency, Washington, D.C.
Edmondson, W.I., and A.H.  Litt. 1982. Daphnia  in  Lake
  Washington. Limnol. Oceanogr. 27: 272-93.
Hall, D.J. 1964. An experimental approach to the dynamics
  of a natural  population  of Daphnia  galeata  mendotae.
  Ecology 45: 94-112.
Hickok,  E.A., and Associates. 1978. Lake Phalen  Restora-
  tion Project. Prel. Rep. to Ramsey County.
Hrbacek, J., B. Desortova,  and J.  Popovsky. 1978. Influence
  of fishstock on the phosphorus-chlorophyll ratio. Int. Ver.
  Theor. Angew Limnol. Verh. 20: 1624-8.
Lampert, W. 1978. Climatic conditions and planktonic inter-
  actions as factors controlling the regular succession of
  spring algal bloom and extremely clear water in Lake Con-
  stance.  Int. Ver. Theor. Angew Limnol. Verh.  20: 969-74.
Lehman,  J.T., and C.D. Sandgren.  1978. Documenting  a
  seasonal change from phosphorus to nitrogen limitation
  in a small temperate lake, and its impact on the population
  dynamics of Asterionella. Int. Ver. Theor. Angew Limnol.
  Verh. 20: 375-80.
Lynch, M., and J. Shapiro.  1981. Predation, enrichment, and
  phytoplankton community structure.  Limnol.  Oceanogr.
  26: 86-102.
Murphy, J., and  J.P.  Riley.  1962. A modified single  solution
  method for the determination  of phosphate  in natural
  waters. Anal. Chim. Acta. 27: 31-6.
Nicholls, K.H., and P.J. Dillon. 1978. An evaluation of phos-
  phorus-chlorophyll-phytoplankton relationships in  lakes.
  Int. Rev. Gesamten Hydrobiol. 63:141-54.
Osgood, R. 1981. A study of the water quality of 60 lakes in
  the seven county Metropolitan Area. Metropolitan Counc.
  No. 01-81-047. St. Paul, Minn.
	1984. Long term  grazing control  of algal  abun-
  dance: A case history. In Lake and Reservoir Management.
  Proc. N. Am. Lake Manage. Soc. Oct.  18-20, 1983.  Knox-
  ville, Tenn.
Rhee, G-Y.  1978. Effects of N:P atomic  ratios and nitrate
  limitation on algal growth, cell composition, and nitrate
  uptake. Limnol. Oceanogr. 23:10-25.
Sakamoto,  M. 1966. Primary production  by phytoplankton
  community in some Japanese lakes and its dependence
  on lake depth. Arch. Hydrobiol. 62:1-28.
Shapiro, J. 1982. The role of physical-chemical conditions in
  affecting algal abundance - Lake Harriet. Pages 235-251 in
  Experiments and Experiences in Biomanipulation: Studies
  of  Biological Ways  to  Reduce  Algal Abundance and
  Eliminate Blue-Greens.  EPA-600/3-82-096.  U.S.  Environ.
  Prot. Agency, Washington, D.C.
Smith, V. 1982. The nitrogen and phosphorus dependence of
  algal  biomass  in  lakes: An empirical and theoretical
  analysis. Limnol. Oceanogr. 27:1101-12.
Smith, V.H., and J. Shapiro. 1981. A retrospective look at the
  effects of phosphorus removal in lakes.  Pages 73-77 in
  Restoration of Lakes and Inland Waters. EPA 440/5-81-010.
  U.S. Environ. Prot. Agency, Washington, D.C.
Standard Methods for the Examination of Water and Waste-
  water. 1981.15th edition. Am. Pub. Health Ass.
Threlkeld, S.T.  1979.  The  midsummer  dynamics of two
  Daphnia species in Wintergreen Lake,  Michigan. Ecology
  60:165-79.
U.S. Environmental  Protection Agency. 1979. Methods for
  chemical analysis of water and wastes. EPA-600/4-79-020.
  Washington, D.C.
Wright, J. 1965. The population dynamics and production of
  Daphnia  in Canyon  Ferry  Reservoir,  Montana. Limnol.
  Oceanogr. 10: 583-90.
                                                      163

-------
                                 Modeling  Techniques
                                              and  Innovations
USE OF A PREDICTIVE PHOSPHORUS MODEL TO EVALUATE
HYPOLIMNETIC  DISCHARGE SCENARIOS FOR LAKE
WALLENPAUPACK
H. KIRK HORSTMAN
ROGER S. COPP
FRANK X. BROWNE
F. X. Browne Associates, Inc.
Lansdale, Pennsylvania
          ABSTRACT

          A vertically-segmented, dynamic phosphorus model program was developed for use on a 16 bit
          microcomputer. The predictive model includes theoretical equations for physical settling, ther-
          mocline diffusion, and sediment release of total phosphorus. The model was developed and
          calibrated using actual data collected from 1980 through 1982. Input loads from combined point
          and nonpoint sources were entered using a time step equal to 1 day. In the case of Lake Wallen-
          paupack, discharge occurs from both the epilimnion and hypolimnion during the latter portion of
          the stratified period. Several discharge scenarios were evaluated to determine if any one would
          produce a significant  reduction in summer epilimnetic phosphorus concentrations.
INTRODUCTION

Lake Wallenpaupack is a 14,000 hectare multipurpose
impoundment located in northeastern Pennsylvania.
The lake has a maximum depth of 16 meters and a
mean hydraulic residence time of 0.6 years. The lake
was built in 1925 by the Pennsylvania Power and Light
Company (PP&L) for hydroelectric generation. A 4.3
meter diameter discharge pipe is located at the base
of the dam.
  Although PP&L owns the lake bottom and some sur-
rounding land, the remainder of the 56,700 hectare
watershed is owned by public and private concerns.
Over the years recreational  use of the lake has be-
come significant to the point where  it is one of the
most important tourist resources in the Pocono Moun-
tain  region.  Many  second-home  developments,
resorts, marinas, and support businesses are located
around  the lake. Land development has accelerated
the eutrophication process and caused serious water
quality problems. The lake suffers from severe dissolv-
ed oxygen depletion, an altered fishery, reduced trans-
parency, and obnoxious  blue-green algal blooms. In
August 1979, although never thoroughly documented,
a bloom of Anabaena reportedly caused numerous
cases of algae-related infections that produced such
symptoms as allergic reactions and gastrointestinal
disorders. This outbreak of illness led to the posting of
warning signs around the lakes.
  After the formation of the Lake Wallenpaupack
Watershed Management District, an EPA Clean Lakes
Program Phase  I Study was performed (F. X. Browne
Associates, Inc., 1982). The study concluded that the
lake is eutrophic and that phosphorus is the primary
nutrient responsible  for  controlling phytoplankton
growth according  to data provided  by algal assays
and chemical ratios. One of the recommendations of
the report was to further investigate the potential ef-
                                          165

-------
 LAKE AND RESERVOIR MANAGEMENT
 fects of  altering  PP&L's  hydroelectric discharge
 policies in order to improve water quality. Since no
 capital expense would be involved, it was hoped that a
 measurable benefit could be achieved at a relatively
 small cost to either the utility or local communities. In
 addition,  these changes  could possibly  be   im-
 plemented almost immediately.
 LAKE CHARACTERISTICS

 Although thermal stratification usually begins in mid-
 May in  Lake Wallenpaupack, a  stable thermocline
 does not form until early June. Anoxic conditions in
 the hypolimnion usually develop by the middle of July
 and last until the fall turnover period which generally
 occurs sometime  in October depending upon mete-
 orological conditions and PP&L's discharge practices.
   As determined by a review of lake temperature data
 collected during the summers of 1981 and 1982, the
 actual thickness and elevation of the thermocline (the
 zone where temperature changed at a rate of 1° Cel-
 sius or more per meter of depth) varied significantly
 according to inflow, outflow, and meteorological con-
 ditions. Attempts to predict thermocline characteris-
 tics based strictly on hydraulic factors or by using em-
 pirical relationships were unsuccessful. A complex
 thermal model was beyond the scope  of this study.
   Therefore, the lake was assumed to consist of tv/o
 hydraulic compartments,  a combined epilimnion/
 metalimnion and a hypolimnion, as shown in Figure 1.
 Based on the thermocline data presented in Figure 2,
 a hypothetical  hypolimnion was defined  as that
 volume  below the elevation  on  1,149 feet (MSL).
 Temperature conditions remained fairly stable below
 this elevation. In addition, after the onset of thermal
 stratification, dissolved  oxygen  was  almost totally
 devoid  in this bottom  layer. The selection of this
 volume to represent the "pure" hypolimnion was also
 supported by a total phosphorus profile developed by
 collecting samples at 1  meter intervals at two stations
 in the lake on Aug. 10,1982. The results of the sample
 analysis showed that  total phosphorus concentra-
 tions increased significantly within 1  to 2  meters of
 the lake bottom (i.e., sediment surface).
   As shown in Figure 1, the large diameter of  the
 outlet  pipe  causes outflow  to occur from bolh
 hydraulic compartments.  The  magnitude  of  the
 outflow from each compartment  was assumed to be
 proportional to the cross-sectional areas of the pipe
 above and below the dividing elevation. A net ground-
 water outflow was assumed for each compartment
 based  on previous hydraulic budgets  calculated  for
 the lake.
LAKE MODEL THEORY

A  computer program  was  developed to  predict
phosphorus concentrations at various  levels in the
lake based on hydraulic factors and phosphorus in-
teractions within the lake. Since most empirical input-
output  phosphorus  models   assume  a  constant
volume, completely mixed lake, they are not sufficient
for  application  to Lake Wallenpaupack.  Instead, a
model which is capable of predicting temporal and
spatial relationships  was  required. Also,  the model
must  address  the  issue of  sediment release of
phosphorus  rather than  simply  considering  the
sediments to be a net  long-term depository.
  Although increased  predictive error can usually be
expected (unless extensive monitoring and calibration
 are performed), a mechanistic ecosystem simulation
 model was used. In addition to the physical transport
 of total phosphorus in conjunction with water flow,
 several other possible phosphorus reactions were in-
 cluded in the model. As shown in Figure 3, these were
 (1) settling, (2) vertical  exchange through  thermocline
 (from both molecular and turbulent diffusion), and (3)
 sediment release.
   In general, these reactions were aggregated, mean-
 ing that physical, biological, or chemical processes
 having similar effects  were combined into three sets
 of coefficients. An attempt to define each of the possi-
 ble processes individually would require a much more
 complex model. Also,  the model was developed for
 total phosphorus only. No distinction was made be-
 tween soluble and insoluble  forms. The type of model
 used was considered detailed enough for the objec-
 tives of this study, and for the amount of background
 data available. Therefore, all of the potential sources
 and sinks for phosphorus as they were assumed for
 the model are shown in Figure 4.
   The model used is  dynamic; a set of differential
 equations are used along with a set of predetermined
 rate coefficients in an attempt to predict the effects of
 various ecosystem interactions  on the  quantity of
 total phosphorus present in the water column. In addi-
 tion to the starting conditions (the lake elevation and
 initial phosphorus concentrations), the input data re-
 quired are inflows, total phosphorus loads, and out-
 flows. This information was developed  and entered in-
 to the computer program using a time  step equal to 1
 day. The  model was developed based on work by
 Figure 1.—Schematic diagram of lake profile showing ver-
 tical compartments and flow components.
                  Bottom of Hetallmnl
                                        • - Lake Bottom
                              Aug.    Sept.
Figure 2.—Elevations of lake surface and bottom of metalim-
nion at a mid-lake station for 1981 and 1982.
                                                 166

-------
                                                                     MODELING TECHNIQUES AND INNOVATIONS
                Epi1imnion/Hetalimnion
        Settling
    Vertical
    Exchange
                Hypo]imnion
        Settling
    Sediment
    Release
         Sediment
         Release
                      Sediments
Figure 3.—Possible aggregate phosphorus reactions due to
physical, chemical, and biological processes.
   Tributary
 £ Wastewater
   Inputs
                        Atmospheric
                        inputs
                    Epi]imnion/
                    Metal imnion
                0 is charge
         Sediment
          Release
                  ettlIng
 Vertical
 Exchange
                (upper portion
                 of pipe)
                                    Groundwater
                                     Output
Inter-
 flow
                    Hypolimnion
             Sediment
             Release
Settling
                                       Di scharge
                                       (lower portion
                                        of pipe)
         Groundwater
           Output
Figure 4.—Potential sources and sinks for phosphorus as
defined in mechanistic lake model.
Chapra and Reckhow (1983). The primary assumptions
of the model are:
   1. Vertical  variations  in  phosphorus  are more
significant than horizontal variations;
   2. Horizontal lake compartments  are completely
mixed;
   3. Phosphorus settles at a constant velocity from
each of the lake compartments;
   4. Phosphorus is  released at a constant rate from
sediments for either oxic or anoxic conditions; and
   5. Vertical exchange of  phosphorus  through  the
thermocline  is a factor of lake depth,  thermocline
thickness, and  concentration differential  between the
epilimnion/metalimnion and the hypolimnion.
   The mass balance equations used for this internally
descriptive phosphorus model (Chapra, 1983) were:

Epilimnion
V^dc^dt) =  L + W-Q-iCT- Qic1-Qaci +
             D12(c2 - Ci)Ah - vs1 AhC! + JS(AS - Ah)
Hypolimnion
V2(dc2/dt) =  QjC-i - Q2c2 - Qbc2 + vs1 A^ - D12(c2 - c^
            Ah-vs2Ahc2 + JsAh

  where Ah  = surface area of hypolimnion (L2)
        As  = surface area of lake (L2)
        c-|  = total phosphorus concentration in
               epilimnion/metalimnion (MIL3)
        c2  = total phosphorus concentration in
               hypolimnion (MIL3)
        D12 = vertical exchange coefficient (LIT)
        Js  = sediment release coefficient
               (M/L.2T)
        L  = tributary and wastewater effluent
               total phosphorus loading rate (M/T)
        QI  = pipeline outflow from epilim-
               nion/metalimnion (L3/T)
        Q2  = pipeline outflow from hypolimnion
               (L3/T)
        Qa  = groundwater outflow from epilim-
               nion/
               metalimnion (L3/T)
        Qb  = groundwater outflow from hypolim-
               nion (L3/T)
        QI  = interflow between epilimnion/
               metalimnion and hypolimnion
               (L3/T)
        t   = time (T)
        vsi  = total phosphorus settling velocity
               in epilimnion/metalimnion (LIT)
        vs2 = total phosphorus velocity rate in
               hypolimnion (L/T)
        V-j  = volume of epilimnion/metalimnion
               (l_3)
        V2  = volume of hypolimnion (L3)
        W  = atmospheric total phosphorus
               loading rate (M/T)
  These equations  were integrated and  solved for the
total phosphorus concentrations on a daily basis for
both the thermally stratified and nonstratified periods.
The rate of sedimentation of total phosphorus  was
considered  to be a function of the hypolimnetic  sur-
face area, Ah (Chapra, 1975). The apparent settling
velocities, vsi  and vs2, were assumed to include the
sedimentation of phytoplankton and organic detritus.
The apparent settling velocity in the epilimnion/metal-
imnion  was considered to be higher  than for  the
hypolimnion because of the larger concentrations of
phytoplankton cells, portions of which are bacterially
degraded  after settling into the hypolimnion. Also, the
homogeneous settling rate for the nonstratified period
was assumed  to be lower because of generally lower
phytoplankton concentrations throughout  the lake.
The following  ranges for apparent settling velocity
were found  in  the literature:
  0.05 -  0.6 m/day (Chapra and Reckhow, 1983)
  0.1 - 0.4 m/day (Imboden, 1974)
  According to Snodgrass and O'Melia (1975), the ver-
tical exchange coefficient  includes  the effects of
molecular and turbulent diffusion, internal waves,  ero-
sion of the hypolimnion, and other fluid  processes on
the transfer of materials across the thermocline. They
present the  following equation:
     D12(m/day)  = 0.005 Z
where
     Z equals  mean lake depth (m).
  For Lake  Wallenpaupack, therefore, D^2 is equal to
approximately  0.045  m/day  during the  stratified
period. This  coefficient was assumed to increase by a
factor of three for the nonstratified period,  since the
thermal barriers against diffusion are  not present.
                                                   167

-------
 LAKE AND RESERVOIR MANAGEMENT
   The amount of phosphorus which is released from
 lake  sediment is a complex function  of  physical,
 chemical,  and  biological reactions.  Besides  the
 degree of turbulence at the sediment-water interface,
 the stoicheometry of iron  usually has the largest im-
 pact  on the significance of phosphorus release.  For
 noncalcareous lakes such as Lake Wallenpaupack,
 phosphorus recycle is usually less significant during
 both  oxic and anoxic periods because relatively high
 amounts of iron  are available for precipitation.  For
 noncalcareous lakes, Stauffer (1981) reported release
 rates for soluble reactive phosphorus of less than O.CiO
 mg/m2/day for oxic conditions and up to 3.8 mg/m2/day
 for anoxic conditions. For sediments comprised main-
 ly of refractory  organic  silt, Mawson, et al.  (1983)
 reported release  rates for total  phosphorus of 0.04
 mg/m2/day for oxic conditions and 3.05 mg/m2/day for
 anoxic conditions. Based on this data, the release rate
 of total  phosphorus  from the  sediments  in Lake
 Wallenpaupack during oxic periods was assumed 1o
 be 1/50 of that during anoxic periods.


 MODEL CALIBRATION

To refine the values for the rate coefficients which are
available in  the literature, the coefficients for Lake
Wallenpaupack were calibrated using watershed and
lake data collected for the 4-month period from June
through Ocotober 1982. Lake operations were prac-
tically normal  for this   period  and  no  extreme
hydrologic  conditions  occurred,  although  overall
tributary inflows  were lower than usual. After  11
calibration  runs,  the  following  coefficients  were
selected:
   Settling rate (epilimnion)  = 0.19 m/day
   Settling rate (hypolimnion) = 0.09 m/day
   Vertical exchange        = 0.15 m2/day
   Sediment release (anoxic) = 0.0025 g/m2/day
  The results obtained using these coefficients are
presented in Table 1. The "actual"  concentrations
shown for the epilimnion/metalimnion represent the
means for all surface and middle samples collected
on each respective date. The "actual" concentrations
for the hypolimnion represent the means for all bot-
tom samples collected below 1149 feet (MSL). The
mean predicted values for the epilimnion/metalimnion
and the hypolimnion were  both within 5 percent of
their  respective actual  mean values. The predicted
concentrations for the final date were also quite close.
Only two of the 14 predicted concentrations were oil
by significantly more than 50 percent. This is well
within the range  of acceptability for this  type of
model. Although dynamic, the model still does not ac-
 count  for  many  important  factors  such  as
 phytoplankton  dynamics  (for example, rapid algae
 blooms), meteorological effects (e.g., sun, wind, rain),
 and physical water  conditions (e.g.,  currents  and
 unusual temperature fluctuations). Therefore, the total
 phosphorus concentrations as predicted by the model
 changed gradually and did not correspond directly
 with the actual daily conditions experienced  in the
 lake. Model accuracy was better for the epilimnion/
 metalimnion, which is the primary compartment of
 concern.

 INPUT FLOWS AND LOADS

 A "normalized" hydrologic year was developed using
 long-term outflow information provided by PP&L. The
 generated inflow information was used in conjunction
 with total phosphorus loading equations developed as
 part of  past studies in the watershed (Horstman and
 Browne, 1982). Total phosphorus loads for both base
 loads and storm loads were therefore  estimated for
 the normalized hydrologic year (see Table 2). These
 loads  were  checked  using  literature values  for
 phosphorus  export coefficients  for  different  land
 uses.
   Different input files could be created and entered in-
 to the computer program  to evaluate other types of
 hydrological patterns (either real or theoretical). Also
 the effects of certain watershed management prac-
 tices could be evaluated using the model.
CURRENT LAKE OPERATIONS POLICY

Although the 44 megawatt hydropower station com-
prises only a small fraction of PP&L's total generating
capacity, it is still an  important source of electricity
during these times of expensive energy resources. It is
particularly  valuable  during  peak  energy  demand
periods. PP&L has had to consider many factors in the
development of their current lake operations  policy.
Some of these considerations are:
  1.  Lake elevation
     a.  Dam safety (including  both minimum and
       maximum water surface elevations)
     b. Recreational lake user requirements
     c. Ice cover
  2.  Hydroelectric power generation
     a. Maximum capacity of turbines
     b.  Reserve  capacity (e.g., to prevent  pipeline
       freeze-up during winter)
     c.  Avoidance of loss of water over spillway
     d.  Peak power demand periods
         Table 1.—Predicted versus actual total phosphorus concentrations for 1982 model calibration period.

                                    Total Phosphorus Concentrations
                        Epilimnion/Metalimnion
                       Hypolimnion
Date
(1982)
6/1 6(a)
7/6
7/21
8/3
8/10
8/26
9/15
10/19
Mean
Actual
44
27
37
39
28
17
40
21
31.6
Predicted
(MO/D
44.0
40.7
36.4
33.2
31.7
29.0
25.7
22.1
32.8
Difference
0.0
50.7
-1.6
- 14.9
13.2
70.6
- 35.8
5.2

Actual
55
32
90
62
77
69
113
21
64.9
Predicted
55.0
69.0
63.6
77.0
74.0
68.2
64.3
24.2
61.9
Difference
0.0
115.6
-29.3
24.2
-3.9
-1.02
-43.1
15.2

 (a) This date used to establish initial conditions.
                                                168

-------
                                                                  MODELING TECHNIQUES AND INNOVATIONS
  3.  Costs to utility
     a. Seasonal value of energy
     b. Workload requirements
  4.  Downstream requirements
     a. Flow augmentation for Delaware River (during
       low flow periods only)
     b. Fishermen's  safety  in  Lackawaxen  River
       during fishing season
     c. Dissolved oxygen levels in Lackawaxen River
       during low flow conditions
     d. Ice blockage in Lackawaxen River
     e. Flooding prevention

  When hydrologic conditions permit, all of the water
discharged from the lake is withdrawn through the
hydroelectric outlet  pipe. The elevation of the top of
the spillway roller gates  is 1,190 feet MSL According
to their  Federal  Energy  Regulatory  Commission
license, PP&L  must  maintain  the  water elevation
below  1,182  feet  MSL  between  August  1  and
November 15 to allow for the storage of potentially
large quantities of runoff which may be produced dur-
ing the hurricane season.
  Based on all of these considerations, PP&L has
established the target elevations shown for Scenario
A in  Figure 5.

ALTERNATIVE DISCHARGE SCENARIOS

In addition to the  discharge  rates necessary to
achieve the elevations for Scenario A, several hypo-
thetical discharge scenarios were developed (see Fig.
5 for lake surface elevations). These scenarios were
directed at  reducing the summer epilimnetic/metal-
imnetic total phosphorus concentration. If a predicted
reduction  could be  achieved, its possible effects on
algal concentrations and other response parameters
(such as transparency and dissolved oxygen) would be
analyzed.
  In Scenario B, an increased drawdown rate  was
assumed for the  period of August through October
which  would  theoretically take advantage of  high
nutrient concentrations in the hypolimnion, as well as
the mixing effects of the fall turnover process. June
and  July elevations were held constant at  approx-
imately 1,182 feet  MSL Heavy discharge was still
assumed for the peak winter heating months, as is the
current practice.
  Scenario C was similar to Scenario B except that
the increased drawdown rate was  assumed to start
  Table 2.—Flow and phosphorus loading input data for a
              normalized hydrologic year.
                                                                 A = Current Pol icy
Month
January
February
March
April
May
June
July
August
September
Ocotober
November
December
Total
Inflow (a)
(m3)
2.83x10?
2.86
5.67
5.67
3.77
2.07
1.56
1.33
1.08
1.67
2.69
3.17
Total Phosphorus
Load (b)
(kg)
1,425
1,453
2,724
2,934
1,868
1,079
817
683
612
879
1,398
1,703
                                   Figure 5.—Lake surface elevations for various discharge
                                   scenarios.
                                   earlier in the summer. This required that some of the
                                   winter and spring inflows be stored in the lake, rather
                                   than being discharged immediately.
                                     Scenario D assumed that the lake would be kept at
                                   a fairly  constant  elevation (approximately 1,182 feet
                                   MSL) throughout the year. This would  require that
                                   storm flows,  and presumably storm loads,  be dis-
                                   charged as soon as possible after they enter the lake
                                   (allowing some time delay for flow routing through the
                                   lake).
                                     Since  a  considerable  quantity of the water dis-
                                   charged through  the outlet pipe can come from the
                                   epilimnion/metalimnion, it was not considered advan-
                                   tageous to evaluate a scenario involving the release of
                                   water over the spillway.

                                   RESULTS

                                   Somewhat peculiarly, regardless of the initial concen-
                                   trations  or  discharge scenario  used, the  model
                                   predicted the final total phosphorus concentrations
                                   for the epilimnion/metalimnion and the hypolimnion to
                                   be equal at 24 ^g/l. The model, as it was structured, ap-
                                   parently predicts some sort of equilibration process in
                                   the  lake. Therefore, this concentration  was used as
                                   the  initial concentration for  another  set of  program
                                   runs involving each of the four scenarios.
                                     The model revealed several important factors about
                                   phosphorus reactions in the  lake. For example, sedi-
                                   ment release is estimated to be a significant source of
                                   phosphorus to the water column. For all scenarios,
                                   sediment release was estimated to account for 15 per-
                                   cent (3,000 kg) of the total input load (20,570 kg). Also,
                                   although the seasonal timing was different, the model
                                   predicted that   40  percent  of  the annual   input
                                   phosphorus load  was discharged via the outlet pipe
                                   for each of the scenarios. The remaining amount was
                                   transferred  to the sediments.
                                     As shown, the  model indicated that no significant
                                   reduction in the mean summer epilimnetic/metalim-
                                   netic concentration would  be achieved by altering
                                   PP&L's  current  discharge policy to any  of the alter-
                                   native scenarios evaluated:
                                                                            Predicted Mean Summer
                                                                            Epilimnetic/Metalimnetic
                                                                               Total Phosphorus
                                                                                   (ngl\ as P)
 Total
34.37x10?
17,575
 (a) Includes tributary flows and direct rainfall
 (b) Includes all external point and nonpomt sources.
Scenario A
Scenario B
Scenario C
Scenario D
25.2
25.0
25.1
24.9
                                                 1RP

-------
 LAKE AND RESERVOIR MANAGEMENT
   These predicted mean concentrations are in an ap-
 propriate range as determined by analyzing lake data
 for previous summers.
   Possible explanations for these results are:
   1. The volume of the "pure"  hypolimnion  is quite
 small compared to the volume  of the total lake (ap-
 proximately 1/20 at full pool elevation). Therefore, the
 effectiveness of any scenario tested will  depend
 primarily on epilimnetic factors.
   2. The lake has a fairly high flushing rate (approx-
 imately twice per year), meaning  that conditions at the
 beginning of the year do not play a significant role in
 determining summer phosphorus concentrations.
   3. Because  of  its  size and  shape, the  lake  is
 basically an  efficient  settling  basin.   Therefore,
 whatever phosphorus  is not  discharged quickly set-
 tles to the bottom sediments.
   4. One or more of the model assumptions may not
 have been valid.
   The sensitivity of the model was tested by arbitrarily
 increasing the inflows, input loads, and outflows by 25
 percent for Scenario A. The result was a 9 percent in-
 crease in the mean summer epilimnetic/metalimnetic
 total  phosphorus  concentration. No  uncertainly
 analyses were attempted since the model did not i i-
 dicate  any  significant  difference   between  the
 scenarios.
CONCLUSIONS

Assuming that the model theory and rate coefficient
used were correct, no alternative discharge scenario
produced a  significant change in the mean summer
epilimnetic/metalimnetic phosphorus  concentration
as compared to the current discharge policy. There-
fore, no  recommendations were made to the utilily
company at this time. The primary factors governing
total phosphorus concentrations in the lake  are the
timing and  magnitude of  input  flows  and  loads.
Although it may be advantageous to model the poten-
tial effects of assuming different hydrological input
scenarios, obviously there  is  no way to control the
weather;  and  therefore the  only way to  reduce
 phosphorus  concentrations  in the lake is via water-
 shed management techniques.
   The complexity of the model could be increased by
 segmenting the two vertical  compartments in a long-
 itudinal direction. Other improvements could perhaps
 be obtained by altering the model to account for solu-
 ble versus insoluble phosphorus reactions. This would
 allow  one to better  address the issue  of  bioavail-
 ability, at least at a preliminary  level of analysis.


 ACKNOWLEDGEMENTS: This study was performed under a
 private contract with the Lake  Wallenpaupack Watershed
 Management District, funded by the District's constituent
 members.  Special  appreciation is  owed to Dr. Steven C.
 Chapra for his assistance on this project.


 REFERENCES

 Chapra, S.C. 1975. Comment on "An empirical method of
  estimating the retention of phosphorus in  lakes," by W.B.
  Kirchner and P.J, Dillon. Water Resourc. Res. 2(6):1033-4.

 	1983. Personal communication. Feb. 5. Texas A&M
  University, College Station.

 Chapra, S.C.  and  K.H.  Reckhow.  1983. Engineering Ap-
  proaches for Lake Management. Vol.  2:  Mechanistic
  Modeling. Ann Arbor Sci. Publ., Ann Arbor, Mich.

 F. X. Browne  Associates, Inc. 1982. Lake Wallenpaupack
  water quality management study.  Lansdale, Pa.

 Horstman,  H.K., and F.X. Browne. 1982.  A watershed man-
  agement plan for Lake Wallenpaupack. In W.K. Johnson,
  ed. Proc. 1982 Am. Soc. Civil Eng. Natl. Conf. Environ. Eng.
  New York.

 Imboden, D.M. 1974. Phosphorus model of lake eutrophica-
  tion. Limnol. Oceanogr. 19(2):297-304.

 Mawson, S.J., H.L Gibbons, W.H. Funk, and K.E. Hartz. 1983.
  Phosphorus  flux rates in lake sediments. J. Water Pollut
  Control Fed. 55(8): 1105-10.

Snodgrass, W.J., and C.R. O'Melia. 1975. Predictive model for
  phosphorus  in lakes. Environ. Sci. Technol. 9(10):937-44.
Stauffer, R.E. 1981. Sampling strategies  for  estimating the
  magnitude and importance of internal phosphorus sup-
  plies in lakes. EPA 660/3-81-015. U.S. Environ. Prot. Agency,
  Corvallis, Ore.
                                                 170

-------
WATER QUALITY SIMULATION OF THE  PROPOSED
JORDANELLE RESERVOIR,  UTAH
DAVID L. WEGNER
U.S. Bureau of Reclamation
Upper Colorado Region
Salt Lake City,  Utah
            ABSTRACT

            Jordanelle Reservoir, a proposed component of the Bonneville Unit of the Central Utah Project, will
            be a major source of municipal and industrial water for the Salt Lake Valley. The final environmental
            impact statement identified the need to evaluate Jordanelle in terms of its downstream impact on Deer
            Creek Reservoir and the use of the water. To evaluate the impact, a combination of temperature and
            water quality simulation models, anaerobic simulation of the sediments, empirical nutrient loading
            models, and estimates of primary productivity were utilized. As a result of these efforts, Jordanelle
            Reservoir is expected to experience seasonal excesses of eutrophication with resulting low dissolved
            oxygen levels, potential mobilization of mine tailings, and seasonal recycling of nutrients from the
            sediments. To restrict the impact of these factors on downstream productivity and to allow efficient
            routing of nutrients and water through Deer Creek Reservoir, a multiple level withdrawal outlet struc-
            ture and variable operation scenarios were designed for Jordanelle. The simulation period covered
            the entire stagnation period with the coefficients defined from previous Deer Creek simulations, the
            proposed Jordanelle operation plan, and empirical relationships. The procedures defined will be in-
            tegrated with a watershed management plan to provide for efficient use of Jordanelle water with the
            resulting least impact to the downstream use of water. This analysis has applications to other im-
            poundments and the determination of efficient operation plans.
INTRODUCTION

Jordanelle Reservoir is proposed as a segment of the
Bureau of Reclamation's Central Utah Project, Bonne-
ville Unit. The proposed reservoir will be located on the
Provo River and  Drain Tunnel Creek area, approx-
imately 5  miles upstream from Heber City, Utah (Fig.
1). Water from the Jordanelle Reservoir will be used as
a portion of the municipal and industrial water supply
for communities along the Wasatch Front Area. Jor-
danelle Reservoir was identified in the Municipal and
Industrial  Supplement to the Final Bonneville Unit En-
vironmental  Impact Statement (1973), as potentially
having a major impact on  the water released to  Deer
Creek Reservoir and to the  municipal water supply.
Because the Jordanelle Reservoir may affect this seg-
ment of the Central Utah  Project, this study was in-
itiated to  evaluate the project  and  determine the im-
pacts of its reservoir design and management options.
  To do that effectively it was necessary to address
the Deer Creek-Jordanelle complex on a systems ap-
proach. The objectives were  determined to be:
  1. Evaluate the limnological environment of the pro-
posed Jordanelle Reservoir  in regards to its trophic
state, productivity, and reservoir dynamics.
  2. Predict the impact of Jordanelle Reservoir on the
downstream water quality  in  the Provo River and Deer
Creek Reservoir, and
  3. Evaluate the design and water quality criteria for
Jordanelle Reservoir in order to effectively develop  a
reservoir basin management plan that would provide
the least impact on downstream water quality and the
aquatic environments.
  The analysis of the  proposed  reservoir deals with
the system after  a period of 5 to 7 years. This time
period will allow the reservoir to progress through in-
itial filling  and nutrient  release from the  reservoir
basin  sediments.


PREDICTED TROPHIC STATE

The trophic state of the proposed Jordanelle Reservoir
was projected based on empirical models utilizing the
amount of phosphorus availability and sedimentation
rate as the critical  criteria.  Studies performed by
Mueller (1981) concluded that for proposed Bureau of
Reclamation reservoirs, a combination model of both
the Vollenweider (1975) and Canfield and Bachmann
(1981) approaches provided the best estimate of con-
ditions that could  occur  in Western reservoirs, the
primary components of this analysis can be outlined
as follows:
                P =
where:

    P

    L
    z
    a
    t
                     Z * (a + 1/r)


                      and

                o  = 6.67 * (L/z)0.589
in-lake total phosphorus concentration
(mg/l)
areal phosphorus loading rate (g/m.2 yr)
mean lake depth (m)
empirical sedimentation coefficient (yr~1)
hydraulic detention time (yr)
                                                 171

-------
LAKE AND RESERVOIR MANAGEMENT
Because of the configuration of Jordanelle Reservoir,
phosphorus and water budgets were computed  for
each of the major arms of the reservoir. The results are
presented  in Table 1.
   The average annual loading under average year pro-
ject conditions is projected to be approximately 6,000
kg of total phosphorus (TP) per year. This represents
approximately 2.0 g TP/m2 yr that will be added to the
reservoir. The  loading levels for the proposed reservoir
will  occur throughout the year, with the major inputs
occurring during the spring runoff. Figures 2 and 3
depict the average annual flow levels and phosphorus
inputs for the  reservoir basin.
   Based on the combination  model of Vollenweider
and  Canfield/Bachmann, the in-lake total phosphorus
concentration  under average project conditions in the
Drain Tunnel Creek arm is projected to be 0.025 mgi/l
and 0.020 mg/l in the Provo River arm. The combina-
tion of the  estimated phosphorus concentrations in
the  two tributaries  and annual flow  levels yields  a
calculated annual areal phosphorus loading rate of
0.82 g/m2 yr and a calculated in-lake total phosphorus
concentration of 0.014 mg/l. Based on this analysis
the probability of eutrophic conditions occurring with
the estimated phosphorus concentrations  is 28 per-
cent  near the dam, 64 percent in the Drain Tunnel
Creek arm, and 50 percent in the Provo River arm.
   Garner (1983) predicted that Jordanelle  Reservoir
would exhibit mesotrophic to slightly eutrophic con-
                                                      Figure 2.—Annual inflow water levels to Jordanelle Reservoir
                                                      under average project flow conditions.
Figure 1.—The proposed Jordanelle Reservoir with the main
inflows.
Figure 3.—Annual total phosphorus inflow levels into Jor-
danelle Reservoir under average year project conditions.
            Table 1.—Average annual inflow and total phosphorus loadings under average project conditions.

                                       Jordanelle Inflow • Average Project Conditions
Provo River

Month
October
November
December
January
February
March
April
May
June
July
August
September
Total
Discharge
** '(Acre- Ft)
5.1
4.9
4.7
4.7
4.7
5.2
27.7
84.5
59.9
9.7
5.3
4.8
221.2
Phosphorus
MG/L
0.02
0.02
0.03
0.01
0.05
0.09
0.09
0.03
0.01
0.01
0.02
0.03

KG**
125.8
120.9
173.9
58.0
289.9
577.3
3,075.1
3,126.9
738.9
119.6
130.7
177.6
8,714.6
Weber Diversion
Discharge
(Acre-Ft)
0.6
0.5
4.3
3.8
3.3
3.7
1 6.9
0.0
0.0
7.9
1.8
1.1
48.4
Phosphorus
MG/L
0.05
0.02
0.02
0.01
0.05
0.04
0.07
0.00
0.00
0.03
0.09
0.04

KG
37.0
123.3
106.1
46.9
203.5
182.5
1,459.2
0.0
0.0
292.3
199.8
54.3
2,704.9
'Drain Tunnel Creek
Discharge
(Acre-Ft)
1.4
1.6
1.6
1.6
1.6
1.8
2.0
2.7
1.8
1.5
1.4
1.3
20.3
Phosphorus
MG/L
0.7
0.5
0.5
0.1
0.12
0.21
0.22
0.12
0.7
0.11
0.19
0.15

KG
120.9
98.7
98.7
19.7
236.8
466.3
542.7
399.7
155.4
205.5
323.1
240.5
2,911.0
 Notes. 'Includes McHenry and Ross Creek Flows
      "•KG = (Ac-Ft) (MG/L) (1.23349)
      ""Thousand Acre - Feet
                                                  172

-------
                                                                  MODELING TECHNIQUES AND INNOVATIONS
ditions, based on calculations of the Larsen and Mer-
cier model (Mueller, 1981). This application indicated
that  the  reservoir  will be  sensitive  to  phosphorus
levels and loading  rates.  The values calculated with
the combination model also indicated that the  reser-
voir will be mesotrophic to eutrophic, with the  Drain
Tunnel Creek arm being more sensitive to eutrophica-
tion.
  It is anticipated  that Jordanelle will experience a
seasonal succession of algal species as temperature,
nutrient levels,  and  species competition vary. Deer
Creek  Reservoir, located  9 miles downstream,  ex-
periences a spring diatom pulse, followed by progres-
sion  into a  green-dominated complex.  If nutrient
levels,  anoxic conditions,  and temperatures  last for a
significantly long  enough  period,  the green species
will give way to  blooms of blue-green species, usually
Aphanizomenon flos aqua. Jordanelle Reservoir will
not be light  limited but algal  populations will be
limited  by  the  proportions  of the nitrogen,
phosphorus, and carbon available  during the growing
season.  Phosphorus will  be  the primary limiting
nutrient early in the growing season with the potential
for nitrogen limitation later in the summer or early fall.
The potential for the shift in nutrient limitations plus
the ability of blue-green algae to fix nitrogen from the
atmosphere, will allow the blue-greens to outcompete
the green algal species  in the reservoir.  It  is  an-
ticipated  that the occurrence of blue-green algae will
be more prevalent in the Drain Tunnel Creek arm  of the
reservoir. The probability of eutrophic conditions oc-
curring will expand  under low flow years to approx-
imately 76 percent reservoir wide.
  To determine the  probability  if  Jordanelle will be
dominated  by  blue-green  or  non-blue-green  algal
types,  a discriminant function was calculated based
on the methodology of Reckhow and Simpson (1980).
The main components of this analysis are the level of
median summer in-lake inorganic nitrogen concentra-
tions and the average influent phosphorus concentra-
tions. These  data are combined  with the hydraulic
detention times of the main reservoir  areas  to define
the principal component  function. Based on this
analysis,  it was determined  that the Drain Tunnel
drainage had a probability of 98 percent dominance by
blue-green species, the Provo River drainage had a 95
percent probability and the  area  near the dam was
calculated to be 85 percent probability. Based on
previous studies on  other high elevation reservoirs, it
is suspected that these values may be overestimated.
In any  sense  however, Jordanelle  Reservoir shows a
propensity for being dominated by  blue-green species
later in the growing season.


THERMAL STRATIFICATION  AND
HYPOLIMNION OXYGEN DEFICITS

The prediction of thermal stratification and hypolim-
netic oxygen  concentrations is important in the rela-
tionships established between the water column and
the  chemical  and  nutrient  recycling  from the
sediments and  mobilization from abandoned  mine
tailings. Craft (1982)  has indicated that the composi-
tion of the mine tailings in the upper end of the  Drain
Tunnel Creek arm will be sensitive  to anoxic/reducing
environments. Based on calculated detention times,
inflow/outflow rates,  seasonal shifts, and basin mor-
phology, it is anticipated that stratification will  occur
in each section  of Jordanelle Reservoir with more pro-
longed and  intense stratification occurring in the
Drain Tunnel Creek area.
  To more accurately define the reservoir thermal pro-
files that will occur near the withdrawal structure in
Jordanelle, the Corps of Engineers Water Quality for
River and  Reservoir Systems (WQRRS) computerized
mathematical simulation  model was used (Wegner,
1983). This type of application had  been utilized in
past Bureau studies by Yanke (1981). The results of
the utilization of this model are depicted in  Figure 4
which defines the seasonal changes in  dissolved ox-
ygen levels that could be expected to occur as thermal
stratification intensifies during the summer season.
This analysis indicates that the lower portion of the
hypolimnion will go  anoxic   30  to  40 days after
stratification begins and  is projected to last for ap-
proximately 40  to  50 days. The length of time that
anoxic conditions will be prevalent in the hypolimnion
will be a function of meteorological conditions, inflow
rates, timing of  initial stratification, and  biological ox-
ygen demand.
  The  length of time that anoxic  conditions are ex-
isting in the hypolimnion also has an impact  on the
amount of phosphorus  recycling that can occur from
the reservoir sediments. Messer (1983)  projects that
up to 18 percent of the total phosphorus budget of
Deer  Creek Reservoir  may be attributable  to  the
phosphorus recycling from the reservoir sediments.
 RESERVOIR INTERRELATIONSHIPS AND
 DOWNSTREAM IMPACTS

 Initially, Jordanelle Reservoir will experience substan-
 tial productivity as the reservoir basin fills and the
 available  nutrients are leached from the sediments.
 The purpose of this  analysis was to evaluate  Jor-
 danelle Reservoir based  on  how it will  interrelate
 under project conditions. The development of the Cen-
 tral Utah Project will alter the flow and downstream re-
 quirements for water supplies. Realizing that there is
 no such thing as "average" reservoir conditions and
 that reservoir equilibrium  may take 15 to 20 years to
 achieve, this analysis was  based on the best probable
 estimate of how the reservoir systems will be operated
 under project conditions. Jordanelle Reservoir will col-
 lect phosphorus and heavy metal inflows. It has been
 estimated by Garner (1983) that approximately 49 per-
 cent  of the inflowing phosphorus amounts  to  Jor-
 danelle will  be retained in  the sediments.  This
 calculates out  to be approximately 4,700 kg/yr of re-
 tained phosphorus.
   This level of phosphorus represents approximately
 19 percent of the total phosphorus load to Deer Creek
 Reservoir. This trapping of phosphorus  in the  Jor-
 danelle Reservoir will reduce loading  to Deer Creek
 and should help improve its trophic status. Knowing
 that  Deer Creek  Reservoir  is  a major  source of
 municipal and  industrial water to the Wasatch Front,
 an objective of the Deer Creek and Jordanelle Reser-
 voir Management Plan and the  Bureau of Reclama-
 tion's design,  has been to use the Jordanelle Reser-
 voir as a  water quantity/quality  control for the Deer
 Creek system.  To effectively  manage  the  two reser-
 voirs, a multiple level withdrawal structure and aera-
 tion have  been evaluated for Jordanelle.
   Seasonal control of the quality of water released
 from  Jordanelle would control phosphorus loading to
 Deer  Creek. Summer releases of low level phosphorus
 water would help reduce the biological stimulation of
 algal  blooms and subsequent poor water quality. It is
 anticipated that once thermal  stratification occurs,
 the reservoir releases from Jordanelle would be moved
                                                 173

-------
 LAKE AND RESERVOIR MANAGEMENT


             JUNE



 Surface
 Bottom
                                S.
JULY
                                                          S.
                        AUGUST
                                                                                   S.
                                                  SEPTEMBER



                     10
                                              10
                                                                        10
                                                                                                 10
                                       Dissolved Oxjtgen  Levels ( mg/L  )

 Figure 4.—Hypolimnetic oxygen depletions over the sunrmer in the proposed Jordanelle Reservoir—average project condi-
 tions.
to a  higher elevation to avoid high nutrient  levels.
As thermal stratification breaks down in the fall, the
reservoir releases can be moved farther down into the
water column to flush the nutrients downstream and
through Deer Creek.
   To alleviate  the  potential problems  of  anox c
stimulation of heavy metal movement from the mire
tailings in the Drain Tunnel Creek  arm, an aeration
system to break down thermal stratification may be
necessary. Further studies are being conducted in this
area.


SUMMARY

The development and prediction of  the water quality
environment  for the proposed Jordanelle Reservoir
has   combined empirical,   mathematical,  time-
independent,  and  time-dependent  models and ap-
proaches.  No  one  method has  proven to  be the
ultimate model or  to  provide all  of  the  answers.
Developing an effective reservoir management plan  or
estimating reservoir response to  development  re-
quires careful analysis  of  the seasonal trends and
downstream requirements. Effective utilization of Jor-
danelle Reservoir to maximize water quality  condi-
tions in Deer Creek  Reservoir will  require some in-
novative techniques and  management  procedures.
Based on  our  analysis, we project that Jordanelle
Reservoir  will  fluctuate between  mesotrophic  and
eutrophic  productivity levels; the  best management
practices  may be achieved with a  multiple level with-
drawal structure and seasonal utilization of reservoir
dynamics.
                 REFERENCES

                 Canfield, D.E., Jr., and R.W. Bachmann. 1981. Predictions of
                   total phosphorus concentrations, chlorophyll a, and Sec-
                   chi depths in natural and artificial lakes. Can. J. Fish.
                   Aquat. Sci. 38: 414-23.

                 Craft, D. 1982. Estimate of heavy metal movement from the
                   mine tailings in the Jordanelle  Reservoir basin. Internal
                   memo.

                 Garner, L.  1983. Phosphorus budgets of Deer Creek and
                   Jordanelle Reservoirs, Utah. M.S. Thesis. Brigham Young
                   Univ., Provo, Utah.
                 Messer, J.J. 1983. Personal communication.
                 Mueller,  O.K.  1981. Mass  balance  model  estimation of
                   phosphorus concentrations in reservoirs.  Water Resour.
                   Bull, (in press).
                 Reckhow, K.H., and J.T. Simpson. 1980. An empirical study of
                   factors affecting blue-green versus non blue-green algal
                   dominance in lakes. Inst. Water Resour., Mich. State Univ.,
                   East Lansing, Mich.
                 U.S. Dep. of the Interior. 1973. Final  Environmental Impact
                   Statement, Bonneville Unit, Municipal and Industrial Sup-
                   plement, Central Utah Project. Bur. Reclam.
                 Vollenweider, R.A. 1975. Input-output models with special
                   reference to the phosphorus loading concept in limnology.
                   Schweiz. Z. Hydrol. 37 (1): 53-84.
                 Wegner,  D.L 1983. Development of the water quality plan
                   for Jordanelle Reservoir. Internal memo., U.S. Dep. Inter.
                 Yahnke, J.W. 1981. Water  quality of  the proposed Norden
                   Reservoir, Nebraska, and its  implication  for  fisheries
                   management. REC-ERC-81-8, U.S.  Bur. Reclam., Denver,
                   Colo.
                                                   174

-------
TIME SERIES  MODELING OF RESERVOIR WATER QUALITY
ROBERT  H. MONTGOMERY
U.S. Army Corps of Engineers
Waterways  Experiment  Station
Vicksburg, Mississippi
             ABSTRACT

             Time series models of reservoir water quality are water resource management tools and provide in-
             sight into reservoir dynamics since they account for autocorrelation, seasonality, and trends present
             in the data. Autoregressive integrated moving average (ARIMA) models were applied to describe pat-
             terns in selected water quality variables of Red Rock Lake, Iowa. Models were generated for total
             phosphorus, total nitrogen, and suspended solids for an 8-year period (1972-1979). Patterns in lake
             concentration and the usefulness of time series models to lake and reservoir management are described.
 INTRODUCTION

 Mathematical modeling of water quality advances our
 understanding of aquatic ecosystems and provides
 potentially valuable  information for  water resource
 managers  (Chen,  1970;  Spofford,  1975;  Reckhow,
 1979). One stochastic method of  modeling reservoir
 water  quality variables  over time  is  time  series
 analysis (Box and Jenkins, 1970). Time series models
 may be used to  (1) define mathematically underlying
 processes (McKerchar and  Delleur,  1974; Tiao and
 Delleur, 1976), (2) forecast future conditions (Newbold,
 1970; Cogger, 1979),  (3) determine the effect of inter-
 ventions, natural or man-induced (Box and Tiao, 1975;
 Hipel et  al. 1975), (4) detect trends (Lettenmaier et al.
 1978), and (5) relate a response variable to a set of in-
 put variables (Jenkins, 1979). Time series models are
 advantageous in modeling water quality because they
 can incorporate  the  effects of hydrologic and limno-
 logical   phenomena  (for example,  autocorrelated
 values, seasonality, and trends).
   Traditional methods of time series  analysis involve
 decomposing a  series into trend, seasonal or cyclic
 variation, and other  irregular fluctuations (Chatfield,
 1975). A trend is  usually defined as a long-term change
 in the mean level of  a time series. Seasonal or cyclic
 changes are variations that  occur at some fixed fre-
 quency (such as yearly  temperature variations). By
 removing trends and/or cyclic  variations, residual
 variations can be described in terms of probability
 models (like autoregressive or moving average). These
 probability models assume that the time series data
 are stationary, that is, no systematic changes occur in
 the mean or variance and all periodic variations are
 removed.
   Time series analysis is divided into three phases:
 identification, estimation, and forecasting  (Box and
 Jenkins, 1970;  Box  and  Tiao, 1973). First, the time
 series data are plotted to identify possible trends and
 seasonalities. If  a trend is present it may be removed
 by (1) fitting a simple mathematical function to the
 curve, (2) using a linear filter (Kendall, 1973), or (3) dif-
 ferencing a series until it becomes stationary. Season-
 ality may also be removed by filters or differencing
 techniques.  Transformations  of  the data may be
 necessary either to stabilize the variance or to make
 the seasonal effect additive. After seasonal or cyclic
 variations are removed and any necessary transforma-
 tions are performed, autocorrelation and partial auto-
correlation coefficients are calculated. Plots of auto-
correlation  and partial autocorrelation coefficients
against their lags (correlogram) provide a preliminary
suggestion  as to which time series models and lags
are appropriate (Chatfield, 1975). The correlogram may
also provide a check to see if the series is stationary.
  After the  identification phase has led to a tentative
formulation of the model, the estimation phase yields
estimates of the parameters via  the Box and Jenkins
least-squares method (Box  and Jenkins, 1970). The
models are subjectively evaluated for goodness of fit
based on parameter estimates, sum of squares error,
T-ratio, and  standard error  estimate of  model. If a
model is adequate according to defined  criteria, the
final phase, forecasting, is used to estimate predicted
values and confidence limits. Model residuals are then
graphically and  statistically tested for goodness of fit
with actual data (Montgomery and Johnson, 1976). If
more than one model is acceptable, the models may
be compared by the previously mentioned statistics or
the Akaike Information Criteria (AIC) (Akaike, 1974) to
choose the model with the best fit. Once a final model
is selected the  predicted values and the upper and
lower confidence intervals may  then be plotted with
the actual values to  provide  a visual examination of
model fit.
  The intent of this paper is to show the application of
time series  analyses in  modeling  important limno-
logical variables (total phosphorus, total  nitrogen,
suspended solids) in the Lake Red Rock, Iowa. Also, to
discuss potential applications of time series analysis
to lake/reservoir limnology and management.


STUDY SITE

Lake Red Rock, a U.S. Army Corps of Engineers flood
control reservoir, was created in  1969 by the impound-
ment of the Des Moines River 96  km downstream from
Des Moines, Iowa. Red Rock Dam controls discharges
from the Des Moines River, a 32,000 km2 drainage
basin consisting predominately  of agricultural lands
with point sources from the city of Des Moines. Lake
Red Rock has a mean depth of 3  m, a maximum depth
of 11 m, a surface area of 36 km2, and an average theo-
retical hydraulic residence time of 7 days at normal
pool level. Perceived water quality problems are most
often associated with excessive  suspended sediment
                                                 175

-------
 LAKE AND RESERVOIR MANAGEMENT
 concentrations even though the reservoir is classified
 as eutrophic with respect  to  both nutrients (phas-
 phorus and nitrogen) and chlorophyll concentrations
 (Bauman et  al. 1979). Sediment surveys indicate an-
 nual  sedimentation  rates (mainly clay) ranging from
 4.9 cm/yr near the dam to 27.1 cm/yr (mainly fine grain
 silts) in  upstream areas over the old flood plain and
 river channel or thalweg (Kennedy et al. 1980).
 METHODS

 Total  phosphorus, total  nitrogen,  and suspended
 solids concentrations were determined approximately
 weekly during the period  1972-1979 (Bauman et al.
 1979). Surface samples, collected at a deepwater sta-
 tion approximately 0.8 km upstream of the dam, were
 preserved  in  the  field, stored  in  acid-washed poly-
 ethylene bottles, and  returned to the laboratory for
 analyses in accordance with Standard Methods (1975).
 While the  use of  weekly measurements would he.ve
 been most appropriate based on the hydraulic deten-
 tion time (approximately 7 days), weekly records wore
 often  incomplete.   Therefore,  mean  monthly
 measurements were   used  for  statistical  analyses
 reported here.
   Time series models were generated using the PROC
 ARIMA procedure of the Statistical Analysis  System
 (SAS) (Version 79.4b) (Stat. Anal. Syst., 1979). ARIMA
 (Autoregressive Integrated Moving Average)  models
 given an observed value as a linear combination of
 past values (autoregression) and past errors (moving
 average) from equally spaced time series data with no
 missing values. The integration part of the model ac-
 counts for nonstationary sources of variation by dif-
 ferencing  the series into a stationary series,  that is,
 the stationary model fitted to the differenced data has
 to be summed or integrated to provide a model for the
 nonstationary data. The general ARIMA  model is of
 the form (Box and Jenkins, 1970):
    X  = 0  -
          -... -0qa,_
+ 4>pXt-P + a,
                                      a  -
where:  X, = the original data or a difference of
             degree (d) of the original data (use W,
             if data were differenced)
         0 = constant term
     fi...<|>p = autoregressive parameters (order p)
    0-i...0q = moving average parameters (order q)
        a, = random error

ARIMA models may be presented by the level or order
(i.e., (p,q,d)), where p,q, and  d are the number of aulo-
regressive, moving average, and seasonal differencing
parameters of the model, respectively.
RESULTS

Total Phosphorus. Epilimnetic total phosphorus con-
centrations varied seasonally and were, in  general,
highest in spring  and fall  (0.6 mg/l),  and lowest in
winter and summer (0.2 mg/l) (Fig. 1). This seasonal
pattern is evident by the occurrence of significant par-
tial autocorrelations at lag 1 and 12 (Fig. 2). However,
no definite model type (i.e., AR or MA) could be deter-
mined from these plots. Thus, ARIMA models of order
(1, 0, 0) (0, 0, 1), and (1, 0,1) were estimated for a non-
seasonal case and a seasonal  case. The seasonal
case  consisted  of a  seasonal difference =  1 and
period = 12, with the use of different combinations of
                           order 1 seasonal autoregressive and seasonal moving
                           average parameters. The models were first examined
                           to see if the parameter estimates were significant at
                           the a =  0.05 level using the T-ratio. The four ARIMA
                           Models with significant parameter estimates were of
                           order (1, 0, 0) x (1, 0, 0), (0, 0,1) x (0, 0,1), (1, 0, 0), and
                           (0, 0,  1) (Table 1).  Model residuals were assessed sta-
                           tistically  for  goodness  of fit  (Montgomery  and
                           Johnson, 1976), with all four models being significant.
                           Model predictions, including upper and  lower 95 per-
                           cent  point  confidence intervals,  and actual values
                           plotted for the four ARIMA models of order (0, 0,1) x
                           0, 0,1), (1, 0, 0) x (1, 0, 0), (1, 0, 0), (0, 0,1) are plotted in
                           Figure 3.
                             Total Nitrogen.  Seasonal patterns in epilimnetic
                           nitrogen concentrations were more variable (1.5-12.0
                               Ob.
                                   1972  1973  1974   1975   1976   1977  1978  1979
                                                     YEAR
                                   1972  1973  1974  1975   1976   1977   1978  1979
                                                    YEAR
                             700


                             600


                             500


                             400


                             300


                             200 -


                             100 -


                               0 i
                                  1972  1973  1974   1975  1976  1977  1978  1979
                                                    YEAR
                          Figure  1.—Time  series  plot  of  mean  monthly  total
                          phosphorus, total nitrogen, and suspended solids for Red
                          Rock Reservoir.
                                                 176

-------
mg/l) than those for total phosphorus (Fig. 1). The
autocorrelation does not truncate, but rather dampens
out, suggesting the presence of autoregressive terms
(Fig. 2). The partial autocorrelations were significant
at lags 1, 11, and 13, suggesting an autoregressive
model of order 13 with 4>2> fio and $12 confined to zero.
To check for model parsimony, an AR (1) was also
estimated. Both, the AR (1) and AR (1,11,13) (this rep-
                                                          MODELING TECHNIQUES AND INNOVATIONS

                                           resentation is used for AR model with 2 - 
-------
LAKE AND RESERVOIR MANAGEMENT
   Suspended Solids. Plots of mean monthly suspend-
ed solids concentrations and estimated autocorrela-
tion and partial autocorrelation coefficients  plotted
through lag 24 (Fig. 1 and 2, respectively) suggesl a
stationary series, no seasonality, and the use of lags
1, 11, and  13. Therefore, ARIMA models  of differemt
combinations of orders (i.e.,  1, 11, 13) were estimated
and evaluated.  Models with significant  paramei-er
estimates were MA (1,  11, 13), ARIMA (0, 0, 1) and
ARIMA (1, 0, 0) (Table 1). All  three models provide a
reasonable fit of the data with the MA (1,11,13) having
the lowest model standard error (Table 1). Predicted
values, upper and lower 95 percent  point confidence
intervals, and actual values for models MA (1, 11, 13)
ARIMA (0,  0, 1), and ARIMA (1, 0,  0) are plotted in
Figure 5.

DISCUSSION

All four time series models  of total phosphorus are
sufficient in modeling the total phosphorus concen-
tration in  Lake Red Rock (Table 1). All  models had
significant parameter estimates (a = .001) and similar
model standard errors (range .703-721).  However, a
check of residual autocorrelations (those not signifi-
cantly different  from zero),  and the lowest model stan-
dard error (.703), suggest the use of the seasonal mov-
ing average model. In general, seasonal oscillations in
predicted values tended to be more dampened than
those for actual values (Fig. 3). The models also tend-
ed  to overestimate actual total phosphorus  during
times of sustained low values and peaks (or troughs)
at one lag after a peak (or trough). One case of model
 lack of fit occurred in the winter of 1977. The extreme-
 ly high concentration of total phosphorus coincided
 with an algal bloom that occurred because of a period
 of no snow cover on the lake. These values may be
 considered nonindicative of the normal phosphorus
 population; however, it was not considered an outlier
 since the value was the result of an uncommon occur-
 rence and not sampling error. The seasonal pattern
 and  model tendencies  of the  models may provide
 good forecasts if annual means are  desired, but are
 questionable if attempting to identify  peak spring con-
 centrations. Hence, if conditions in the watershed re-
 main the same, one may be reasonably confident that
 total phosphorus concentrations will  be below 2 mg/l.
  Of the  two total nitrogen  models that were esti-
 mated to be sufficient, only one, AR (1,11,13), provid-
 ed a reasonable fit based on the check of  residual
 autocorrelations (Table 1). The AR (1) model may be
 usable, depending on the level of significance accep-
 table by the user (i.e., a •- .109 at lag 12 for residual
 check). The total nitrogen series had strong fluctua-
 tions in the data (summer peak, winter trough) and
 resulted in wider confidence limits than those for total
 phosphorus. As with total phosphorus, the predicted
total nitrogen values tended to be damped, especially
 during an extended period of low (as in 1977) and high
(as in 1979) actual values. The middle section of the
 predicted series (1975-1977)  seems  to show nearly
twice the variability in comparison to predicted values
 in years  1972-1974 and 1978-1979. The large varia-
 bility and damped predictions limit the reliability of
forecasting annual and seasonal total  nitrogen con-
centrations.
     6


     5


     4

   ex
   01 3


   H 2

     1


     0


     -1
       1972  1973  1974  1975  1976  1977  1978 1979
                        YEAR
     1972 1973  1974  1975  1976 1977  1978 1979
                      YEAR
     7

     6

     5

     4
   Cx
   01 3
   E

   a.' 2
   i-


     0

     -1

     -2
       1972  1973  1974  1975  1976 1977  1978 1979
                        YEAR
    7

    6

    5

    4

  oi3
  E
   -2
     1972 1973  1974  1975 1976 1977  1978  1979
                      YEAR
Figure 3.—Plot of predicted, 95 percent point confidence limits, and actual values for total phosphorus models: ARIMA (0, 0,
1) x (0,0,1), ARIMA (1, 0, 0,) x (1,0, 0), AR(1), and MA(1). Diamond  =  predicted, * = actual, dashed line = 95 percent point
confidence limits.
                                                 178

-------
                  MODELING TECHNIQUES AND INNOVATIONS












co
S
••£
o
c
o
a
1
co
S
•o
c
CO

CD
i
1
i-
1
co
I--




























2 0>
oJS
"CM
3 r-
If
fit
'ai t
52
55
Model
Variance
81
c.2
o To
1 I
55 •£













£
CO
•c
CO
i— O CO h~
co CM f^ in
o> co co r^-

CD co co t^
r^ cvi 05 co
co co co co

8 CM CM i-
i- i- CM

C» O O CM
•«t in in in


§ §

r- T- T- T-
8888

§888

o o co co
CM CO O f-
cxi CNJ co •*
•*•««••*•*



2 cf
1 .^
o Si

+ ^

in

i
•S
co T
in ^
+ in

CO* "^
-j. _j_

o o
5 S? T ff
11 II * t
x" x" E —
^ & + o
o- ^_- CO •*
X X 5 m
^ g- i-: in
° P II ||
8. C. - L
<<><>£
S E  00
8 "

o o
in *-t

||

CO GO
o in
co co
8
T—
8

,_ .^
8 8

§ 8

o> co




c6~
CO

1
x"
O>
m
CM
r
T-
1
X*
in
«


T ctT
S T
r». ^
CN 1
O5
II ^
X S
CO" ||

jf x"
fir nr
< <

z
1-

O> O) O>

oo co co
g 18 S

^ CD CO
CM CO CO

o •* cq
CM CD 1^
^ ^ CO
in co r*-
m co oo
8
in
CM
8

i- CN CJ
p p 8

in
CM i- T-
888

P CM in
1^ CM CM
O CO CO



m
1
3

co
1
'

1
ccT
CO
CM
/

1
g

n- j. +
"t. in i
a w *r
+ n- CM
CM + $)i
$ * •
S + +
n g S
X" 00 S
S" n M
i— "
""- — —
? 5" 2"
2 5 <

CO
CO

     Of the  three  variables,  the  ARIMA  models  for
   suspended solids  provided the  best  fit (residual
   check),  but  the parameters  were  less  significant
   (Table 1). While all three models fit the data excellent-
   ly, the MA(1,11,13) had the lowest model standard er-
   ror (124.79) and is therefore recommended. Suspended
   solids concentration in Lake Red Rock tended to be
   event-oriented due to the effect of storms. All of the
   models  underestimated  the storm-related  peak
   suspended solid concentrations.  Hence, if these
   peaks are of interest,  the use of a transfer function
   model, with flow as the input, may be  more useful.
   Also,  with all  three  models, the lower confidence
   limits were almost always negative, which obviously
   is impossible in limnological data of this kind; and
   botfr lag  1 models consistently overestimated the
   baseline conditions.
     One problem with using monthly instead of weekly
   data is a decreased sample size and a loss of informa-
   tion. This is very important at Lake Red Rock because
   the residence  time is approximately 1  week;  thus
   changes in water quality may occur rapidly. The use of
   monthly instead of weekly data is of special concern
   when  modeling spring (limiting)  nutrient conditions
   because monthly observations will  dampen  weekly
   peaks and thus provide misleading information.  How-
   ever, this same dampening of weekly peaks by month-
   ly data  may enhance the application of the event-
   oriented suspended solid data by dampening the large
   storm event impulses.
   CONCLUSIONS

   Observed total  phosphorus concentrations in  Lake
   Red Rock had a strong seasonal pattern, exhibited by
   time series and autocorrelation plots, with peak con-
   centrations occurring in spring and fall. Only lag  1 and
   12 autocorrelations were significant. This suggests
   that within  a year phosphorus concentrations can be
   modeled as a function of only the prior month's con-
   centration,  while over all years, observed concentra-
   tions for individual months are extremely similar. Ob-
   served total nitrogen concentrations  exhibited  large
   fluctuations during a year. Peak concentrations tend-
   ed to occur in the summer; however, in some years oc-
   curred yearly. Significant autocorrelations occurred
   both lags 1  and 2 within a year. A change from positive
   to negative autocorrelations  between  successive
   years suggests  that high concentrations in one year
   imply lows the next year. Very little seasonal patterns
   were evident in  observed suspended solid concentra-
   tions  except for extremely  large spring peaks, pro-
   bably from storm events. Autocorrelation between
   months within a year were small while significant bet-
   ween the same  month over all years.
     For each variable, at  least  one model provided a
   reasonable fit to the data. The models in general were
   very good at predicting mean concentrations, as well
   as in generating synthetic series, and may be valuable
   in examining limnological processes. Important con-
   cerns of the ARIMA models generated in this applica-
   tion are the following: (1) predicted values are damped
   and result  in under and  over estimation of predicted
   values, especially during extended periods of  cons-
   tant  actual values; (2) predicted values tended  to be
   1-2 lags forward of the actual values in the series (this
   is especially important if the  concern is with peaks
   and troughs); (3) the models may be  limited  in fore-
   casting ability when the model  incorporates only lag 1,
   the series shows high variability, or the series has in-
   consistent  changes in  patterns; (4) the lower  point
179

-------
 LAKE AND RESERVOIR MANAGEMENT
 confidence limits may be negative, an impossibility in
 nature; and  (5) when modeling  event-oriented data
 transfer function models are suggested. While thsse
 concerns may be viewed as problems in some cases,
 they should be used in the application of the models
 to enhance the overall information generated.
   The application of time series models  to reservoir
 water  quality management has excellent  potential  in
 the areas of:  (1) forecasting, (2) generating synthetic
 series, (3) examining limnologic processes, (4) input-
 output formulation (that is, transfer function models,
 and (5) determining the effect of interventions  and
 detecting trends. Forecasting of water equality condi-
 tions is probably the single most important concern
 for most resource managers. Time series analysis s a
 quantifiable forecasting technique that allows predic-
 tion of future concentrations  given that  conditions
 (watershed) remain constant, or change at a constant
 rate, or if conditions are suddenly altered (intervention
 analysis). For example, the time series model selected
 may be altered, or input altered  (if transfer function
 model) and examine potential changes in predictions.
 Also, a proposed reservoir water quality  may  be ex-
 amined based on models developed at similar reser-
 voirs.
   A second  application of  time  series models,  one
 dealing more  with research than  management, is the
 generation of synthetic series  (model predictions for
 the  same  time  period  as  the  original  data),  "he
 generated series can be used as a simulation itself, an
 input variable for an ecosystem model, and a valida-
 tion of a second model predicting the same variable.
      972   1973   1974   1975   1976   1977   1978   1979
                     VEAR
Figure 4.—Plot of predicted, 95 percent point confidence
limits, and actual values for total nitrogen models: AR(1,11,
13)and AR(1). Diamond = predicted,* = actual, dashed line
= 95 percent point confidence limits.
   Probably the most significant feature of generated
series is the ability to greatly increase the temporal
frequency of data. The development and type of time
series model generated will also provide useful limno-
logical information on the variable of concern, such as
seasonality,  trends, type, and structure  of  autocor-
relation within the variable. Time series plots provide
visual evidence of patterns, trends, and seasonality in
observed concentrations. The autocorrelation shows
the correlation structure between  observations  and
partial autocorrelations show the significance of indi-
vidual lags in model development. This limnological
information will assist in the application of the models
(say,   forecasting),  sampling  design,  and  overall
understanding of reservoir ecology.
                                                              1973  1974  1975  1976  1977  1978  1979
                                                                               YEAR
                                                           600


                                                           400
                                                        o> 200
                                                        E
                                                          -200


                                                          -400
                                                              1973  1974   1975   1976  1977   1978   1979
                                                                              YEAR
                                                          800 r
   -200	
       1973  1974  1975  1976   1977   1978   1979
                        YEAR

Figure 5.—Plot of predicted, 95 percent  point  confidence
limits, and actual values for suspended solid models: MA(1,
11, 13), MA(1), and AR(1). Diamond =  predicted,  * =  actual,
dashed line  = 95 percent point confidence limits.
                                                  180

-------
                                                                        MODELING TECHNIQUES AND INNOVATIONS
  A fourth  application, transfer function models,  is
the forecasting of an output variable which is related
to  an input  variable(s).  An  example is predicting
nutrient  loadings (output) to  a reservoir that  is im-
pacted significantly by storm events, as a function of
flow (input). Other potential  examples are modeling
lake nutrient  concentration(s) as a function of input-
concentration^); modeling outflow oxygen concentra-
tions as a function of lake oxygen concentration just
above the  dam; and modeling  phosphorus release
from sediments as a function of hypolimnetic oxygen
concentration (internal loading/recycling). Finally, the
time series  models have been shown to be applicable
in detecting trends and examining the effect of inter-
ventions. For example,  the effects of dam construc-
tion and/or proposed new management  plans (e.g.,
new operating scheme for dam or addition of a waste
treatment plant) on  inflow, lake, and outflow water
quality.
REFERENCES

Akaike, H. 1974. A new look at statistical model identifica-
  tion. IEEE Trans. Contr. AC-19(6): 716-23.
Bauman, E.R., C.A. Beckert, D.L Schulze and D.M. Soballe.
  1979. Water Quality Studies-EWQOS Sampling Red Rock
  and Saylorville Reservoirs, Des Moines River, Iowa. Annu.
  Rep. Eng. Res. Inst., Iowa State Univ., Ames.
Box, G.E.P., and G.M. Jenkins. 1970. Time Series Analysis:
  Forecasting  and  Control.  Holden-Day, San  Francisco,
  Calif.
Box, G.E.P., and  G.C.  Tiao.  1973. Bayesian  Inference in
  Statistical Analysis. Addison-Wesley, Reading, Mass.
	. 1975.  Intervention  analysis with applications to
  econometric and  environmental problems. J. Am. Stat.
  Ass. 70:70-119.
Chatfield, C. 1975. The Analysis of Time Series: Theory and
  Practice.  Chapman and Hall, London.
Chen, C.W.  1970. Concepts and utilities of ecologic models.
  J. San. Eng. Div. Am. Soc. Civil  Eng. 96(SAS): 1085-6.
Cogger, K.0.1979. Time-series analysis and forecasting with
  an absolute error criterion. In S.  Makridakis and S.C.
  Wheelwright, eds. Forecasting. North-Holland, New York.
Hipel, K.W., W.L  Lennox, T.E. Unny, and A.I. McLeod. 1975.
  Intervention analysis in water resources.  Water  Resour.
  Res. 11(6):855-61.
Jenkins, G.M. 1979. Practical Experiences with Modeling and
  Forecasting Time Series. G. Jenkins & Partners  Ltd., St.
  Helier, England.
Kendall, M.G. 1973. Time-Series. Hafner Press. New York.
Kennedy, R.H., K.W. Thornton, and J.H. Carroll. 1980. Sus-
  pended sediment gradients in Lake Red Rock. Presented
  at Am. Soc. Civil Eng. Symp. Surface Water Impound-
  ments, Minneapolis, Minn. June 2-5.
Lettenmaier, D.P., K.W. Hipel, and A.I. McLeod. 1978. Assess-
  ment of environmental  impacts. Part two: Data collection.
  Environ. Manage. 2(6):537-54.
McKerchar, A.I., and J.W. Delleur. 1974.  Application of sea-
  sonal parametric linear stochastic models to monthly flow
  data. Water Resour. Res. 10(2):246-55.
Montgomery, D.C., and LA. Johnson. 1976. Forecasting and
  Time Series Analysis. McGraw-Hill, New York.

Newbold, P. 1979.  Time-series model  building and fore-
  casting: a survey. In S.  Makridakis and S.C. Wheelwright,
  eds. Forecasting. North-Holland, New  York.
Reckhow, K.H. 1979. Empirical lake models for phosphorus:
  Development, applications, limitations and uncertainty. In
  D. Scavia and  A. Robertson, eds. Perspectives on Lake
  Ecosystem Modeling. Ann Arbor Science, Ann Arbor, Mich.
Spofford, W.O.  1975. Ecological  modeling  in  a  resource
  management framework: an introduction. In C.S.  Russell,
  ed. Ecological  Modeling. Resources for the Future, Inc.,
  Washington, D.C.
Standard Methods for the Examination of Water and Waste-
  water. 1975. 14th ed. Am. Pub.  Health Ass. Washington,
  D.C.
Statistical  Analysis System. 1979. Statistical Analysis Sys-
  tem Inst. Inc., Gary, N.C.
Tiao, P.C., and J.W. Delleur.  1976. Seasonal and nonseasonal
  ARIMA models in hydrology. J. Hydraul. Div. Am. Soc. Civil
  Eng. HY10:1541-59.
                                                     181

-------
 MODELING DEVELOPMENTS ASSOCIATED WITH
 THE UNIVERSITY  LAKES RESTORATION PROJECT
RONALD F.  MALONE
DANIEL G.  BURDEN
CONSTANTINE E. MERICAS
Department of Civil Engineering
Louisiana State  University
Baton Rouge,  Louisiana
            ABSTRACT

            Water quality models have been applied to the six University Lakes of Baton Rouge, La., to
            assist in the analysis of effects of restoraticn efforts. Phosphorus was identified as the nutrient
            limiting algal growth in these lakes. A relationship between total phosphorus and fishkills was
            identified. Initial applications of a modified Vollenweider model indicated that the model was
            capable of projecting long-term average late conditions but that it was limited by its inability to
            represent short-term variations. These variations were crucial for  the projection of fishkill
            episodes. Intensive surveys failed to identify practical modifications to the model within a deter-
            ministic format and a stochastic approach was undertaken. Uncertainties were partitioned into
            terms representing  lumped modeling errors and sampling variability. This approach produced
            results that were suitable for interfacing with objective functions based  upon total phosphorus
            levels. The need for  more widespread application of stochastic techniques was evident from ex-
            periences associated with this project.
INTRODUCTION

As  the  engineering and  scientific community in-
creases its understanding of the processes controll-
ing  eutrophication,  lake  restoration projects havs
become more commonplace. Engineers involved with
such projects will frequently find that accurately pro-
jecting system responses to management alternatives
is difficult. Eutrophication modeling responds to this
need  by  identifying  mathematical  relationships
among  loading, morphological,  and water  qualit/
parameters based on both theoretical considerations
and empirical evidence.
  This paper discusses  enhancing eutrophication
models through considering uncertainties associated
with modeling lake systems. This approach considers
uncertainties stemming from natural variability in lake
concentrations of  nutrients, model  limitations, and
the  statistical nature of monitoring programs. The
resulting  models  can  accurately quantify  lake
responses in a format compatible with the evaluation
of management alternatives.
SITE DESCRIPTION

The University Lake System (Fig.  1) consists of six
small urban lakes ranging in size  from 1.17 to 87.89
hectares. The lakes were formed in the 1930's when
low lying cypress swamps  were timbered and darr-
med. The expansion of the Louisiana State University
campus to the west and rapid residential development
to the east led to the development  of causeways and
drainage systems that subdivided the original lake in-
to its present configuration  of six lakes.
  These lakes are representative of a large number of
small urban waterbodies in  the south that have been
adversely affected by intense development of surroun-
ding lands. Increased nutrient loading  caused by
deteriorating runoff quality has led to highly eutrophic
(or hypereutrophic) conditions. In turn,  these condi-
tions reduce the lakes' recreational value and overall
aesthetic quality. A lake restoration project sponsored
by the U.S. Environmental Protection Agency and the
City-Parish Government of East Baton Rouge was in-
   College Lake,
Figure 1—Schematic of the University Lakes.
                                               182

-------
                                                                   MODELING TECHNIQUES AND INNOVATIONS
itiated in 1977 to deepen the lakes, correct sanitary
sewer leaks, and correct  runoff problems. It is an-
ticipated that this restoration effort will continue into
the fall of 1984.
restoration period. The deterministic model  used to
model Crest  Lake  during  the  restoration  period
(Mericas and Malone, 1982) may  be summarized as
follows:
BACKGROUND

A predominant problem in managing hypereutrophic
systems is the periodic catastrophic events that result
from the  rapid growth and  subsequent collapse of
algal  populations (Barica,  1978,  1980). The  conse-
quences of such growth and collapse cycles include
massive fishkills and odor problems associated with
floating masses of decaying  algae or fish. The matter
is complicated by the fact that these waters are so
high on the trophic scale that the extreme abundance
of nutrients  minimizes their  limiting  effects (Mur,
1980). Climatological factors  have also been identified
as major factors in summer collapses of algal popula-
tions  in hypereutrophic systems (Barica, 1978;  Sw-
ingle, 1968).
  The  major problem with the University Lakes was
the high frequency of summer fishkills. Three or four
major fishkills were typically observed  in the system
each year prior to restoration. The fishkills are the
result of severe oxygen depletion resulting from the
respiration and decay of dense algal populations. The
kills conditions appear to  be triggered by  abrupt
changes in weather conditions, but have always been
associated with total phosphorus (TP) levels above 0.4
mg/l (Mericas, 1982). Management objectives (and the
corresponding  modeling  efforts) were,  therefore,
directed toward reducing TP  levels below the 0.4 mg/l
threshold  value.
  Modeling  of total phosphorus (TP) has been a cen-
tral theme in most  of the  eutrophication research
published in the past 15 years. Summaries and discus-
sions of these modeling efforts   are presented by
Reckhow  (1981), Mercil et al.  (1980), and Jorgensen
(1978). Much emphasis has been placed on the basic
mass-balance model originally described by  Vollen-
weider  (1969). This model has evolved  into formats
that typically predict mean  annual TP from various
morphological parameters and loading estimates. The
model  is usually used with annual mean TP levels to
predict trophic conditions that will result from propos-
ed lake changes.
  Total phosphorus concentrations in Crest Lake, one
of the smallest lakes undergoing restoration, were
continually  modeled  by a  modified  version of  a
Vollenweider model for the latter part of the 31/4-year
     dP _ Q|P|
     dt =   V
Qex           Q0P
T7(Pu-P)--TT      (D
where
      P  =  mean in-lake TP concentration (mg/l)
      PI  =  inflow TP concentration (mg/l)
      Pu  =  inlake TP concentration in University
            Lake (mg/l)
    Qex  =  wind driven interlake exchange rate
            (m3/day)
      QI  =  inflow rate (m3/day)
     Q0  =  outflow rate (m3/day)
       t  =  time (days)
      V  =  lake volume (m3)
       a  =  net sedimentation coefficient (day-1).
This model describes loading to the lake from  both
rainfall runoff and wind-induced interlake exchange
through a culvert and pipe running under a causeway
that separates the two lakes. A submodel defining the
Qex term was developed  as follows:
    Qex = f(r2)
Figure 2.—Results of the  modified Vollenweider  model
simulations for Crest Lake for the period of July 1979 through
August 1982.
where

       f  =  empirically defined wind exchange con-
            stant (m3/d-mph2)
       r  =  resultant daily wind component along
            the lake axis (mph).

  Results of the modeling effort on Crest Lake were
analyzed to determine the compatibility of the Vollen-
weider model with short-term applications to highly
eutrophic southern lakes.  Input parameters, such as
rainfall and  wind  speed, were considered on a daily
basis to  see if  the model  could represent the short-
term variations that strongly influenced the occur-
rence of summer fishkills. Model  simulations were
compared to total phosphorus observations collected
on the lake at intervals as short as 1 day. Figure 2 il-
lustrates the relationship between  the model projec-
tions and observations for the various phases of the
interim restoration effort (Mericas and Malone, 1982).
These results indicated that the model can be used to
analyze system responses on a seasonal basis, but
does not truly reflect short-term variations of observa-
tions collected on a weekly or daily timeframe.
  Hypereutrophic systems such as  Crest Lake are
particularly sensitive to changes in both the physical
environment and the phytoplankton community (Uhl-
mann, 1980). Modeling is complicated by the inherent
short-term and essentially random variability in water
quality parameters resulting from the unstable nature
of these systems. Lake  problems such as fishkills are
typically associated with extremely high (or low) water
quality fluctuations.
  Yet, the deterministic structure of the basic Vollen-
weider model  prevents any concise estimation of
these variabilities. In the deterministic format, the
model can only project a single state variable value for
any  set  of input  parameters.  Deterministic  lake
nutrient models have proven to be invaluable manage-
ment tools, particularly in analyzing fundamental pro-
                                                 183

-------
LAKE AND RESERVOIR MANAGEMENT
 cesses controlling average lake quality. However, 1o
 accurately  represent short-term variations,  deter-
 ministic models must entail high levels of complexity.
   Higher orders  of  structural complexity increase
 data requirements for calibration which, in turn, result
 in  higher  costs and  extensive  time delays.  Thus,
 although functional relationships between meteoro-
 logical parameters and short-term water quality varia-1
 tions undoubtedly exist, the data and evaluation ef-
 forts required to  identify these relationships would
 limit the practicability  of  the  resulting  model for
 engineering applications.
   Furthermore, many of the most significant model in-
 puts such as rainfall, wind speed, or cloud cover aie
 often stochastic in nature and functionally undefin-
 able under the deterministic modeling format. So, the
 value  of a  deterministic projection, essentially  a
 single system response, is further undermined by the
 random nature of these input parameters and the daily
 variability  observed  in hypereutrophic  systems
 (Uhlmann, 1980).
STOCHASTIC MODELS

As an alternative to a very complex and often uncer-
tain deterministic approach, it has been proposed that
a stochastic modeling approach be developed to ad-
dress the problem. Several considerations make a
stochastic  representation  particularly attractive.  A
stochastic  model recognizes those parameters that
are essentially random by dealing with them directly
as random variables. Additionally, noise terms may be
included to account for numerous minor effects that
tend to have a cumulative impact. Thus, a model may
be constructed based on the major system influence!?
that may be reliably represented as mathematical rela-
tionships, along with stochastic terms representing
the mass of essentially random effects that contribute
to system variability.  System responses that cannot
be reliably predicted, such as population collapses,
may  be  represented in terms of uncertainty.
  Mathematically, representation of  variability or
uncertainty is usually accomplished by  adding an
uncertainty term to an existing deterministic relation-
ship. In  the case of the Crest Lake modeling effort, a
white noise term was introduced as  a  variability
parameter to the Vollenweider model. The white noise
term as applied here represents a lumped uncertainty
term. It  reflects  variability induced by the various
loading  sources as  well  as incidental   errors
associated with the model's structure and calibration
processes. A  white noise term has  the  following
characteristics as a random variable:
                                                      jected TP value,  however, is quantified  through the
                                                      stochastic form of the model:
    E[w] = 0

    E[$ - E(w))2] = S2

where
                                               (2)

                                               (2)
    S^ =  variance of the white noise term
     w =  random variable white noise factor (mg/l-
           day)

A  white  noise variable con-tributes only variability,
represented by the variance, S&, to the model. Thus,
the mean projection of the Vollenweider model, which
has been recognized as being reliable, is not influenc-
ed by the addition. The variability  in the mean prc-
                                                           dP
                                                           dt
                                         + w  (4)
                                                      where
                                                            P = random variable of in-lake TP concentra-
                                                                 tion (mg/l)

                                                        This  differential  equation can be solved  by
                                                      numerical or exact techniques (Malone et  al. 1983)
                                                      similar to those used with deterministic models. The
                                                      resulting  projections  (Fig.  3)  include  statistical
                                                      estimates of the variability induced by the sources of
                                                      uncertainty considered  in the white noise term.  In
                                                      most cases, these projections will estimate average or
                                                      mean in-lake conditions.
                                                        If  the resulting  projections are  interfaced with
                                                      statistical results of the monitoring program they can
                                                      be compared directly to individual observations of a
                                                      water quality parameter (Fig. 4) or  with design objec-
                                                      tives  based upon  individual  observations  (Fig.  5).
                                                      Quantitative estimates for the expected frequency of
                                                      violating in-lake water quality criteria are easily deriv-
                                                      ed  from  stochastic outputs  by  applying  well-
                                                      established probability theories.
                                                         050


                                                         045


                                                         040


                                                         035


                                                         030


                                                         025


                                                         020


                                                         015


                                                         010


                                                         005
                                                                            July 15
                                                                             1982
                                                                                               Aug 10
                                                      Figure 3.—Stochastic model projections and observed mean
                                                      phosphorus levels from an intensive monitoring program of
                                                      Crest Lake.
  05O


  045


S 040


1 035

V)
i 030
o

£ 025


I 020
Q.



f °'5
° 010


  005
                                                                             July 15

                                                                             1982
                                                                                        Aug I    Aug 10
                                                     Figure  4.—Interfaced  stochastic  model projections  and
                                                     representative individual phosphorus observations from the
                                                     intensive monitoring of Crest Lake.
                                                  184

-------
ADVANTAGES

The  authors' experiences  with  stochastic  models
developed  in conjunction with the restoration of the
University  Lakes indicate  that  stochastic  models
could potentially be very useful in lake management.
The principal advantage of their use stems from their
ability to quantitatively represent the variability  in-
herent in lake  systems. Thus  models based upon a
limited understanding of the complex processes con-
trolling lake quality can still be used quantitatively as
a basis for lake management decisions.
  More sophisticated models  can  be enhanced  by
considering residual uncertainties in their projections.
In either case, the lake manager bases decisions with
full knowledge of the reliability of his modeling projec-
tions.
  Since  stochastic  models can be  used to enhance
the performance of the simpler lake models, they may
reduce  monitoring costs often associated  with the
support of more sophisticated stochastic approaches.
Stochastic  techniques can represent  uncertainties
resulting  from  applying  models  across  different
systems (i.e., calibration on one lake and prediction on
another) and may provide a vehicle for assimilating
knowledge gained from numerous lake restoration ef-
forts.

LIMITATIONS

Most of the disadvantages associated with stochastic
models stem from the fact  they are not widely used.
The  foremost  problem  that  will confront  the lake
manager interested  in applying a  stochastic model is
the almost complete absence of computer software in
this  area.  Most  stochastic lake models currently  in
use were written by the user or were modifications of
existing deterministic codes. Those lake managers at-
tempting to develop a site-specific stochastic model
will quickly learn that the terminologies and methods
associated with  stochastic processes  are  initially
foreign to those with no experience or formal training
in the subject. It takes a while to learn to think in terms
of random variables rather than discrete values. Until
this  conceptual  hurdle is overcome, communicating
the technical aspects of the modeling  process is dif-
ficult.
 FUTURE NEEDS

 Without wide application of stochastic models, it will
 be some time before their full potential as predictive
 tools is realized. Applications for the near future will
 be limited to those lake programs  with  ongoing or
                MODELING TECHNIQUES AND INNOVATIONS

historic monitoring  programs  with  a  sufficient data
base to support site-specific determination of coeffi-
cients. Research should be directed  at determining
which factors  affecting lake water quality  are best
handled  deterministically and which stochastically.
Standardized formats for handling selected factors in
a stochastic manner should be developed to facilitate
the development of a broad-based data pool for future
applications.

ACKNOWLEDGEMENTS:  Funding for  this research  was
derived in part from the U.S. Environmental Protection Agen-
cy, City-Parish Government of East Baton Rouge, and the
State of Louisiana through a cooperative lake restoration ef-
fort under the Clean Lakes Program. This paper was not sub-
ject to review by the funding agencies; findings reflect the
opinions of the  authors only. The data base used  in this
paper includes contributions by Glenn McKenna and Andrew
Eversull.
REFERENCES

Barica, J. 1978.  Collapse of aphanizomenon  flos-aquae
  blooms resulting in massive fish kills in eutrophic lakes:
  effect of weather. Verh. Int. Ver. Limnol. 20: 208-13.

	1980. Why hypereutrophic systems? Pages ix-xi in
  J. Barica and L Mur, ed. Developments in Hydrobiology.
  Vol. 2. W. Junk, The Hague, The Netherlands.

Jorgensen, S.E.  1978. State of  the  art in  eutrophication
  models. Pages 293-8  in State-of-the-Art  in  Ecological
  Modelling.  Vol.  7.

Malone, R.F., D.S. Bowles, M.P. Windham, and W.J. Grenney.
  1983. Comparison of techniques for assessing effects of
  loading uncertainty upon a long term phosphorus model.
  Appl. Math. Model. 7: 11-18.

Mericas, C. 1982. Phosphorus dynamics and the control of
  eutrophication  in a southern urban lake. Master's Thesis.
  Louisiana  State Univ., Baton Rouge.
Mericas, C., and R.F. Malone. 1982. Short-term application of
  a  modified version of the basic  Vollenweider nutrient
  model  to a hypereutrophic urban lake. Pages 609-614 in
  W.K.  Lauenroth, G.V. Skogerboe, and  M.  Flug,  eds.
  Analysis  of  Ecological  Systems:  State-of-the-Art in
  Ecological Modelling. Elsevier Scientific Publ.  Co.,  New
  York.
Mercil, S.B.,  C.M. Conway, and L.E. Shubert. 1980. Phos-
   phorus stability in a hypereutrophic lake. Pages 179-80 in
   J.  Barica and L.R.  Mur,  eds. Developments  in Hydro-
   biology. Vol. 2. W. Junk, The Hague, The Netherlands.
 Mur, L. 1980. Concluding remarks. Pages 331-3 in J. Barica
   and L. Mur, ed. Developments in Hydrobiology. Vol. 2. W.
   Junk, The Hague, The Netherlands.
 Reckhow, K.H. 1981. Lake data analysis and nutrient budget
   modeling. EPA-600/3-81-001.  U.S. Environ.  Prot.  Agency,
   Washington, DC.
 Swingle, H. 1968. Fish kills caused by phytoplankton blooms
   and their prevention. FAO Fish. Rep. 44(5): 407-47.

 Uhlmann, D. 1980. Stability and multiple steady states  of
   hypereutrophic ecosystems. Pages 235-47 in J. Barica and
   L.R.  Mur, eds. Developments in Hydrobiology. Vol. 2. W.
   Junk, The Hague, The Netherlands.

 Vollenweider, R.A.  1969. Possibilities and  limits of ele-
   mentary models concerning the budget of  substances in
   lakes. Arch. Hydrobiol. 66(1): 1-36.
                  020     030     040
                  TOTAL PHOSPHORUS (mg/i)
 Figure 5.—Interfaced model projections and the zone of fish
 kill occurrences.
                                                    185

-------
  A CROSS-SECTIONAL MODEL  FOR  PHOSPHORUS  IN
  SOUTHEASTERN  U.S. LAKES
 KENNETH H. RECKHOW
 J. TREVOR CLEMENTS
 School of Forestry and Environmental Studies
 Duke University
 Durham, North Carolina
             ABSTRACT

             Data from 42 lakes and reservoirs in Virginia,, North Carolina, South Carolina, and Georgia were
             used to develop and evaluate cross-sectioneil regression models for phosphorus. For all models
             under consideration, the endogenous variable was lakewide average annual phosphorus con-
             centration, and the exogenous variables were phosphorus input to the lake and various geomor-
             phologic  and hydrologic  variables.  A linear  model was developed  for  logarithmically-
             transformed terms that resulted in a substantially lower prediction sum of squares than did
             other models under consideration. Among the other models examined were those specified and
             fitted to north temperate lakes by other investigators. Parameter characteristics and error terms
             for the proposed model were studied with particular attention directed to the a priori assumption
             of a strictly linear relation between input ard concentration of phosphorus. The calculation of
             prediction uncertainty is illustrated in an application of the model using an empirical Bayes ap-
             proach and a prior model based on linearity.
INTRODUCTION

Most  simple  input-output, black-box,  or  cross-
sectional regression models for phosphorus that have
been widely used were developed from data on north
temperate lakes (e.g., Dillon and Rigler, 1975; Vollen-
weider, 1976; Larsen and Mercier, 1976; and Reckhow
and  Simpson, 1980).  Under the  quite  reasonable
assumption (supported by the work presented herein)
that climate affects model specification and/or model
parameters, this state of affairs means an abundance
of options for the temperate climate model users but
an uncomfortable choice for potential model  users
desiring to work  in other  climates.  This latter in-
dividual could  choose to develop a model (a choice
few have elected), or choose a temperate zone model
under the assumption that climate is unimportant. The
work presented  herein was motivated by the  im-
mediate needs for another option and also by a desire
on the part of the investigators to examine the proper-
ties of cross-sectional data and models  under non-
temperate conditions. The objective of the work,  there-
fore, was to develop, test, and apply a cross-sectional
model for phosphorus in southeastern U.S. lakes.
CROSS-SECTIONAL MODELS

Cross-sectional data sets and models have proven to
be effective in understanding certain aspects of lim-
nological behavior common to many lakes. Particular-
ly for small lakes, the cross-sectional study provides a
basis for projecting aggregate behavior that would
simply not be available in most cases because of the
funding and effort required to study even a single lake
over a long time  period.  Frequently,  these cross-
sectional models are used  to predict behavior in a
single lake at  different points in time. A necessary
assumption for this use is  that the cross-sectional
behavior described in the model is equivalent to the
time series  behavior for a single  lake (in  discrete
steps, under the steady state assumption).
  Stated another way, it is assumed that behavior in
an application lake at a different point in time is simp-
ly another realization of the population for which the
model development  data set is a (random?) sample.
Thus if the model is representative of a group of lakes
from which the application lake is selected, then it is
assumed that the model may be used on the applica-
tion lake to  predict  behavior at other points  in time
(providing that the scenarios evaluated do not reflect
conditions that  effectively remove the  lake from the
applicable population).
  Using  cross-sectional regression in the manner
described, the modeler  gains  a  model that has a
relatively wide inferential base. This means that the
model, if properly developed, may be used to predict
behavior in all lakes in the population represented by
the model.  In  contrast,  a model specified  and
calibrated for a single  lake may not be applied to
another  lake without scientific confirmation that it is
appropriate (a nontrivial exercise).
  Model interpolation (the permissible use of a model
without  additional information) is clearly preferred to
extrapolation (use of the model outside the bounds of
the model development data set), as extrapolation in-
cludes  additional  error that may  be difficult  to
estimate. The cross-sectional  model  is  attractive
because multi-lake applications may involve  strictly
interpolation. The single  lake model,  on  the other
hand, cannot be interpolated to another lake unless it
is re-specified and re-calibrated for that lake.
  The "costs" of the aforementioned attributes of the
cross-sectional  model are  larger  error terms  than
those to be  expected  from equivalent single lake
models.  The  large error terms result from natural dif-
                                                186

-------
                                                                   MODELING TECHNIQUES AND INNOVATIONS
ferences or variations among lakes; to some extent
the errors represent variables  not included in  the
model. In another sense, the additional error in  the
cross-sectional model is the (additional) extrapolation
error that must  be considered when a single lake
model is extrapolated. So, in effect, by using a cross-
sectional model, the modeler accepts the larger cross-
sectional error in exchange  for the wider  inferential
base (and shorter "inductive leap") that the  model pro-
vides (in comparison to that for single system simula-
tion models).
  Of course, it is conceivable that the cross-sectional
error is so large that the model has little predictive
power.  This  will  occur when  behavior  variations
among lakes overwhelm the similarities. In  that situa-
tion, the uniqueness of each  lake would be para-
mount, and limnologists would be able to make  few
general statements about limnological  behavior. For-
tunately, limnologists have found that lakes have both
similarities and uniqueness,  and that it is possible to
generalize  abut   the   similarities.  Cross-sectional
models are simply one way to express the similarity in
behavior across lakes.
  The usefulness of the cross-sectional models for
lake management will  ultimately depend on the size
of the cross-sectional  error attributable to lake  uni-
queness. If this error is thought to be large and the
cost of developing a single lake model is less than the
perceived value of new information  to be obtained
from the single lake model, then the modeler would be
wise to  specify and calibrate  a single lake model.
However, if  the  cross-sectional similarities  are
relatively great, and the cost  of  single lake  model
development cannot be justified (in terms of new infor-
mation obtained on unique lake characteristics), then
the  modeler  may wisely use the  cross-sectional
model. This  work proceeded on the belief that the lat-
ter of the two cases is probably  true in  many lake
management situations.
           Table 1.—Data set characteristics
 Variable
Minimum
Median
Maximum
A (km2)
z(m)
TW (yr)
L (g/m2-yr)
qs (m/yr)
R
P,n (mg/l)
P (mg/l)
P0 (mg/l)
0.81
1.50
0.016
0.06
2.30
-0.11
0.015
0.007
0.007
19.74
9.35
0.118
4.23
66.65
0.41
0.063
0.033
0.040
447.59
41.30
1.65
93.3
650.20
0.89
0.259
0.143
0.145
THE MODEL DEVELOPMENT DATA SET

The cross-sectional model calibration and evaluation
were based on a data set of 42 lakes and reservoirs
located in Virginia, North Carolina, South  Carolina,
and  Georgia.  By selecting lakes  in only these four
southeastern U.S. coastal States, it was believed that
the influence  of climate on lake behavior  could  be
maintained at an acceptably low level. The data were
taken from working papers based on the U.S. Environ-
mental Protection  Agency's National Eutrophication
Survey conducted in the mid-1970's (U.S. Environ. Prot.
Agency, 1976).
  The sampling  programs  were relatively uniform
across lakes. Lake samples were taken at one or more
stations  for  three  dates  spread over the  growing
season. Depth samples were taken from surface to
bottom at regularly spaced intervals. Measurements
of tributary concentrations occurred 12 to 14 times
during the 1  year  of tributary sampling, on approx-
imately a monthly basis. Flow estimates were provid-
ed by the U.S. Geological Survey; these were usually
obtained from gauging stations, and if necessary in-
volved extrapolation to ungauged  sites.
  Table 1 contains statistics summarizing the data
set for variables relevant to this  study. Most of the
cross-sectional histograms of these variables were
skewed right and could be approximated by lognormal
distributions.  The  one exception was  phosphorus
retention (R) which yielded a histogram that was ap-
proximately normal in shape. Probably the  one note-
worthy feature of this data set (in comparison to data
sets of similar size from temperate regions) is the
absence of lakes with high values of hydraulic deten-
tion time (TW). This is probably because the Southeast
has few large, deep natural lakes. Most of the lakes in
the data set are artificial impoundments with relative-
ly large flow rates  and correspondingly low detention
times.
  Sample correlation coefficients are presented  for
the data set in Table 2. All variables except  R are log-
transformed.  Bivariate  plots  indicate  approximate
bivarlate normality  in most situations; of more impor-
tance, though, is the fact that exploratory  graphical
studies reveal no  evidence that overly  influential
cases exist in the  data set. This point is  again ad-
dressed in the model development section.
   For the model, lake phosphorus concentration (P) is
the  dependent variable.  Based on  previous  ex-
perience, log  P was selected as the metric for model
estimation,  since  the  logarithmic transformation is
likely to lead to model residuals that satisfy the or-
dinary least squares assumptions. Correlation coeffi-
cients in Table 2 and bivariate plots indicate that areal
water loading (qs), areal phosphorus loading  (L),
hydraulic detention time (TW) and influent phosphorus
concentration (Pin) are the best choices for predictor
                                  Table 2.—Sample correlation coefficients.
             logP
         logAs
       log z
       logq.
       logL
logrH
log Pir
logP
log As
log z
logqs
logL
logrw
log P,n
R
logP0
-0.187
-0.386
0.563
0.819
-0.716
0.858
- 0.082
0.925
0.090
- 0.548
-0.346
0.550
0.114
0.524
-0.232

0.116
0.001
0.404
-0.176
0.286
-0.367


0.895
- 0.861
0.346
- 0.395
0.621



- 0.823
0.728
-0.130
0.828




-0.408
0.513
-0.760





0.339
0.785 -0.285
                                                  187

-------
LAKE AND RESERVOIR MANAGEMENT
variables (P0 is a response variable, like P). Given the
extent  of  multicolinearity among these cross-
sectional  variables, a model with more  than two ex-
ogenous variables will  cause parameter  estimation
and prediction difficulties.  Therefore, either a one
variable model, or more likely, a two variable model in-
volving log TW and log Pin appears to be  the model of
choice. It is to be noted that this conclusion is consis-
tent with prior beliefs about model  specification (bas-
ed  on north temperate models), so the  data are not
rigorously being asked to both specify and calibrate
the model (which could raise questions in significance
tests; see Learner, 1978).
MODEL  DEVELOPMENT AND EVALUATION

Parameter estimation, using ordinary least  squares
(Ray, 1982), yielded the following two-variable model:

  log(P) = -0.887 + 0.717log(Pin) - 0.278log(rw) (1)

In the original metric, this model is:

  P = 0.130 (Pin)0.717/Tw0.278                      (2)

  Table 3 contains  summary  information  for this
model. The root mean square  error (Root  MSE), cr
model error (sy/x), is  0.116; this  is smaller than most
model error terms calculated for north temperate lake
models  (see Reckhow and Chapra, 1983, chapter 8).
For the purpose of  prediction,  however, we recom-
mend that the root mean square error of prediction
(Root PRMSE) be  used  in place of the root  mean
square error.
  The  root PRMSE  is calculated  as  follows. One
observation is removed from the data set, the regres-
sion model is estimated,  and then this model is used
to calculate a prediction  for the "removed" observa-
tion. The difference between this prediction  and this
observation is a "residual." This exercise is repeated
for all n = 42 observations; thus in each case 41 obser-
vations are used to estimate the  model and one obser-
vation is used to calculate a residual. The residuals
are then squared, summed, and divided by the number
of degrees of freedom. The result is the PRMSE (take
the square root to determine  the root PRMSE). This is
considered a  prediction error term  since the  residual
is determined for a single  observation not in the model
development data set (which  is a characteristic of ac-
tual predictive applications).  Note that, as expected,
the  root PRMSE is greater than  the root MSE for the
model presented in Table 3.
                                        Most (cross-sectional) north temperate lake models
                                     specify a linear relationship between lake phosphorus
                                     concentration (P) and phosphorus input (L or Pin). For
                                     example, the model developed concurrently by Vollen-
                                     weider (1976) and Larsen and Mercier (1976) is of this
                                     type:
                                                Pin
                                       P =
                     d
                                                 TwO-5)
                                                                                    (3)
                                     A  priori  we  expected that this  linear relationship
                                     would  also hold for the chosen  model  for south-
                                     eastern U.S.  lakes. In fact, it is  apparent from the
                                     statistics for the model presented in Table 3 (and from
                                     statistics for  other model specifications considered),
                                     that the hypothesis of linearity between phosphorus
                                     input and phosphorus lake concentration must be ten-
                                     tatively rejected for the  southeastern  U.S.  lakes
                                     studied. This  conclusion is  based on the fact that the
                                     parameter  for  log  Pin (0.717) differs from  1.0  (the
                                     parameter value necessary for linearity) by almost five
                                     times its standard error (0.060). Under assumptions of
                                     normality and random sampling, it is highly unlikely
                                     that the true  P/Pin relationship for the population of
                                     southeastern  lakes involves linearity.  Note, however,
                                     that we are examining a cross-sectional model. This
                                     informal  hypothesis  test  says nothing  about  the
                                     plausibility of a linearity  hypothesis for a single  lake
                                     time series.
                                       A plot of predictions versus observations for the
                                     chosen model (Equation  1) is  presented in Figure 1.
                                     Few if any points stand  out from the trend and  can
                                     readily  be called  "influential" (in  terms of  their
                                     singular effect on the specification and/or parameters
                                     of the regression model). If we want the model to be as
                                     representative as possible  of  the selected group of
                                     lakes,  then several  diagnostic statistics  and plots
                                     should be examined. We briefly discuss these now to
                                     underscore the  point that the chosen model (Equation
                                     1) is in fact a  good model for our data set.
                                       Figure 2  is  a  plot of the model residuals versus the
                                     predicted P values. It is typical of the form of other
                                     residuals plots  for this model. Residuals plots for a
                                     regression  model are studied to look  for outliers (in-
                                     fluential observations) or unmodeled patterns in the
                                     data. Absense of these two problems should give the
                                     modeler more confidence that the regression equation
                                     is a reasonable model (in terms of representativeness
                                     and certain assumption violations). In Figure 2, we see
                                     that the data are randomly scattered, indicating no
                                     discernible pattern has been missed.  Perhaps one in-
                                 Table 3.—Southejistern lakes and reservoirs.

                                 Reckhow-Clements log-log regression model

  MODEL: Log P =  -0.887 - 0.278 Log TW + .717 Log Pin
  Root MSE
  F-value
  pred. ESS
  Root PRMSE
  0.116
168.014
  0.592
  0.123
adj. R-Sqiare
Prob>F
Sample size
 0.89
 0.0001
42
  variable
         parameter
          estimate
         standard
           error
                                                                        T-value
                                                                             ProblTI
  Intercept
  Logrw
  Log Pin
           -0.887
           - 0.278
            0.717
           0.090
           0.036
           0.060
                  -9.829
                  -7.760
                   11.999
0.0001
0.0001
0.0001
                                                 188

-------
                                                                     MODELING TECHNIQUES AND INNOVATIONS
fluential  observation  at the bottom of the plot may
warrant further study. Otherwise, the model appears
good on the basis of this (and other) residuals plot(s).
  Influential  observations may be more clearly iden-
tified using one of several diagnostic statistics pro-
posed  by  Belsley et al. (1980) and  by  Cook  and
Weisberg (1982).  For  example, Figure 3 presents in-
fluence statistics for the model parameters (the |3's)
for TW and Pin. The D|3's in Figure 3 basically represent
the effect on a parameter resulting from each observa-
tion  (one observation  per data point in Figure 3), scal-
ed by the parameter variance (see Ray, 1982 for the ex-
act equation for the D/3's). The X in the center of the
cluster is the baseline parameter value (on each axis)
for all 42 observations. Points far from the center (the
X) represent single observations in the model develop-
ment data set that particularly influence the value of
either or both parameters. The vertical or horizontal
distance between a  point and  the X is  the change
(scaled by  the variance) to be expected  in j) by remov-
ing the single observation represented by that point.
  Points that are far from the X may be deemed in-
fluential  according  to  a statistical  criterion (see
Belsley et al. or Cook and  Weisberg for details).
Perhaps  more important,  however, this plot is useful
for identifying influential  observations  that can later
be examined on a substantive basis. We prefer not to
reject observations strictly for statistical reasons, but
rather to search  for a substantive (e.g., limnological)
basis for rejection, if it exists. Therefore, the influence
diagnostics  are  used  here  simply to  identify
"outliers." The  outliers  identified with  the aid of
Figure 3 were found to be acceptable on a substantive
basis (i.e.,  no unusual features of the lake to suggest
faulty  data,  for  example, based on the  U.S. EPA
reports) and therefore all observations were left in the
data set  for final  model estimation. From a statistical
basis, the two observations at the extreme left and the
two  observations at the extreme right were found to
    0.100
     0.063
     0.040
     0.02S
     0.016
     0.010
                      0.717
               P = 0.130 Pln / -,
             0.006  0.01   0.016 0.02S  0.040  0.063  0.100

                   Observed P (mgfl)
    0.30


    0.25


    0.20


    0.15


    0.10


    0.05


     0


   -0.05


   -0.10


   -0.15"


   -0.20


   -0.25


   -0.30
            -1.9   -1.7   -1.5    -1.3    -1.1   -Oi9

                       Predicted Value* (Log P)
Figure 2.
                                                       Of,
                                                               -030   -0.20  -0.10
 Figure 1
Figure 3.

exceed the statistical rejection criterion proposed by
Belsley et al. (Dps > 2/n°-5  = 0.309 for this data set,
where n = 42 observations). Nevertheless, we chose to
leave these points in the data set, because we found
no substantive basis for removing them.
  Probably  the  most  widely  used cross-sectional
regression model is that  presented in Equation 3.
While  it was originally developed, in one case, from a
data set composed primarily of nutrient-poor north
temperate  lakes (Larsen  and Mercier, 1976),  it  has
since  been applied  all  over the world,  in a range of
climatic zones and trophic states. Further, the TW rela-
tionship in the denominator of Equation 3 has assum-
ed almost  a transcendental basis in some applica-
tions. We therefore thought it important to study this
model specification, and in particular, evaluate the 0.5
exponent on TW for our southeastern U.S. lakes data
set.
                                                   189

-------
 LAKE AND RESERVOIR MANAGEMENT
   To evaluate the fit of the model specified in Equa-
 tion 3 (with the single parameter, 0.5, allowed to vary),
 nonlinear regressions were run on logarithmic trans-
 formations of each side of the equation  (the trans-
 formations were necessary to  obtain well-behaved
 residuals). For values of the parameter between O.D5
 and 1.0 (at 0.05 intervals), the criterion of fit, root mean
 square error (sy/x),  was calculated. The results are
 presented in Figure 4.
   Figure 4 has several interesting features. First, tie
 optimal value  (minimum sy/x) for the parameter is not
 0.5  (with Sy/x = 0.215), but rather, 0.11 (with sy/x = 0.175).
 Further, 0.5 is not within  the 95 percent asympto:ic
 confidence interval  of 0.11 (which has an asympto:ic
 standard error of 0.0628).  Note  that the root MSE lor
 this model is  considerably higher than the root MSE
 (sy/x = 0.116) for the  model  in Equation 1. Thus, for the
 southeastern  U.S. lakes data set, the 'V^-model" is
 not  the best choice, based on a mean  square error
 criterion.
   This last point raises  an issue: on what basis
 should a model  be  selected for a particular applica-
 tion? Reckhow and  Chapra (1983, chapter 1) discuss
 several criteria for model selection that could be help-
 ful.  For example, it  is desirable (but often  difficult  in
 practical  situations) to apply  a decision theoretic
 criterion  of minimum  risk for  model choice. If we
 restrict our  candidate models  to cross-sectional
 phosphorus lake models,  there are two primary
 statistical  model   selection   criteria:  represen-
 tativeness and  minimum  error.  Representativeness
 means that the model was developed from (or ade-
 quately tested for) a data  set  that contains lakas
 similar to (for important characteristics) the applica-
 tion lake. Thus one would  be unwise to apply a model
 developed for  oligotrophic north temperate lakes to a
 highly  eutrophic southeastern  lake without  prior
 testing. Prior  testing involves running (and possibly
 refitting) the model on several eutrophic southeastern
 lakes.  Evaluating the model on only the application
 lake is unsatisfactory; the  model may happen to fit ex-
 isting lake data by chance  and yet be an unacceptable
 predictive  model for that lake  (because it  is not
 representative).
   The  second model selection criterion is minimum
 error. For all  representative models, the model with
 the  minimum  error (minimum  root  MSE),  calculated
 using a representative data set, is to be preferred. The
 root PRMSE  may  then be used to  determine the
             VollenweMw (Laraen-Merctor) Model
     024|
     O.ZS
 Root
 Mean
Square
 Error
 (Styx)
     0.20!


     0.191



     0181


     0.171







Figure 4.
             Loj (P) - Log (  r» )
                    1+T
0.10    0.25    040    0.55    0.70    0.85    10)
                                                     prediction error for application of the model. The two
                                                     model selection criteria, representativeness and error,
                                                     are equivalent  to  the two terms, bias and variance,
                                                     that are used in statistics to  help  identify  good
                                                     estimators. In the long run, it is better to use these two
                                                     criteria for model selection than to rely on the seren-
                                                     dipity of model fit solely on the application lake.
                                                     AN APPLICATION

                                                     To illustrate use of the proposed southeastern U.S.
                                                     phosphorus lake model and uncertainty analysis, we
                                                     apply an empirical  Bayes approach to pool informa-
                                                     tion from more than one source. The first information
                                                     source (called  the  "prior") is based on  the  a priori
                                                     assumption of linearity between input and lake con-
                                                     centration. The second information source (called the
                                                     "likelihood") is the  southeastern U.S.  lakes model in
                                                     Equation 1. Predictions from these two  sources are
                                                     pooled using Bayes theorem (to calculate the final or
                                                     "posterior"  result).  Details  on the  procedure  are
                                                     presented in Reckhow and Chapra (1983,  section 8.5).
                                                        Lake Rhodhiss, the application lake, is an impound-
                                                     ment in western North Carolina. Based on 1982 data, it
                                                     was found  that:

                                                          TW = 0.040 yr
                                                          Pin =  0.110 mg/l
                                                          P = 0.065 mg/l

                                                        For the  prior  model in the empirical Bayes ap-
                                                     proach,  it was assumed that  a linear relationship ex-
                                                     ists between Pin and P. This means that lake phos-
                                                     phorus retention is a constant, so an appropriate prior
                                                     model is:
                                                          P = Pln(1 - R)
                                                                                        (4)
                                                     To use this model in a empirical Bayes application, we
                                                     need a mean and standard deviation for R. The mean
                                                     (Rm) was  obtained  from  1  year  of  input/output
                                                     measurements, while the standard deviation (Rsd) was
                                                     calculated from measurement errors using first order
                                                     error analysis  (as  demonstrated in Reckhow  and
                                                     Chapra). The results are:

                                                          Rm = 0.41
                                                          Rsd = 0.06

                                                       For  this application, we estimate the change in
                                                     phosphorus concentration to be expected from a re-
                                                     quirement that the major wastewater treatment plant
                                                     discharging into Lake Rhodhiss install a phosphorus
                                                     removal procedure. It is projected  that this will reduce
                                                     phosphorus loading by 15 percent. If W is the total an-
                                                     nual mass loading of phosphorus to a lake, and Q is
                                                     the total annual volumetric water  flow through  a lake,
                                                     then Pin is:
                                                          Pin = W/Q
                                                                                                    (5)
                                                      Using estimates of mean and variance for Q and for
                                                      the projected  reduction  in W,  the projected  mean
                                                      reduction  in Pin and standard deviation for the Pin
                                                      reduction (determined using first order error analysis)
                                                      are:

                                                          (Pin)red = 0.0165 mg/l  +/- 0.005 mg/l

                                                     The "prior" (denoted by P') mean and standard devia-
                                                     tion for the projected concentration are calculated by
                                                  190

-------
                                                                      MODELING TECHNIQUES AND INNOVATIONS
applying the prior phosphorus model (Equation 4) to
predict the projected reduction in P, and by applying
first order error analysis to estimate the uncertainty in
this change (it is assumed that R and Pin are indepen-
dent). The uncertainty in the projected reduction must
be combined with  the  standard error (calculated as
0.0057 mg/l) of the present measured phosphorus con-
centration  to yield the  uncertainty  in  the prior
predicted new phosphorus concentration. This results
in a predicted phosphorus concentration  reduction
(pred +/- standard error) and  prior predicted  new
phosphorus concentration (P' +/- standard error) of:
     Pred = 0.0097 mg/l +/- 0.0031 mg/l

     P' = (0.065 - 0.0097) +/- (0.00312 + 0.00572)0.5

     P' = 0.055 mg/l +/- 0.0065 mg/l

   The  southeastern U.S.  lakes  model  (Equation  1)
yields  the  second source of information (the "likeli-
hood") on  projected  phosphorus concentration.
Because of the additive nature of influent phosphorus
concentration, we can say that:

     (Pin)projected = (PinJpresent + (Pin)red            (6)

Therefore:

     (Pin)proj = 0.110 -  0.0165  = 0.0935 mg/l

Thus the mean likelihood predicted phosphorus con-
centration  (P") is (applying Equation 1):

     P" =  0.130(0.040)-.278(Q.0935).717
     P" =  0.058 mg/l

   Since Equation 1  is not linear in the Pm/P relation-
ship, calculation of the uncertainty in the phosphorus
                concentration  reduction  is  not  entirely straight-
                forward. In fact, the error term cannot be determined
                for the concentration reduction; instead it must be
                calculated for the projected concentration. Thus, total
                phosphorus loading  error is  (applying the variance
                operator, Var(x), to Equation 6):
                                                          Var((Pin)proj) = Var((Pin)red) + Var((Pin)pres)
                                                                (7)
                Since the present phosphorus loading was measured
                (not estimated from the literature, using, for example,
                export coefficients), the error in this loading estimate
                (Var((Pin)pres)) is effectively already part of the model
                (Equation 1) standard error (see Reckhow and Chapra,
                1983, for an explanation). Therefore, it is appropriate
                to assume that the additional loading error (for P") is
                Var((Pin)red), which is (0.005 mg/l)2. If the error in TW is
                assumed to be 0.008 yr, and Pin and TW are assumed to
                be  independent,  first order error analysis yields the
                following  prediction plus/minus the standard error:

                     P" = 0.058 mg/l  +/- 0.016 mg/l

                The model error (root PRMSE), which dominated the
                preceding calculation, was combined with the error
                contributions from Var (TW) and Var((Pin)red) in concen-
                tration (i.e., untransformed)  units; parameter errors
                were negligible and therefore ignored. The likelihood
                prediction analysis is complete with the addition  of
                the standard error in  the measured phosphorus con-
                centration:

                     P" = 0.058 +/- (0.0162 +  0.00572)0.5
                     P" = 0.058 mg/l  +/- 0.0170 mg/l

                  The prior (P') and the likelihood (P") predictions are
                pooled to yield a "posterior" (P*) prediction using the
                conjugate normal distribution approach presented in
                                            Quality of Life
  Water
  Clarity
 Fish:
 Type
  and
Quantity
            Navigation
             Effects
            Pathogen
          Concentrations
Toxic
Contaminants




|
'
Toxic
Contam-
atlc inant
ds: Levels
Particulates:
Reservoir
Fill-in
and
Treatment
Effects
(e.g., filter
clogging)
                                                    Cost
                                                     of
                                                 Wastewater
                                                  Treatment
                          Economic
                            Effect
                             on
                           Regional
                           Economy
                                                                              Regional
                                                                             Agricultural
                                                                                Profit
Reservoir
 Volume
Reduction
  Water
Treatment
  Costs
                                                 Overall
                                                  Cost
                                                   of
                                                 Clothes
                                                 Cleaning
                                                                                                  Clean
                                                                                                 Clothes?
Figure 5.
                                                   191

-------
LAKE AND RESERVOIR MANAGEMENT
Reckhow and Chapra (1983). Under this procedure, P*
is:
     P* = (1-k)P'+ kP"

where:


               Var(PO
                  (8)
     k    Var(PO + Var(P")

and the error in P* is:
                  (9)
     Var(P*) =
                    1
   1
                                        -1
                 Var(P')
Var(P")
                 (10)
Carolina. The arrows at the bottom of the figure point
to attributes that are intended to measure the success
(or failure) of any management program.
  The point contained in Figure 5 is that phosphorus
alone is not  a  meaningful attribute.  Decisionmakers
typically cannot assess the success of a management
plan on the basis of projected phosphorus concentra-
tion.  Therefore,  the modeler must  extend  his/her
model to include the meaningful attribute(s); the alter-
native is possibly a mental extrapolation on the part of
the decisionmaker from phosphorus  to a meaningful
attribute (like fish type and quantity). Since the scien-
tist/engineer is better able to make this extrapolation,
we would be wise to try to incorporate the meaningful
attributes  into the  scientific  analysis  when  this
analysis is used to aid in water quality management
planning.
For the Rhodhiss Lake application, this yields a pro-
jected phosphorus concentration ( + /-  the standard
error) of:

     P* =  0.0554 mg/l +/- 0.0061 mg/l

Note that, as is to be expected, the posterior predicted
concentration is between the prior and the  likelihood,
and  the uncertainty in the posterior is less than the
uncertainty in either the prior or the likelihood. Fur-
ther, by modeling the change in  concentration only,
we are able to reduce the impact of model error on the
final prediction error.


CONCLUDING COMMENTS

A model was developed and successfully evaluated
for  predicting  total  phosphorus  concentration  in
southeastern U.S.  lakes. This model was  compared
with other model  specifications,  indicating, among
other things, that the linear relation between input and
lake concentration is not entirely consistent  with
these southeastern U.S. lakes data. An application il-
lustrated   how  error  reduction methods  might  be
employed  with the model.
   As a final comment, Figure 5 is presented to under-
score the need to extend our modeling efforts beyond
phosphorus. Figure 5 is an objectives hierarchy. It is
intended to comprehensively identify planning objec-
tives for a reservoir eutrophication analysis in North
                        REFERENCES

                        Belsley, D.A., E. Kuh, and R.E.  Welsch. 1980.  Regression
                          Diagnostics. J. Wiley & Sons, Inc., New York.
                        Cook, R.D., and S. Weisberg. 1982. Residuals and Influence in
                          Regression. Chapman and Hall, New York.
                        Dillon, P.J., and F.H. Rigler. 1975. A simple method for predicting
                          the capacity of a lake for development  based on lake trophic
                          status. J. Fish. Res. Board Can. 32:1519-31.
                        Larsen,  DP.,  and  H.T. Mercier.  1976. Phosphorus  retention
                          capacity of lakes. J. Fish. Res. Board Can. 33:1742-50.
                        Learner, E.E. 1978. Specification Searches: Ad Hoc  Inference
                          with Nonexperimental Data. J. Wiley & Sons, Inc., New York
                        Ray, A., ed. 1982. SAS Users Guide. SAS Institute, Inc. Cary,
                          N.C.

                        Reckhow,  K.H., and S.C. Chapra. 1983. Engineering Approaches
                          for Lake Management. Vol. I: Data Analysis and  Empirical
                          Modeling. Ann Arbor Science, Woburn, Mass.
                        Reckhow,  K.H.,  and  J.T. Simpson. 1980. A procedure using
                          modeling and error analysis for the prediction of lake phos-
                          phorus concentration from land use information. Can. J. Fish.
                          Aquat. Sci. 37:1439-48.
                        U.S. Environmental Protection Agency. 1976. Various National
                          Eutrophication Survey Working Papers. Corvallis, Ore.
                        Vollenweider,  R.A. 1976. Advances in  defining critical loading
                          levels for phosphorus in lakes eutrophication. Mem. Inst. Ital.
                          Idrobiol. 33:53-83.
                                                   192

-------
PHYTOPLANKTON-NUTRIENT  RELATIONSHIPS IN  SOUTH CAROLINA
RESERVOIRS:  IMPLICATIONS  FOR MANAGEMENT STRATEGIES
JEFFREY PEARSE
South Carolina Department of Health and Environmental  Control
Columbia, South  Carolina


            ABSTRACT

            Using existing data selected from a  large government data base, an analysis of the factors
            governing algal standing crop in 40 South Carolina reservoirs revealed that chlorophyll a produc-
            tion was considerably lower than that reported for northern temperate lakes. This observation
            was attributed to the high levels of nonalgal turbidity frequently exhibited by the study reser-
            voirs.  The accuracy of eight literature models in predicting chlorophyll was examined and most
            were found to overestimate actual ambient values. To account for the various moderating fac-
            tors affecting chlorophyll production in turbid reservoirs, the use of multivariate models is sug-
            gested. The results of this  study indicate that local conditions should be taken into account
            when  selecting a predictive model for management purposes. One feasible option for govern-
            ment  agencies is to develop predictive models using readily available data.
INTRODUCTION

As population and industrial development pressures
increase, Government agencies are increasingly call-
ed upon to approve increases in nutrient loadings to
lakes  and reservoirs. To  efficiently  manage  these
water bodies, it is necessary to determine the effects
additional nutrients  will have upon  them. Several
methods are available,  but because  of time, man-
power, and cost  constraints,  the methods that  are
ultimately used must be simple,  rapid, and inexpen-
sive.
  The method of choice is usually a simple empirical
model  relating  ambient nutrient  levels to phyto-
plankton production. Numerous  models have been
developed during  the last decade,  most of which
predict chlorophyll a from ambient total  phosphorus
levels  (i.e.,  Dillon  and Rigler,   1974;  Jones  and
Bachmann, 1976; Carlson, 1977; Schindler et al. 1978).
Phosphorus is generally considered to  be the most im-
portant nutrient associated with eutrophication, while
chlorophyll a is considered to be the most reliable
measure of response to eutrophication (Taylor et  al.
1979). Unfortunately,  numerous factors can affect the
algal  response to nutrient loading, the result being
that each lake may respond  differently (Smith and
Shapiro, 1981). Therefore, lake planners and managers
must understand the limitations and applications of a
model before using it.
  Most of the models presented in the literature have
been developed using data from  naturally formed
temperate lakes with low to moderate phosphorus and
chlorophyll a  levels (Allan,  1980). The application of
these models to lakes with characteristics differing
from those in the  original data set may  be  ques-
tionable (Reckhow, 1981). Lakes in warmer regions
respond  differently  from natural  temperate  lakes
(Smith,  1982),  as  do artificial  lakes (Canfield and
Bachmann, 1981) and lakes with high  inorganic  tur-
bidities caused by  sediment inputs (Williams  et al.
1978; Taylor et al.  1979; Hern et al. 1981; Jones and
Lee, 1982).
  Reservoirs   in  South  Carolina are  not  well
represented in the development of literature models.
Moreover, they exhibit characteristics that may make
them incompatible with  the  applications  of  these
models. Publicly owned water bodies in the State are
typically shallow, run of the river impoundments with
short  retention  times,  exhibiting weak  thermal
stratification,  and  possessing  large  suspended
material inputs (Table 1). Consequently, they are very
frequently subject to high nonalgal turbidities.
  Williams et al. (1978) found that compared to lakes
in the northeastern United States, southeastern lakes
exhibited significantly less  chloiophyll a production
per unit phosphorus. Experience in South Carolina
                           Table 1.—Characteristics of South Carolina reservoirs.
Parameter
Surface area (km2)3
Mean depth (m)a
Hydraulic retention time (yr)b
Specific conductance (/xmohs/cm)
Total alkalinity (mg/l)
Total hardness (mg/l)
pH (standard units)
Turbidity (FTU)
Source Available Storet data, 1978-1982
aModified from Kimsey et al (1982)
bData from 14 major lakes (Nat Eutroph Surv , 1978)
Minimum
0.16
0.9
0.02
12.5
1
10
5.1
0.3



Median
2.0
4.2
0.2
70
15
11
7.1
3.8



Maximum
450
47.8
1.1
375
77
36
10.4
200



                                               193

-------
LAKE AND RESERVOIR MANAGEMENT

has verified this observation (S.C. Dep. Health Environ.
Control,  unpubl. data). In light of these  points, this
study was undertaken to examine, using available
data,  the  relationships between nutrients  and
phytoplankton  standing  crop  and  to  determine
whether a simple model could be applied to reservoirs
in South Carolina.

METHODS

The  South Carolina Department  of Health  and E.n-
vironmental Control, as the water pollution regulatory
agency in the State, samples a large number of water
quality parameters on a regular basis from a large not-
work of monitoring stations (S.C. Dep. Health Environ.
Control, 1983). I was interested in evaluating the utility
of this data base and the parameters routinely col-
lected for an analysis of nutrient-phytoplankton re a-
tionships. From this data base, 13 stations located in
nine impoundments were selected because of the
completeness of chlorophyll a and nutrient data. For
this analysis, data collected at these stations during a
5-year  period  (1978-82)  were used. Additional data
came from 93 stations in 40 publicly owned reservoirs
that  were sampled during 1980-81  as part of the South
Carolina Clean Lakes Classification Survey (Kimsey et
al. 1982).
  Parameters included  in  this  analysis  were
phaeophytin-corrected  chlorophyll a,  total  phos-
phorus, total nitrogen (calculated from summing total
Kjeldahl nitrogen and nitrate-nitrite nitrogen), and Sec-
chi disk  transparency. Growing season means (May
through  October collections  including some  ea'ly
November data for several Clean Lakes stations) were
calculated for each station  parameter, yielding 147
seasonal  means  from 101  different  stations.  (Five
sample locations were the same for both data bases,
so their data were pooled for 1981.)
                       Nutrient samples were collected  as surface grab
                    samples  for the  13 trend  stations and  as depth-
                    integrated composites for the Clean Lakes Classifica-
                    tion Survey sample sites. All chlorophyll a collections
                    were made from surface grab samples. The collection
                    frequency for the 13 monitoring network stations was
                    two to seven times (mean of five) during the growing
                    season for nutrient parameters and two to  five times
                    (mean   of three)  for  chlorophyll  and Secchi  disk
                    transparency. For the Clean Lakes study stations, col-
                    lections were made between two and three  times dur-
                    ing the growing season for all parameters.
                       After compilation, the  data  were partitioned into
                    nitrogen-  and phosphorus-limited data sets because
                    chlorophyll-phosphorus relationships may  be altered
                    in nitrogen-limited lakes (Jones and  Lee, 1982; Smith
                    and Shapiro, 1981). Nitrogen limitation was assumed
                    when the  total-nitrogen-to-total-phosphorus ratio was
                    less than  10. Parameter values were log  transformed
                    and subjected  1o a  stepwise multiple regression
                    analysis  to  examine the   relationship   between
                    variables.
                    RESULTS

                    The sample data set was composed of a wide variety
                    of stations covering every major lake and 26 minor
                    lakes in the State. Considerable variation was evident
                    for the levels of all parameters (Table 2). Most (52 per-
                    cent) of the stations were apparently nitrogen  limited
                    (total nitrogen:total  phosphorus ratio  less than 10),
                    predominately as a result of excess total phosphorus
                    rather than low nitrogen levels, a situation frequently
                    noted by Wiliams et al. (1978) in a study of 418 eastern
                    lakes.
                       Considerable variation  was evident  in the yield  of
                    chlorophyll a per unit  phosphorus in South Carolina
      Table 2.—Summary data for the parameters used in multivariate analyses. Data presented are mean growing
         season values and reported as mg/ms except for Secchi disk transparency which is reported in meters.
Data Set
All data
n = 147


Phosphorus limited
n = 70


Nitrogen limited
n = 77


Parameter
Chlorophyll a
Total phosphorus
Total nitrogen
Secchi disk transparency
Chlorophyll a
Total phosphorus
Total nitrogen
Secchi disk transparent
Chlorophyll a
Total phosphorus
Total nitrogen
Secchi disk transparency
Mean
11.8
119
973
1.5
10.8
57
997
1.5
12.7
176
951
1.5
Standard
Deviation
12.4
119
704
1.1
11.3
30
867
1.0
133
139
517
1 1
Range
1.0
20
190
0.3
1.6
20
240
0.3
1.0
30
190
0.4
-72.4
- 810
-6970
-6.5
-59.2
-210
-6970
- 4.4
- 72.4
-810
-2620
-6.5
       Table 3.—Summary statistics for regressions of log chlorophyll a on log total phosphorus (log TP), log total
              nitrogen (log TN), and log Secchi disk transpanency (log SD) for South Carolina reservoirs.
Data Restriction
Variable
None (all data)
    n = 147
Phosphorus limited
    n = 70
Nitrogen limited
    n = 77
 logTP
 logTN
 log SD

 logTP
 logTN
 log SD

 logTP
 logTN
 log SD
0.12
0.18
0.33

0.13
0.13
0.28

0.16
0.24
0.36
 20.6
314
 70.3

  98
 10.2
 270

 14.2
 23.2
 42.8
0.0001
0.0001
0.0001

0.01
0.01
0.0001

0.001
0.0001
0.0001
                                                 194

-------
                                                                     MODELING TECHNIQUES AND INNOVATIONS
reservoirs (Fig.  1). When  compared  to  the model
regression line reported by Dillon and Rigler (1974), it
is  apparent   that  chlorophyll  production  is  con-
siderably lower than that of the Canadian lakes includ-
ed in the  model. One would expect this and similar
models such as  Jones  and  Bachmann (1976)  to
overestimate chlorophyll  a  values in South Carolina.
  As a result of the scatter in the  chlorophyll-phos-
phorus relationship, it is not surprising that the regres-
sion analyses revealed that total phosphorus was not
the best predictor of chlorophyll a, accounting for only
12 percent of the variance  for all lakes combined
(Table 3). Of the parameters included in the regression
analyses,  Secchi disk transparency appeared to  be
the best single predictor  of chlorophyll a,  accounting
for 33 percent of the variance observed for all lakes.
When  considered separately,  both  the nitrogen- and
phosphorus-limited data sets exhibited similar predic-
tor accountabilities, with  the nitrogen-limited data set
capturing  slightly more overall variance.
  The regression line of  best fit for all lakes was:

     Log chlorophyll a =  0.14 log total phosphorus +
0.231  log  total   nitrogen  -  0.66  log Secchi  disk
transparency +  .003

     Where n  =  147 r2 = 0.37

A significant amount of variance was unaccounted for
by this equation. Although Secchi  disk transparency
was the best predictor of chorophyll a it did not cor-
relate strongly with nutrient levels. Specifically, for all
data combined,  Secchi transparency contributed 10
percent and  20 percent  of  the  variance for  total
phosphorus and total nitrogen, respectively.
  Because of the low amount of variance accounted
for by the  equation of best fit and because previous at-
tempts to apply literature  models to South Carolina
reservoirs have resulted in  overestimating chlorophyll
a levels (S.C. Dep. Health Environ. Control,  unpubl.
data), eight literature models were compared (Table 4).
The  prediction accuracy of  each model on the data set
was compared  by examining  the  standard  error of
estimate  indices (Table  5). The prediction equation
from this study  was included for comparative pur-
poses.
  Generally,  those models that included only  total
phosphorus  in the formula were  poor predictors of
chlorophyll a levels. These  models exhibited less error
when  applied to  the  phosphorus- rather than the
nitrogen-limited  data set. This outcome was expected
since  most of the models were developed from pre-
dominantly  phosphorus-limited lake  data.  No one
                       literature model possessed the  lowest error  rates
                       across both data sets, indicating that perhaps model
                       selection should take into account the nutrient limita-
                       tion  in  the  water body  under  study.  For the
                       phosphorus-limited data set, the model developed by
                       Williams et al. (1978) from 418 eastern lakes was most
                       accurate. Interestingly, this  model and  model  No. 5
                       were the only  models  tested  that  included South
                       Carolina reservoir data in their development.  For the
                       nitrogen-limited data  set, the  model  described by
                       Smith  (1982)  was  most  accurate. This  model  was
                       developed using 101  Florida lakes,  many of which
                       were nitrogen-limited.

                       DISCUSSION

                       Transparency, as reflected in Secchi disk values, is a
                       measure of  light penetration  and  is  controlled by
                       organic and  inorganic constituents in the water col-
                       umn, the most important being algal  biomass and in-
                       organic suspended  sediments (Brezonik, 1978). Light
                       attenuation by sediment particles is a frequent occur-
                       rence in South  Carolina reservoirs during periods of
                       runoff. Therefore, the observation that Secchi  disk
                       transparency was apparently the best single predictor
                               •  TN:TP>10

                               *  TN:TP<10
                                        / • I    9     *
                                                 100


                                       TOTAL  PHOSPHORUS
                                                         ng/m3)
                       Figure 1.—Relationship between chlorophyll a  and  total
                       phosphorus concentrations for 40 South Carolina reservoirs.
                       The line is from Dillon and Rigler (1974).
      Table 4.—A listing of regression formulas and their development data bases for the chlorophyll-nutrient models
                compared in the text. The dependent variable is log chlorophyll a (mg/m3) for all models.
  Model
Formula
Data Base
Source
    1        1.449 log(TP) - 1.136
    2        146log(TP) - 1.09

    3        3 27 log(log(TP)) + 0.542
    4        064log(TP) + 1.87

    5        OJOIog(TP) + 1.99

    6        0.249 log(TP) + 1.06 log(TN) - 2 49

    7        0.653 log(TP) + 0.548 log(TN) - 1.517
    8        0.374 log(TP) + 0.935 log(TN) % 1 2.488

    9        0.14 log(TP) + 0.231 log(TN) - 0.66 log(SD) + 0.003
 TP = total phosphorus concentration {mg/m3), TN - totat nitrogen concentration
 (mg/mj), SD = Secchi disk transparency (m)
                             Canadian lakes
                             N. temperate lakes
                             Canadian lakes
                             Eastern U.S. lakes
                             Eastern U.S. lakes
                             Florida lakes
                             N. temperate lakes
                             Florida lakes
                             S.C. lakes
                 Dillon & Rigler, 1974
                 Jones & Bachmann, 1976
                 Hickman,  1980
                 Williams et al. 1978
                 Williams et al. 1978
                 Canfield, 1983
                 Smith, 1982
                 Smith, 1982
                 This study
                                                  195

-------
LAKE AND RESERVOIR MANAGEMENT

of phytoplankton standing crop  was  not  an unex-
pected result. Brezonik (1978) observed a strong cor-
relation  between  Secchi  disk  transparency and
chlorophyll a in 55 colored Florida lakes, while Wright
and Soltero (1973) noted that the depth of the euphotic
zone was  the most significant parameter controlling
chlorophyll a in Yellowtail Reservoir.
  The  low variance in chlorophyll a levels explained
by total phosphorus levels may be partially attributed
to sediment-nutrient interactions. It is known that the
adsorption reactions between phosphorus  and sedi-
ment particles reduce the bioavailibility of nutrients 1o
algae  (Canfield and  Bachmann, 1981;  Hoyer and
Jones, 1983;  Williams et  al. 1978), often producing a
negative correlation between chlorophyll a  and total
phosphorus (Taylor et al.  1979).
  As mentioned previously, the chlorophyll  produc-
tion per unit  phosphorus in South Carolina  reservoirs
was found to be quite low in relation to the production
observed  in  other temperate lakes. Williams et  al.
(1978)  noted  that southeastern lakes produced con-
siderably  less chlorophyll than northeastern lakes and
attributed this to the higher levels of inorganic turbidi-
ty present in the southeastern  lakes. Waters high  in
nonalgal turbidities are known to support lower levels
of chlorophyll  a than nonturbid  lakes (Smith, 1982;
Hoyer and Jones, 1983).
   Other factors have been shown to affect the rela-
tionship  between chlorophyll  and  phosphorus   in
lakes.  These include variability of chlorophyll content
in algal cells (Nicholls and Dillon, 1978), macrophyte
abundance (Canfield, 1983), and  lake type (Canfield
and Bachmann, 1981). Nitrogen concentration, zco-
plankton  abundance, and flushing rate may also be
important factors. However, Hoyer and Jones (1963)
noted  that for a wide range of midwestern reservoirs,
significant variance could not  be attributed to these
three variables.
  Another important factor affecting the study results
may be the data set itself. Productive lakes, like many
of those found in the study data set, have widely fluc-
tuating summer chlorophyll a levels (S.C. Dep.  Health
Environ. Control, unpubl.  data; Prepas and Trew, 1963;
Allan,  1980), often with the maximum  chlorophyll level
being  three to four times the mean  (Canfield, 1983).
Therefore, the frequency of sampling used in this
study  may not be sufficient to fully capture the rela-
tionships  between variables, although Canfield (19S3)
determined  that  three  collections  during  a  1-year
period   were  sufficient  to  delineate the
nutrient-chlorophyll relationship in Florida lakes. The
slightly higher variance accounted for by the nitrogen-
limited data set in this study may reflect the sampling
frequency differences noted in the two data sources
used for this study.
  The nitrogen-limited data set had the lowest propor-
tion of Clean  Lakes data and therefore, the highest
sample frequency. Also, nutrient data collected from
the Clean Lakes stations were depth-integrated water
column  composites  that tend  to  overemphasize
nutrient Bevels during stratification periods as a result
of nutrient releases from  bottom sediments. Conse-
quently,  use of the Clean Lakes data base may help
explain part of the observed  low level of chlorophyll
production per unit phosphorus.  Further analysis of
the data sets and the use of other data is planned.
  Apparently, the selection of a chlorophyll-nutrient
model for use in South Carolina reservoirs should be
made  with  care, since the  error estimates  of  the
literature models tested varied so widely. As lake pro-
ductivity  increases,  control  of  algal  populations
becomes more complex with physical factors becom-
ing more important (Hickman, 1980).  It follows then,
that multivariate models may be necessary to predict
chlorophyll responses satisfactorily in South Carolina
reservoirs and other lakes subject to high inorganic
turbidities. A recently developed multivariate model
incorporating  phosphorus  and  inorganic suspended
solids (Hoyer  and Jones, 1983) could have  con-
siderable potential  in  such lakes.
  Because the phosphorus-chlorophyll relationship
in lakes subject to high nonalgal turbidities is affected
by many factors, applying  ambient  total phosphorus
limits such as those proposed by Dillon (1975), Larsen
and Mercier (1976), and the U.S. Environmental Protec-
tion Agency (1976)  to such lakes may  be unwise. If
these limits were stringently applied to South Carolina
reservoirs, significant  nutrient removal costs would be
incurred  without  corresponding  increases  in water
quality. At the recommended level of 20 mg/m3 of total
phosphorus, chlorophyll a summer means did not ex-
ceed 6.8 mg/m3 in this study. As suggested by Mancini
et al. (in  press), it  is recommended  that Government
agencies develop  acceptable nutrient  levels based
upon local conditions.
CONCLUSIONS

In  water bodies subject to high  nonalgal turbidities,
the relationship  between  chlorophyll  a  and nutrient
levels is not as clear as in other lakes. Consequently,
local conditions should be taken into consideration
before using literature models to predict chlorophyll
levels.  Although the results presented  here are not
conclusive,  I feel that it  is feasible for Government
agencies, especially on the State level, to develop
region-specific  predictive  models  based upon data
already available  to them. These data bases can in-
clude trend-monitoring data,  Clean  Lakes studies,
       Table 5.—Standard error of estimates for nine chlorophyll-phosphorus models calcualted for each data set.
Model
1
2
3
4
5
6
7
8
9
Independent
Variable
logfTP)
log(TP)
log(logfTP))
log(TP)
log(TP)
log(TP) + log(TN)
log(TP) + log(TN)
log(TP) + log(TN)
log(TP) + log(TN) + log (SD)
All Data
172.3
206.6
29.3
14.2
17.0
18.8
305
13.6
10.7
Phosphorus
Limited
27.9
33.6
15.2
10.5
10.7
21.2
18.6
144
99
Nitrogen
Limited
235.7
282.7
37.7
16.8
21.1
16.2
38.1
12.8
11.2
 TP = total phosphorus concentration (mg/m1); TN = total nitrogen concentration (mg/m3), SD = Secchi disk transpaiercy (m)
                                                 196

-------
                                                                           MODELING TECHNIQUES AND INNOVATIONS
consultant and  industry  reports, as  well as  other
miscellaneous studies.
   To make management decisions rapidly and ac-
curately, monitoring agencies should plan future lim-
nological studies with the intent of developing predic-
tive models. The emphasis of these studies should be
frequent summer sampling  when algal  biomass  is
most closely  related to nutrient  concentrations  and
when public interest is at its peak (Smith and Shapiro,
1981).
REFERENCES

Allan, R.J. 1980. The inadequacy of existing chlorophyll  al
  phosphorus  concentration  correlations  for  assessing
  remedial measures for hypertrophic lakes. Environ. Pollut.
  (Ser. B) 1:217-31.

Brezonik, P.L 1978. Effect of organic color and turbidity on
  Secchi disk transparency. J. Fish. Res. Board Can. 35:
  1410-16.
Carlson,  R.E. 1977. A trophic state index for lakes.  Limnol.
  Oceanogr. 22: 361-9.

Canfield, D.E., Jr. 1983. Prediction of chlorophyll a concen-
  trations in  Florida lakes: The importance of phosphorus
  and nitrogen. Water Res. Bull. 19: 255-62.
Canfield, D.E., Jr., and R.W. Bachmann. 1981. Prediction of
  total phosphorus concentrations, chlorophyll a, and Sec-
  chi depth in natural and artificial lakes. Can. J.  Fish.
  Aquat. Sci. 38: 414-23.
Dillon, P.J. 1975. The phosphorus budget of Cameron Lake,
  Ontario: The importance of flushing rate to the degree of
  eutrophy in lakes. Limnol. Oceanogr. 20:28-39.
Dillon,  P.J., and F.H. Rigler. 1974.  The  phosphorus-chloro-
  phyll relationship in lakes. Limnol. Oceanogr. 19:767-73.
Hern, S.C., V.W. Lambou, LR. Williams, and W.D.  Taylor.
  1981. Modifications of models predicting trophic state of
  lakes: Adjustment of models to account for the biological
  manifestations  of nutrients.  EPA-600/3-81-001. U.S. En-
  viron. Prot. Agency, Las Vegas, Nev.

Hickman, M.  1980. Phosphorus, chlorophyll and eutrophic
  lakes. Arch. Hydrobiol. 88: 137-45.
Hoyer.  M.V., and J.R. Jones.  1983. Factors affecting the
  relation between  phosphorus and chlorophyll a  in mid-
  western reservoirs. Can. J. Fish. Aquat. Sci. 40: 192-9.
Jones, R.J., and R.W. Bachmann. 1976.  Prediction of phos-
  phorus and  chlorophyll levels  in lakes. J. Water Pollut.
  Control Fed. 48: 2176-82.
Jones, R.A., and G.F. Lee. 1982. Recent advances in assess-
  ing impact of phosphorus loads on eutrophication-related
  water quality. Water Res. 16: 503-15.
Kimsey, C.D.  Jr., et al. 1982.  South Carolina  Clean Lakes
  Classification Survey. Tech.  Rep. No. 019-82. Bur. Water
  Pollut.  Control. S.C. Dep. Health Environ. Control,  Colum-
  bia.
 Larsen, D.P., and H.T. Mercier. 1976. Phosphorus retention
   capacity of lakes. J. Fish. Res. Board Can. 33: 1742-50.
 Mancini, J.L., P.A. Mangarella, G. Kaufman and E.G. Dnscoll.
   In  press.  Technical  Guidance  Manual  for  Performing
   Waste  Load Allocations. Book IV.  Lakes and Impound-
   ments. Chap. 2. Eutrophication. U.S. Environ. Prot. Agency.
   Washington, D.C.
 National Eutrophication Survey. 1978. A compendium of lake
   and reservoir data collected by the National Eutrophica-
   tion Survey in eastern,  north-central, and southeastern
   United  States. Working Pap. No. 475. U.S. Environ. Prot.
   Agency, Corvallis, Ore.
 Nicholls, K.H., and  P.J.  Dillon. 1978. An evaluation of phos-
   phorus chlorophyll-phytoplankton relationship for lakes.
   Int. Rev. ges. Hydrobiol. 63: 141-54.
 Prepas, E.E., and D.O. Trew. 1983. Evaluation of the phos-
   phorus-chlorophyll relationship for lakes off the Precam-
   brian Shield in western Canada. Can. J. Fish.  Aquat. Sci.
   40: 27-35.
 Reckhow, K.H. 1981. Lake Data Analysis and Nutrient Budget
   Modeling.  EPA-600/3-81-011. U.S.  Environ. Prot. Agency,
   Corvallis.  Ore.
 Schindler, D.W., E.J. Fee, and T. Ruszczynski. 1978. Phos-
   phorus input and  its consequences  for  phytoplankton
   standing crop and production in the Experimental Lakes
   Area and in similar lakes. J. Fish.  Res.  Board Can. 35:
   190-6.
 Smith, V.H. 1982. The  nitrogen and phosphorus dependence
   of algal biomass in lakes: An empirical  and  theoretical
   analysis. Limnol.  Oceanogr. 27: 1101-12.
 Smith, V.H., and J.  Shapiro. 1981. Chlorophyll-phosphorus
   relations  in  individual lakes.  Their importance  to  lake
   restoration strategies. Environ. Sci. Technol. 15: 444-51.
 South  Carolina  Department of Health and  Environmental
   Control. 1983. State of South Carolina monitoring strategy
   for fiscal year 1984. Tech. Rep. No. 021-83. S.C. Dep. Health
   Environ. Control, Columbia.

	Unpubl. data from 1975 to 1983.
Taylor, W.D., LR. Williams, S.C. Hern, and V.W. Lambou.
  1979. Phytoplankton Water  Quality Relationships in U.S.
  Lakes. Part VII: Comparison of some new and old indices
  and measurements of trophic state. EPA-600/3-74-079. U.S.
  Environ. Prot. Agency, Las Vegas,  Nev.
U.S. Environmental Protection  Agency. 1976. Quality Criteria
  for Water. Off. Water Hazard. Mater., Washington, D.C.
Williams, LR., V.W. Lambou, S.C.  Hern, and R.W. Thomas.
  1978. Relationship of productivity and problem conditions
  to ambient nutrients: National Eutrophication Survey fin-
  dings for 418 eastern lakes. EPA-600/3-78-002. U.S. Environ.
  Prot. Agency, Las Vegas, Nev.

Wright, J.C., and R.A. Soltero. 1973. Limnology of Yellow-
  tail  Reservoir and the Bighorn River. EPA-R3-73-002. U.S.
  Environ. Prot. Agency, Washington, D.C.
                                                      197

-------
 RELATIONSHIPS BETWEEN  SUSPENDED SOLIDS,
 TURBIDITY, LIGHT ATTENUATION, AND ALGAL  PRODUCTIVITY
RUSS BROWN
Tennessee Valley Authority
Morris,  Tennessee
            ABSTRACT

            The effects on algal productivity of changes in light availability because of suspended materials entering
            a lake from storm runoff or turbidity generated by wind mixing of sediments are often discussed. The
            relationships between suspended solids, turbid ty, Secchi disk depth, light attenuation, and the resulting
            photosynthetic response of algae are not well documented. Data from several TVA reservoirs indicate
            that variations in the available light due to turbidity of the water are large and must be described in
            more detail before accurate predictions of algal dynamics and macrophyte growth can be achieved.
            Basic relationships between the various measurements of turbidity, suspended solids, Secchi disk
            depth, and light attenuation are given. This allows frequent measurements of turbidity or Secchi disk
            depth to be combined  with daily solar radiation and less frequent chlorophyll and C14 productivity
            data to yield a detailed estimate of algal procuctivity within a reservoir or lake throughout the year.
            For long lakes and many reservoirs, significant longitudinal gradients in suspended solids will result
            as the inflowing materials settle and the water column clears. The importance of this gradient of available
            light for modeling algal dynamics is demonsTated with data from several TVA  reservoirs.
INTRODUCTION

Productivity  of phytoplankton (or macrophytes) arid
subsequent algal growth is dependent on several envi-
ronmental factors as well as algal physiology. This
paper  shows  that  by accurately  modeling  these
physical  factors, the simulation  of algal dynamics
within  a  lake or reservoir is significantly improved.
Photosynthesis can be described as the conversion of
light energy into algal  biomass with  accompanying
uptake of nutrients and release of dissolved oxygen
and dissolved organics. The dominant environmental
influence  is the available photosynthetically active
radiation (PAR). Temperature governs the rate of bio-
chemical  processes and  temperature gradients  in-
fluence mixing characteristics of the surface layer.
Near-surface mixing modifies the light environment of
algae.  The PAR  is also affected by absorption arid
scattering of light which are controlled by suspended
solids and by algae.
  A solid basis for understanding algal productivity
and growth can be achieved by properly describing the
environmental factors of temperature,  turbidity, light,
and  near-surface mixing.  It  is not  sufficient  ':o
describe the limnologic environment with a seasonal
time scale,  because many shorter term  variations
have significant  effects on algal dynamics (Harris,
1980). Neither will reservoir or lake average conditions
be adequate. Temperature patterns within lakes and
reservoirs have been well described,  including the
near-surface wind  mixing processes  (Gulliver arid
Stefan, 1982; Imberger  et al. 1978; Ford and Stefan,
1980; and Harleman, 1982).
  Descriptions of the light environment are much less
developed, particularly with regard to variations caus-
ed by turbidity events in reservoirs and lakes (Ritchie
et al.  1976).  Storm  inflows produce high turbidity in
lakes and reservoirs (Johnson et al. 1981). Rivers may
produce high turbidity  in  the  upstream portion of a
lake or reservoir (Stewart and Martin, 1982). Relation-
ships  between suspended solids, turbidity, and light
attenuation  will  be described, based on  data from
several TVA reservoirs. These relationships will then
be used to determine PAR availability for algal produc-
tivity.
SUSPENDED SOLIDS AND LIGHT
ATTENUATION MEASUREMENTS

Suspended solids concentrations  can be measured
directly by filtering a water sample and drying and
weighing the residue. The size distribution of the par-
ticles can also be determined with various methods.
These determinations require considerable laboratory
time  and expense.  Less  expensive  indirect
measurements of suspended solids can also be made.
Turbidity (light  scattering) is  the  most common  in-
direct measurement of suspended solids. Some trans-
missivity meters (light extinction) have been used in
the field  to  gather  depth  profiles  of  suspended
materials (Stewart and Martin, 1982). Secchi disk dep-
ths and light  penetration devices  are used to deter-
mine near-surface light attenuation. Data on suspend-
ed materials and light attenuation often have not been
combined to obtain the relationship between suspend-
ed solids and light attenuation (Stefan et al. 1983).
SUSPENDED SOLIDS AND TURBIDITY

Frequent turbidity measurements can provide data on
the temporal and spatial variations in suspended
solids concentrations. Water supply treatment plants
are a potential source of this information that have not
been frequently used. Turbidity and suspended solids
concentrations  must be related empirically for  each
field  situation,  because  turbidity  is  an  optical
parameter that  varies with particle size, shape, and
the spectral absorption  of the particles (Duchrow and
Everhart, 1971; McCluney, 1975). The relationship bet-
ween turbidity and suspended solids may remain con-
                                                198

-------
                                                                   MODELING TECHNIQUES AND INNOVATIONS
stant for a particular field situation and variations in
turbidity then provide a convenient and fairly accurate
record  of suspended solids concentrations (Truhlar,
1976).
   Figure 1 shows the relationship between turbidity
and suspended solids obtained from three TVA reser-
voirs. Data from Fort Loudoun Reservoir (near Knox-
ville, Tenn.) during a 1971 seasonal survey include a
wide range of suspended solids concentrations caus-
ed by high inflow periods and clearing in the down-
stream portion of the reservoir. The ratio of turbidity to
suspended solids varied from 0.5 to 1.5. Data from the
TVA Biothermal Research Facility reflect variations in
turbidity of water pumped  from  Wheeler Reservoir
(near Huntsville,  Ala.). Travel time through the 1 m
deep channels is approximately 12 hours, and con-
siderable settling of suspended solids occurs. The in-
flow and outflow samples are easily distinguished.
Inflowing-turbidity-to-suspended-solids  ratio  varied
from 0.5 to 1.0 and the outflowing ratio was about 1.0.
Data from an intensive storm inflow survey of Norman-
dy Reservoir (near Nashville, Tenn.) reveal a fairly con-
sistent  relationship between turbidity and suspended
solids.  Turbidity was about  twice  the  suspended
solids  concentration. In  all  cases the ratio between
turbidity and suspended solids appears  to fall  bet-
ween about 0.5  and  2.0 for turbidities less than 100
NTU.
    100


 = 80
 t-
 z
  . 60
    40
 ffi
 IT

 ^  20
                                 III TURB-SS RATIO
      0    20   40    60   80   IOO  I20   I40   I60
                 SUSPENDED SOLIDS, mg/l

      A FT LOUDOUN RESERVOIR, 197!
     0      25     50      "0     25     50     75
      TURBIDITY, NTU                SS, mg/l
      B. BIOTHERMAL, 1979     C. NORMANDY RESERVOIR
                              STORM INFLOW SURVEY,
                              1983

Figure 1.—Relationship  between turbidity and suspended
solids concentration.
TURBIDITY AND SECCHI DEPTH

The relationship between turbidity and light penetra-
tion in natural waters has not been  well established
despite the importance of light for aquatic productivi-
ty and the relative ease of measuring turbidity and
light penetration with a Secchi disk. Chandler (1942)
measured the fluctuations of available light in western
Lake Erie due to changes in water transparency (tur-
bidity). He measured turbidity, suspended solids, Sec-
chi depth, and  light penetration at  weekly intervals
and noted large changes in turbidity  following storms
due to wind mixing of the sediments. Figure 2a shows
Secchi depth inversely related to turbidity in Lake Erie:

               Secchi = C/Turbidity

where C is a constant between 15 and 25.

Data from the TVA Biothermal facility show a similar
relationship, with C between 5 and 10 (Figure 2b).
Similar relationships have been found for Fort  Lou-
doun and Normandy Reservoirs (Fig. 2c and 2d), with C
also between 5 and 10.

LIGHT PENETRATION AND SECCHI  DEPTH

Secchi depth is a simple measure of light penetration,
but the  relationship  between  Secchi  depth  and
percentage  of  light penetration has  not been  well
established. The  basic relationship is obscured by
problems  of inconsistent  observation and differing
measurement devices used to determine light penetra-
tion (Tyler, 1968). An example of the spatial variation in
light penetration  and  the  relation  between Secchi
depth and light penetration is seen in data from Lake
Huron-Saginaw Bay   collected  by  Beeton (1958).
Figure 3a shows that average summer Secchi depth
varied from 12 m in the lake to 1  m in the shallow por-
tion of the bay. There was variability in the Secchi
depth-light relationship (Fig. 3b), but the Secchi depth
corresponded to 10 or 20 percent of surface light over
a wide range of Secchi depths (1 to 15 m). The gradient
of available  light within Saginaw Bay was significant
and either Secchi depth or light penetration measure-
ments provided this information.

LIGHT EXTINCTION COEFFICIENT

The  various measures  of light  attenuation  and
suspended solids or turbidity data can be  related to
the light  extinction  coefficient,  kT. Light attenuation
can be described as an exponential decay,
                                                    where l(z) is light at distance z below reference depth
                                                           I0 is light at reference depth.

                                                    This relationship has been found adequate to describe
                                                    PAR light levels once surface effects  become small
                                                    (below 50 percent surface light  depth). Secchi depth
                                                    measurements will always include various surface ef-
                                                    fects, but light  penetration measurements can avoid
                                                    this if they are  analyzed from slightly  below the sur-
                                                    face. The extinction coefficient can be estimated us-
                                                    ing light penetration  measurements or light trans-
                                                    mission data as
                                                          kT =
                                                 199

-------
LAKE AND RESERVOIR MANAGEMENT
Figure 4a shows  the  relationship between  turbidit/
and extinction coefficient at TVA's Biothermal facility.
There remains  quite a lot of variation but the extinc-
tion coefficient increases with turbidity.

       kT = C  • Turbidity

with C between 0.15 and 0.5.
                                                         The depth of 1 percent light penetration is often us-
                                                       ed to define the euphotic zone. The 1 percent depth
                                                       can be related to the extinction coefficient as
                                                              1% light depth  =
                                                                                -1n(.01)     4.6  .
  Stefan et al. (1983) studied very turbid Lake Chicct
along the Mississippi in Arkansas and found that kT -
2.0 + .05 • SS. The large intercept was similar to that
found at the TVA Biothermal facility. In both cases;,
this may be the result of not having turbidities of less
than 10 NTU. Figure 4b shows the Lake Erie data, with
C between 0.05 and  0.10.  Similar relationships were
found for Fort Loudoun and Normandy data (Fig. 4c
and 4d) with C about 0.10. The variation in the relation-
ship is not unexpected because  the attenuation  cf
light is an extremely complex process involving ab-
sorption, scattering, reflection, and refraction. All the
data indicate a nonzero kT at low turbidity.
  The relationship between Secchi depth and extinc-
tion coefficient depends on the level of surface  light
penetration at which the observer can still distinguish
the disk (Williams, 1981). Since this critical light  level
is variable, the relationship also varies, as
       Secchi depth =  —
                        kj

where I/I0 is the percentage light penetration at the
  Secchi depth. For I/I0 of 10 percent, Secchi depth is
  2.3/kT; if I/I0 is 20 percent, Secchi depth is 1.6/kT; and
  if I/I0 is 30 percent, Secchi depth is 1.2/kT.
Several examples of this relationship  are shown in
Figure 5.
                       20   40    60   80  IOO
                           TURBIDITY, NTU
                   A LAKE ERIE, 1940
                                                120
                                                          A. AVERAGE SUMMER SECCHI DEPTH IN SAGINA BAYIBEETON, 1998)

                                                           25|	
                                                         O
                                                         - 20
                                                         vt
                                                         u.
                                                         O
                                                         Ul 10
                                                         O
                                                                            «      *      12
                                                                             SECCHI  DEPTH, m
                                                       Figure 3.—Variation in light penetration in Sagina Bay.
                                                                i o
                                                             E

                                                             i"
                                                             UJ
                                                             o
                                                             _ 50
                                                             O
                                                             o
                                                                25
                                                                               • INFLOW
                                                                               O OUTFLOW
                                                                                  10/TURB
                                                                  0             25
                                                                         TURBIDITY, NTU
                                                                  B BIOTHERMAL, 1979
                                                                                             50
         20i
         I 5
       £; i o
       o
       5  5
       U
                            -10/TURB
                                            5/TURU
           0          5         10         15
                         TURBIDITY, NTU
            C FT LOUDOUN,  1982
                                                    20
                                                                     •5/TURB

                                                                     -10/TURB
                                                               0     20     40     60    80
                                                                         TURBIDITY,  NTU
                                                               D NORMANDY, 1983
                                                                                                 100
Figure 2.—Relationship between Secchi depth and turbidity.
                                                   200

-------
                                                                   MODELING TECHNIQUES AND INNOVATIONS
  Secchi depth can then be related to the 1 percent
light depth. If Secchi depth corresponds to the 10 per-
cent light level, then 1 percent light depth = 2.0»Sec-
chi  depth. If the Secchi depth is at 20 percent light
depth, then 1 percent light depth =  2.9»Secchi depth,
and if the Secchi depth equals the 30 percent light
depth, then 1 percent light depth =  3.8»Secchi depth.
Defining the euphotic zone as three or four times the
Secchi depth  is  a good estimate,  if actual  light
penetration data are not available.
APPLICATION OF TURBIDITY-LIGHT
RELATIONSHIPS FOR ESTIMATING ALGAL
PRODUCTIVITY POTENTIAL

Algal productivity  estimates cannot  be made from
available PAR calculations without also knowing the
algal biomass,  light adaptation or productivity-light
relationshp (P-l curve), and availability of nutrients.
But data on water clarity (turbidity, Secchi depth, or
light  penetration)  should  be  used along  with
temperature measurements to describe the physical
habitat of the algal assemblage. Without this detailed
characterization of the physical environment, algal
response cannot be accurately analyzed. Variations in
light  penetration  will  not  always be significant; in-
deed, many useful algal evaluations utilize only an
average summer  euphotic zone depth. But in other
cases variability  in light penetration  is significant.
Data should be collected to document the variations
or lack of variability. Viewed another way, there is no
sense in formulating a detailed description of near-
surface wind mixing and temperature structure if the
equally  important light  fluctuations  are not con-
sidered  in corresponding detail. Two examples will il-
lustrate these ideas.
FORT LOUDOUN RESERVOIR-1971

Daily turbidity measurements were obtained from the
Knoxville  water  treatment  plant  located  at  the
upstream end of Fort Loudoun Reservoir. Solar radia-
tion data were also available. These two time series,
which represent the two sources of temporal PAR
variation in the upstream portion of Fort Loudoun, are
shown in Figure 6. The solar data have a seasonal pat-
tern  as well  as short-term fluctuations caused by
variable cloud cover.  The turbidity ranged between
about 10 and 50 NTU, with a few short episodes of tur-
bidity between 50 and  100 NTU. Variations in turbidity
were not as rapid as the solar fluctuations since the
turbidity variations are associated with hydrological
events following storm runoff. Turbidity can be used
to estimate the light extinction coefficient by

      Extinction Coefficient (m-1)  = 0.05»Turbidity

A  low coefficient was chosen to avoid exaggerating
the effect of  turbidity (see Fig.  6). Turbidity fluctua-
tions between 10 and 50 NTU represent a variation in
the extinction coefficient between 0.5 and 2.5  m-1.
This produces a large variation in the euphotic zone,
defined  as the 1  percent light penetration depth, with
a range from 1.8m to 9.2m. For the 1971 Fort Loudoun
turbidity data, the resulting pattern of euphotic zone
depth is shown in Figure 6b. Several windows of high
PAR were apparent during low turbidity periods. Since
algal response may be triggered by periods of high
light availability or by  periods of low PAR, light varia-
tions must be accurately described.
   The pattern of available PAR  resulting from both
surface light  fluctuations and turbidity variations is
shown in Figure 7a. Daily average  PAR is shown  at
depths of 1 m, 3 m, and 5 m. Available light is severely
£  6

   5

   4

   3
         X
           2.5
           2.0
        UJ
        o
           1.5
        UJ
        o
        o
        t-
         X
         Ul
            .5
                            -0.5TURB
                        10         20         30
                                TURBIDITY NTU
                BIOTHERMAL FACILITY
                        0.20 TURB
                                                      40
       5.0
     ^
     ~ 4.0
     uJ
     UJ
     030


     020

     o
     Z I.O
     t-
     X
     UJ o
             0         5         10         15
                            TURBIDITY  NTU
              C. FT. LOUDOUN  RESERVOIR
                                                     20
                                                                            0.08 TURB
                                                                                   0.04 TURB
              20   40   60   80  100  120
                   TURBIDITY  NTU
          B. LAKE ERIE
      -I
      —_ 5
      u_'
      UJ4

      O

      o
      P 2
      (J
                                                                        -.20 TURB
                                                                                 IOTURB
                                                                                  .05 TURB
          0     20     40     60

                   TURBIDITY NTU
          D. NORMANDY RESERVOIR
                                                                                        80
Figure 4.—Relationship between extinction coefficient and turbidity.
                                                 201

-------
LAKE AND RESERVOIR MANAGEMENT
 restricted in the winter by low ambient PAR and high
 turbidity. Available light reaches 5 m only during a few
 periods of low turbidity. For comparison, light levels
 calculated with a constant extinction coefficient of 1.0
 m - 1, corresponding to an average turbidity of 20 NTU,
 are shown in Figure 7b. There are no distinct periods
 of high available light and available PAR does not ex-
 tend below the 3 m depth. Simulation of this seasonal
 envelope of PAR, with no variations due to turbidity
 changes, cannot provide an adequate basis for detail-
 ed algal  population predictions.
   Potential productivity of algae can be calculated for
 this pattern of available PAR if the growth response of
 the algae is described as a function of PAR (P-l curve).
 This type of calculation has been made by Fee (1980)
 and Field and  Effler (1983). A simple relationship will
 be used  here. At low light  levels, productivity per unit
 of  chlorophyll  is assumed to be a linear function  of
 PAR.  Above the saturation light  level, l«, the produc-
 tivity  is a constant. The saturating light level is taken
 to be 100 nEinsteins/m2sec (Harris, 1973) and the max-
 imum productivity is between 5 and 15 mg-02/mg«chlii-
 hr (chla =  chlorophyll a) (Jones, 1978). Taking a value
 of 10  and assuming a constant chlorophyll concentra-
 tion of 1 mg/m3 with a 12-hour constant light period,
 the areal  productivity (per unit chlorophyll)  corres-
 ponding to the available PAR patterns shown in Figure
 7 has been estimated. Light levels at 1 m  depth are
 usually above  saturation and largely unused by the
 algae. During  periods of low turbidity, the depth  to
 which saturating light intensity penetrates  increases
 and the  corresponding areal productivity  increases
 significantly as shown in Figure 8a. Calculations with
 a constant extinction coefficient show a more cons-
 tant productivity since the depth of saturating  light
 penetration remains nearly constant (Fig. 8b). There is
300
                                                SURFACE LIGHT
                                                (Kcal/m2hr)
                    >

                    Q
                    m
                    K
                    3

                    O
                    z
                    <
                      250 -
200 -
150 -
                      100 -
                         0   30  60 90  120 150 180 210 240 270 310 330  360
                                          DAY 1971

                         A  SOLAR AND TURBIDITY DATA
                    E  l2

                    I  10

                    t  8
                    *

                       6
                    a.  2

                    Ul  n
                        1% LIGHT DEPTH
                        0  30 60 90  120  150 180  210 240 270 3OO 330 360
                                          DAY 1971

                        B CALCULATED FT.  LOUDOUN EUPHOTIC  DEPTH
                  Figure 6.—Variation in Ft.  Loudoun euphotic zone depth
                  caused by solar and turbidity fluctuations.
       UJ

       O
       u.
       U.
       UJ
       O
       O
         I0i-
                 I/SECCHI
                    I
                            _l_
_L
                                                       25
                                                       2 0
               UJ I  5
               O
               O
                                                       1.0
                                                    O

                                                    ±  5
                                                    X
                                                    UJ
                   .26       .5        75

                  SECCHI DISK DEPTH, m

             BIOTHERMAL FACILITY
                                              1.0
                                   8
                                   O
                                   X
                                   UJ
                                          I/SECCH!
                                                                  -I/SECCHI
                                                                  I
                    0        5        10       15      2.0

                                  SECCHI  DEPTH, m
                    B. FT. LOUDOUN
                                                                                                  2.5
                                                  I
                                                             -t
                                      01234
                                              SECCHI  DEPTH, m
                                      C  LAKE ERIE


Figure 5.—Relationship between extinction coefficient anc Secchi depth.
                                                  202

-------
                                                                     MODELING TECHNIQUES AND INNOVATIONS
little  information  about environmental  fluctuations
that might control algal productivity patterns in the
model based on a constant light extinction. Yet this
approach is currently  used  in  many  limnological
models.
   Several important factors have been left out of this
example. Chlorophyll concentrations will not  remain
constant. As light conditions improve, growth of the
algae will  occur  and  the  increasing  levels  of
chlorophyll  will begin to have an important effect on
the light extinction coefficient. Several workers have
found the light  extinction coefficient to be a linear
function of chlorophyll a concentration (Megard, et al.
1980; Atlas  and Bannister, 1980).
kT = k
                  kc»chla
where kw is due to inorganic factors (turbidity) and
       kc is approximately .015 m-1/mg-chla/m3

This relationship provides an important feedback on
light attenuation by the algae in relatively clear water.
However, for chlorophyll a concentrations of less than
30 mg/m3, the effect on light extinction is less than 0.5
m-1, which is  a  small change relative to changes
caused by turbidity (equivalent to a change of 10 NTU)
(Lorenzen, 1980).
  A second important factor to consider is that there
will be settling of the suspended materials in the
reservoir or lake that will produce longitudinal  gra-
dients in turbidity  and available light.  Relationships
between velocity,  turbulence, and suspended solids
settling rates  have not been well developed for reser-
   I200i-
    900
 U  600
 a.
 §  300
voirs and this should be further investigated. Signifi-
cant longitudinal gradients are often observed (Ken-
nedy et al. 1982; Iwasa and Matsuo, 1981; Kimmel,
1981).
  A third important factor is the effect of surface wind
mixing on the light environment of algae. Stefan et al.
(1976) and Gulliver and Stefan (1982) have shown the
general effect of light limitation due to increasing the
mixed depth. They assume that all algae in the mixed
layer will  be exposed to an  average PAR that ap-
proaches lo/Dmix- Actually the light exposure of algae
circulating  in the mixed layer is more  complex and
may approximate  langmuir circulation  cells  (see re-
cent review by Leibovich, 1983). Response of algae to
these light fluctuations from mixing patterns deserves
further work. The possibility that the algal population
can adapt both  to prevailing light levels and to par-
ticular regimes of fluctuation should be investigated.
As these additional factors are studied  and included
in a temperature-mixing-light description of the algal
habitat, analysis of algal dynamics will become much
less  speculative and more  directly linked  with en-
vironmental  conditions.
                                               NORMANDY RESERVOIR-1982

                                               A water treatment plant began operating in Normandy
                                               Reservoir, Tenn., and provided  a full year of turbidity
                                               data in 1982. Intake location is at a depth of approx-

i
i
(M
2
i
i
(9
2
H
K
^
0
o
cc
a.
1000-
900

80O
700
600
500-
400
300
200

100

      0  30  60  90 120  150 180 210 240270300330 360
                        DAY 1971

       A AVAILABLE PAR WITH VARIABLE TURBIDITY
         (Kr = 05 TURBIDITY)
                                                     0   30  60 90  120 150 180 210 240 270 300 330 360
                                                                       DAY  1971

                                                     A  PRODUCTIVITY POTENTIAL WITH VARIABLE TURBIDITY
                                                        lOOOi-
       0  30 60 90  120 150  180 210 240 270 310 330 360
                         DAY 1971

       B AVAILABLE PAR WITH CONSTANT TURBIDITY
         (20 NTU)
Figure 7.—Calculated PAR in Ft. Loudoun Reservoir.
                                                        30  60  90  120 150 ISO 210 240 270 300 330 360
                                                                       DAY 1971

                                                     B PRODUCTIVITY POTENTIAL WITH CONSTANT TURBIDITY

                                              Figure 8.—Potential productivity in  Ft. Loudoun Reservoir,
                                              per unit chlorophyll (mg/m3).
                                                  203

-------
 LAKE AND RESERVOIR MANAGEMENT
 imately 5 m. Figure 9a shows that elevated winter tur-
 bidities (January through March) decreased to about 3
 to 5 NTU until December. In this case, estimation of a
 single light extinction for the summer period might be
 justified. But  this  should  not be  assumed  without
 measurements, since the  possibility  of spring and
 summer storm inflows always exists.  Effect of algal
 biomass on the extinction coefficient must be caro-
 fully considered whenever turbidities remain this low
 (<5  NTU) since algae may  become the  dominant in-
 fluence on light   attenuation. Some  limited  light
 measurements were  made in 1982 and are shown
 along with temperatures and the depth of the surface
 mixed  layer in Figure 9b. Seasonally, the  1  percent
 light penetration depth increased from about 4 m  in
 early summer to almost 10 m by the end of September,
 suggesting a decrease in turbidity and algae. Surface
 mixed depth remained between 3 m and 5 m. This pro-
 duced  a region in the metalimnion with  moderate
 temperature (20 to  25°C), sufficient light (1  to 10 per-
 cent surface),  and without any light fluctuation caus-
 ed by mixing. A distinct, stable habitat zone resulted,
 which may have allowed algal adaptation and efficient
 utilization of light in this region. This figure illustrates
 the goal  of   the  temperature-mixing-light  analysis
 recommended  in  this  paper:  a description of the
 physical  environmental  factors  that  is sufficiently
 detailed to support investigations of algal adaptation
 and response  to these fluctuating conditions.
 SUMMARY

 A solid basis for understanding algal productivity and
 growth can be achieved by properly describing the en-
        I25
     t  75
     O
     m
     5  50
        25
    890

    880

    870

    860

 -  850

 C  840

 ieso
 5  820
 >
 Uj  810

 U  800
    790
          0  30  60 90  120 150 180 210 240 270 3OO 330 360
                           1982
          A.  NORMANDY RESERVOIR TURBIDITY
LIGHT DEPTH
      0  30  60 90  120  150 180 2IO 240 270 3OO 33O 360
                           1982
       B TEMPERATURE, MIXED DEPTH, EUPHOTIC ZONE



Figure 9.—Variation in turbidity, light penetration and mixed
depth in Normandy Reservoir, 1982.
                                        vironmental  factors of temperature,  light, and near-
                                        surface mixing. Gradients of suspended solids have a
                                        strong  influence on  the available light  for  photo-
                                        synthesis. Turbidity can  be used as  a convenient
                                        measure  of  suspended solids  once the relationship
                                        between  turbidity and suspended solids has been
                                        determined for a particular water body.
                                           Several relationships between  turbidity and light
                                        parameters have been estimated from data for several
                                        reservoirs. Secchi depth is inversely related to turbidi-
                                        ty. The 1  percent light level is three or four times the
                                        Secchi depth. The light  attenuation coefficient  is
                                        directly related to turbidity, so that changes in turbidi-
                                        ty have a strong influence on light availability.
                                           Data from Ft. Loudoun Reservoir indicate  that in-
                                        flowing turbidity can create periods of low and high
                                        light attenuation  that significantly  influence the
                                        seasonal light pattern. Daily turbidity measurements
                                        significantly improve the  description of the light en-
                                        vironment for algal productivity simulations. Simula-
                                        tions of the reservoir productivity should include settl-
                                        ing of turbidity, absorption of light by algae, and wind
                                        mixing.
                                           Data from Normandy Reservoir indicate that turbidi-
                                        ty variations are  not always  significant, but  this
                                        should be determined from data and not assumed. At
                                        low  turbidity levels, the euphotic zone may still vary
                                        significantly  relative  to  surface  mixing depth  and
                                        create a zone of stable low light which may  lead to
                                        algal  adaptation  and   high   productivity.   Direct
                                        measures of light attenuation are preferred at low tur-
                                        bidities.
REFERENCES

Atlas, D., and T.T. Bannister. 1980. Dependence of mean
  spectral extinction coefficient of phytoplankton on depth,
  water color and  species. Limnol. Oceanogr. 25:157-9.
Beeton,  A.M. 1958. Relationship between Secchi disk read-
  ings and light penetration in Lake Huron. Trans. Am. Fish.
  Soc. 88:73-9.

Chandler, D.C. 1942. Limnological studies of western Lake
  Erie. II. Light penetration and its relation to turbidity.
  Ecology 23(1):41-52.
Duchrow, R.M., and W.H. Everhart.  1971. Turbidity measure-
  ment. Trans. Am. Fish. Soc. 100:682-90.

Fee, E.J. 1980.  Important  factors for estimating  annual
  phytoplankton production in the Experimental Lakes Area.
  Can. J. Fish. Aquat. Sci. 37:513-22.
Field, S.D., and S.W. Effler. 1983. Light-productivity model for
  Onondaga Lake, New York. Am. Soc. Civil Eng.  J. Environ.
  Eng. 109(4):830-44.

Ford, D.E., and H.G. Stefan. 1980. Thermal predictions using
  integral energy  model. Am. Soc. Civil  Eng. J. Hydraul.
  106(1):39-55.

Gulliver, J.S., and  H.G. Stefan. 1982.  Lake phytoplankton
  model with destratification. Am. Soc. Civil Eng.  J. Environ.
  Eng. 108(5):864-81.

Harleman, D.R.F.  I982. Hydrothermal analysis of  lakes
  and  reservoirs. Am.  Soc.  Civil   Eng.   J.   Hydraul.
  108(3):302-25.

Harris, G.P.  1980.  Temporal and spatial  scales in phyto-
  plankton  ecology: mechanisms, methods,  models, and
  management. Can. J. Fish Aquat. Sci. 37:877-900.
Harris, G.P., and J.N.A. Lott. 1973. Light intensity  and photo-
  synthetic  rates in phytoplankton. J. Fish. Res. Board Can.
  30:1771-8.
Imberger, J., J.  Patterson,  B.  Herbet, and  I. Loh.  1978.
  Dynamics of reservoir of medium size. Am. Soc. Civil Eng.
  J. Hydraul. 104(5):725-43.
                                                   204

-------
Iwasa, Y., and N. Matsuo. 1981. Estimation of turbidity in
  reservoirs. Pages 25-34 in IAHR 19th Congress.
Johnson, M.C., D.E.  Ford, E.M. Buchak,  and J.E. Edinger.
  1981. Analyzing storm  event data from DeGray  Lake,
  Arkansas, using LARM. Am. Soc. Civil Eng. 1981 Conven-
  tion Exposition, St. Louis.
Jones, R.J. 1978. Adaptations to fluctuating irradiance by
  natural  phytoplankton communities. Limnol. Oceanogr.
  23(5):920-6.
Kennedy, R.H., K.W.  Thornton,  and  R.C. Gunkel. 1982.  The
  establishment of  water quality gradients in reservoirs.
  Can. Water Resour. J. 7(1):71-87.
Kimmel, B.L 1981. Land-Water Interactions: Effects of Intro-
  duced Nutrients and Soil Particles on Reservoir Productivi-
  ty, Okla. Water Resour. Res. Inst.
Leibovich, S. 1983. The form and dynamics of langmuir cir-
  culations. Annu. Rev. Fluid Mech. 15:391-427.
Lorenzen, M.W.  1980. Use of  chlorophyll-Secchi disk rela-
  tionships. Limnol.  Oceanogr. 25(2):371-2.

McCluney, W.R. 1975. Radiometry of water turbidity meas-
  urements. J. Water Pollut. Control Fed. 47(2):252-66.
Megard, R.O., J.C. Settles, H.A. Boyer, and W.S. Combs, Jr.
  1980. Light,  Secchi  disks,  and trophic  states. Limnol.
  Oceanogr. 25(2):373-7.
                MODELING TECHNIQUES AND INNOVATIONS

Ritchie, J.C., F.R. Schiebe, and J.R. McHenry. 1976. Remote
  sensing  of  suspended sediments  in  surface  waters.
  Photogr. Eng. Remote Sens. 42(12): 1539-45.
Stefan, H.G., T.  Skoglund, and  R.O.  Megard. 1976. Wind
  control of algae growth in eutrophic lakes. Am. Soc. Civil
  Eng. J. Environ. Eng. 102(6):1201-13.
Stefan, H.G., J.J. Cardoni, F.R. Schiebe, and C.M.  Cooper.
  1983. Model of light penetration in a turbid lake. Water
  Resour. Res. 19(1):109-20.
Stewart,  K.M., and  P.J.H. Martin. 1982.  Turbidity and  its
  causes in a narrow glacial lake with winter ice cover. Lim-
  nol. Oceanogr. 27(3):510-17.
Truhlar, J.F. 1976. Determining suspended sediment loads
  from turbidity  records. In  Proc. 3rd Federal Inter-Agency
  Sedimentation Conference, WRC.
Tyler, J.E. 1968. The Secchi disc. Limnol. Oceanogr. 13(1):1-6.
Williams, D.T., G.R.  Drummond, D.E.  Ford, and D.L Robey.
  1981. Determination of light  extinction coefficients in
  lakes and reservoirs. Pages 1329-35 in  H.G.  Stefan, ed.
  Symp. Surface Water Impoundments. Am. Soc. Civil Eng.
                                                        205

-------
  VERIFICATION OF THE RESERVOIR WATER QUALITY MODEL,
  CE-QUAL-R1, USING DAILY FLUX RATES
 CAROL  DESORMEAU COLLINS
 Biological Survey
 New York State Museum
 Albany,  New York

 JOSEPH H. WLOSINSKI
 U.S. Army Corps of Engineers
 ^Waterways  Experiment Station
 Vicksburg, Mississippi
            ABSTRACT

            Emphasis on evaluation techniques where measured versus predicted changes of mass are
            compared can produce a model that predicts the correct answers for the wrong reasons. The
            movement of mass between compartments (tlux) should be compared to measured flux as part
            of the model verification. Model verification tschniques need to reflect the same level of resolu-
            tion inherent in the model structure. The U.S. Corps of Engineers water quality model, CE-QUAL-
            R1, was verified for DeGray Lake, Ark., using this procedure. Several processes important in
            determining the mass of the constituents were measured for DeGray Lake. Primary production
            rate and settling rates of algae, total dry weight, and organic matter were compared to predicted
            fluxes. Results indicated that the model was satisfactory in predicting trends in flux values.
 INTRODUCTION

 That mathematical models are used as a technical
 basis for environmental effects and regulatory policy
 issues emphasizes the need  for such models,  but
 more important, the  critical  nature of the model's
 reliability. The  utility of a model in  assisting  the
 researcher or manager is usually measured by the ex-
 tent of the model's performance evaluation. Conven-
 tional evaluation techniques, where measured versus
 predicted changes of mass are compared, can  pro-
 duce a model that predicts the correct answers for the
 wrong reasons (Scavia, 1980; Wlosinski et al. in prep.).
 Concentrating  only on  measured  versus  predicted
 changes in state variable mass could result in the im-
 plementation of inappropriate management practices.
 Predicted fluxes, or the movement of mass between
 compartments, should be compared to measured  flux
 as part of model evaluation. Realistic solutions to en-
 vironmental water quality problems require attention
 be given not only to the mass of the state variable that
 we perceive as being a problem but also the process
 rate that contributes to determining the mass.
  We describe a coordinated effort between data  col-
 lection and modeling phases of a limnological investi-
 gation to obtain  and compare results of important pro-
 cess rates. The U.S. Corps of Engineers Environmental
 Water Quality and Operational Studies has developed
 a one-dimensional, horizontally averaged, reservoir
 water quality model termed CE-QUAL-R1. It is used to
 study preimpoundment and postimpoundment water
 quality and  the effects of reservoir  management
 operations on water quality.
  CE-QUAL-R1 was evaluated using data collected on
 DeGray Lake, a U.S. Corps of Engineers multipurpose
 project located  on the Caddo  River in the Ouachita
 Mountains in southwestern Arkansas. Its length is 32
 km, and it has a maximum depth of 60 meters. The
volume of the power pool is 7.91 x 108 m3 Wjth a sur-
face area of 5.34 x 107 m2. DeGray Lake has a multi-
level outlet structure and was operated with surface
withdrawal prior to March 1979, at which time the
withdrawal level was lowered 12 m. Data collected in
1979 were used for calibration and data collected for
1980 were used for verification. As part of the verifica-
tion, flux values for selected processes measured in
the field were compared to predicted flux values. The
results of this evaluation procedure are reported here.

MODEL DESCRIPTION

CE-QUAL-R1 simulates the dynamics  of  37  water
quality variables computed in the vertical  direction.
The thickness of each layer depends upon the balance
of inflowing and outflowing water which permits ac-
curate mass balancing and reduces numerical disper-
sion.
  Inflowing waters are distributed vertically based on
density differences; this allows simulation of surface
flows, interflows, and  underflows.  Water density is
dependent on temperature and the concentrations of
dissolved and suspended solids.  Outflowing water is
withdrawn from layers, considering density stratifica-
tion, using the selective  withdrawal algorithms of
Bohan and Grace (1973). Reservoir outflows can either
be specified by using  operation  records or the user
may opt to have the model choose flows from ports in
order to match a target temperature.
  Vertical  transport of thermal energy and mass is
achieved through entrainment and turbulent diffusion.
Entrainment (a  one-way transport  process  that
sharpens gradients) determines the depth of the upper
mixed layer and the  onset  of  stratification.  It is
calculated from the turbulent kinetic energy  influx
generated by wind shear and convective mixing using
an integral energy approach (Johnson and Ford, 1981).
                                               206

-------
                                                                    MODELING TECHNIQUES AND INNOVATIONS
Turbulent diffusion is a two-way transport process
that tends to reduce gradients and is incorporated us-
ing a turbulent or eddy diffusion coefficient that  is
dependent on the wind speed, magnitude of inflows
and outflows, and density stratification.
  The interaction of numerous organic and inorganic
variables is a major attribute of CE-QUAL-R1. The in-
teraction  represents the major processes  of decay,
decomposition,  egestion,   diffusion,  convection,
harvest, ingestion, mortality,  photosynthesis, respira-
tion, settling, inflow, and outflow. The interaction bet-
ween a number of modeled components is presented
in Table 1. A detailed description of the model  is
presented in a Users Manual  (Environ. Lab., 1982).

EVALUATION HISTORY

The comparison  of measured versus predicted flux
rates reported here is part of an ongoing evaluation  of
CE-QUAL-R1 which has included two general types  of
          tests (Wlosinski, in press). The first group of tests are
          used to ensure that the coding of the model is correct,
          and the second group to ensure that model predic-
          tions are reasonable. The second group of tests were
          aided by the excellent data sets collected at DeGray
          Lake, specifically for  evaluating a one-dimensional
          model.
            The initial  phase of model evaluation concerned
          temperature predictions which was  reported on  by
          Johnson (in prep.) That work produced a set of coef-
          ficients  that did an excellent job predicting tempera-
          ture. The second phase concentrated on the predic-
          tion of  oxygen. Although the model predicted on the
          initial simulation  a metalimnetic oxygen minimum
          which occurred at DeGray, the timing was not correct.
          A study of the predicted fluxes on a yearly basis  show-
          ed that the major part of the problem dealt with algal
          production, respiration,  and settling, and sediment
          and dissolved and organic  matter decay. During the
          ensuing calibration simulations, as  the predicted
                               Table 1.—Flux between certain modeled variables.
From
algae 1
algae 2
alkalinity
D.O.M.
ammonia-N
nitrite-N
nitrate-N
coliform
detritus
oxygen
ortho-P
T.D.S.
S.S.
zooplankton
inorg. carbon
benthos
sediment
fish-1
fish-2
fish-3
surface
upstream
W_ 0) 7T
CQ to —
B) 0) 5'
CD (D =_-
-* M ><
Y
Y
F

P P

P P


R R
P P



P P






WWW
nitrite-N
ammonia-N
D.O.M.
R
R

F D
F D
F


D
ODD



R

R
D
R
R
R

WWW
1 8 *
as-?.
I i 5





D
F
F
Y
D



Z







WWW
O 0
$ i n
S 9 S
3 TJ y3
P R
P R

D




D
F
F
F

R

R
D
R
R
R
X
WWW
inorg. carbon
zooplankton
S.S.
1 R
1 R

D




1 D
R


Y
F R
F
R
D
R
R
R
X
W W
fish 1
sediment
benthos
S
S






S
R D R





G Z
I G
Z
Z I
Z I
G G

CO
III
i\o co o>
I
i






i
R R X



I
X
I G
I G
H
H
H


downstream
0
O
O
O
0
O
0
O
O
O
O
O
O
0
o







 D  decay or decomposition
 E  egestion
 F  diffusion and convection
 G  gain or loss caused by layer depth change
 H  fishing harvest
 I   ingestion
 N  non-predatory mortality
 O  outflow
P  photosynthesis
R  respiration
S  settling
W inflow
X  exchange at the air water interface
Y  settling, diffusion and convection
Z  egestion and nonpredatory mortality
                                                  207

-------
LAKE AND RESERVOIR MANAGEMENT
fluxes were  brought  more  in  line with  measured
values, the oxygen concentrations greatly improved. A
problem  still  remained,  for  although  temperature,
most other variables, and fluxes on a yearly  basis
were satisfactory, the predictions  of  a couple of
variables, notably orthophosphate and ammonia, were
considered not satisfactory. Their predictions on a
number  of  occasions  were  about  an  order of
magnitude high. Further calibration attempts did not
remedy the problem.
  At this point we decided that the problem had to do
with the structure of the model or the formulation of
chemical and  biological processes, and that in the
future, fluxes would have to be compared closer, us-
ing  the  same  time  interval  they were  measured.
Changes to the model concerned the compartmenls
of algae, benthos, sediment, fish, dissolved  organic
matter, nitrite, nitrate, and orthophosphate. That work
markedly  improved ammonia and  orthophosphale
while having  no appreciable negative effect  on ary
variables. Further information concerning that work
can be found in a report by Wlosinski and Collins (in
prep.)
  One recurrent problem dealing with the comparison
of measured versus predicted  data concerns  that
model's one-dimensional assumption. Because of this
assumption longitudinal or lateral variations in water
quality constituents cannot be predicted  and all in-
flowing   materials are  instantaneously  dispersed
throughout a horizontal layer. Thus, only one value is
predicted for a particular layer.
  However, at DeGray, three sampling stations were
selected  to characterize the  seasonal, longitudinal,
and vertical gradients in water quality (Thornton et s,\.
1980). Station 12  is located in the headwater regicn
and is 15 m in depth. Station 10 is located in the mid-
dle region and is characterized as mesotrophic with a
maximum depth of 25 m. Station 4, the dam region, is
considered oligotrophic with a maximum depth of 47
m. These three stations have shown a wide range for
measured variables. An order of magnitude difference
in photosynthetic rate is not uncommon between Sta-
tions 4 and 12 on a given day. Most comparisons of
measured versus predicted concentrations involved
Station 4, mainly because it  was deeper, therefore,
having more measurements, and represented the
greatest volume in the reservoir. In addition, the Corps
is interested  in  outflow concentrations from reser-
voirs, and the deepest station usually influences out-
flow concentrations more than other stations.
  Still the variation in  measured values across sta-
tions must be kept in  mind when comparing values.
On an aerial basis, Station 12 represents 18 percent of
the lake, Station 10 represents 32 percent and Station
4 represents 50 percent.
RESULTS AND DISCUSSION

Primary production  data were available for  DeGray
Lake in 1979 and 1980 (R. Kennedy, unpubl. data). The
values were taken monthly and reported as mg carbon
m-2 hr~1. Measurements were made at the  surface
and 1 meter intervals  down  to 7 m for incubation
periods ranging from 2  to 3.5 hours, the periods over-
lapping  solar noon.  Depth-integrated hourly carbon
uptake rates were converted to daily rates for compar-
ative purposes.
  The role of sediment transport and sedimentation
was investigated  in 1980 (R. Kennedy, unpubl. data;
James and Kennedy, in prep.) Data were collected &\
                                                     approximately monthly intervals for six periods bet-
                                                     ween February  and August. Sediment traps were
                                                     deployed at 5 and 15 m  from  the  surface of each
                                                     sampling station and at 45 m depth, 2 m from the bot-
                                                     tom, and at Station 4. Paniculate dry weight, total
                                                     organic carbon, and chlorophyll  a concentration were
                                                     among the variables measured. Sedimentation rates
                                                     were calculated by dividing the amount of the variable
                                                     measured (g m-2) by the deployment period (days)
                                                     and reported as g m-2 
-------
                                                                    MODELING TECHNIQUES AND INNOVATIONS
organic matter while chlorophyll a was considered to
be 0.58 percent of algal dry weight (Spangler, 1969).
  Caddo River inputs strongly affected the timing and
magnitude of C14 productivity  at the headwater (Sta-
tion 12). Autochthonous inputs control production at
the dam  station  (James  and  Kennedy,  in  prep.)
Biweekly productivity rates were highest in the head-
water region whereas the dam region had  low rates
throughout the study period.
  The model was run using a time step of 1 day and
primary production rates were reported daily. A com-
parison of measured with predicted fluxes for 1979 is
given in Figure 1. Results of the verification simulation
for 1980 are given in Figure 2.  Common trends and a
similar range of predictions can be observed through-
out the year.
  We feel  the  predictions are quite satisfactory,
especially  considering  the assumptions that  were
needed for  the  simulations  and evaluation.  The
method used for measuring photosynthesis does not
lend itself  entirely  to  comparison with  predicted
values.  Photosynthetic rate was measured during a
2-3.5  hour  period  overlapping  solar  noon.   This
represents a period when photosynthetic rate is often
at a maximum. However, the model predicts a daily
average photosynthetic rate which  is likely to under-
estimate the measured values. The average variation
in cloud cover was obtained from a meterological sta-
tion location 100 kilometers away from DeGray Lake.
Because variation in this driving variable can signifi-
cantly influence predictions, radiation data taken  at
the  location that  the photosynthetic rate  was
measured would have been more  desirable. In addi-
tion, a single chlorophyll a to dry  weight conversion
factor was used for the entire simulation period.  This
ratio can vary throughout the day for a single species.
Furthermore,  at  least three different algal groups
dominated throughout the  year, each composed of a
different average ratio of chlorophyll a to dry weight.
   After the flux values were obtained for photosynthe-
tic rates, the model was then run accumulating flux
values throughout the year and periodically reporting
the results to mimic the experimental method used in
determining sedimentation rate. Predicted flux rates
at the end of each deployment period were subtracted
from the  initial deployment date then divided by the
number of  deployment  days  to  determine the pre-
dicted sedimentation rate: (Trap out - Trap in) / # of
Deployment days  = Sedimentation rate.
   Chlorophyll a concentration in the sediment traps
was measured to give  an estimate of algal settling
rate at 5, 15, and 45 m (Fig. 3a-c). Predicted algal  settl-
ing rates agreed well with the observed  data at 5 m.
Generally, chlorophyll a concentration measured in
the sedimentation trap increased  until mid-June and
then declined. The observed settling rate at 5 m was
quite different from  that observed at 15 m, which ex-
hibits  a decreasing  rate until the end of  July then in-
creases. Given this type of settling behavior it may be
very difficult to model the complexities  in biological
and physical processes observed.
   CE-QUAL-R1 does not model particulate dry weight
explicity. We approximated the settling rate of this
variable by summing  the settling rates of algae,
detritus,  suspended  solids,  absorbed  ammonia-
 nitrogen, nitrite/nitrate-nitrogen, and phosphate. We
did not consider the contribution by metals or dissolv-
ed compounds. The observed particulate dry weight
 settling rate for 1980 is compared to the predicted
 rates for 1980 at 5,15, and 45 m (Fig. 4a-c). The model
 behaved in a manner characteristic of the system.
  The  settling rate of  total organic  matter was
estimated based upon the settling rate of total organic
carbon and compared to the predicted total organic
matter settling rate at 5, 15, and  45 m (Fig. 5a-c).
CONCLUSION

The results of the model predictions presented in this
paper generally seem to have  underestimated the
measured values.  However, yearly output of results
did show that the range of predicted flux values fell
                  DEGRAY LAKE 5m
                                    • 1980 Station 4

                                    O 1880 prediction

                                    • 1979 prediction
                  DEGRAY LAKE 15m
                                    • 1980 Station 4

                                    01980 prediction

                                    • 1979 prediction
                 DEGRAY LAKE 45 m
                                     • 1980 Station 4

                                     O 1980 prediction

                                     • 1979 prediction
Figure 3 a-c.—Results of measured algal settling rates and
model simulation values at (a) 5 m, (b) 15 m, (c) 45 m.
                                                  209

-------
 LAKE AND RESERVOIR MANAGEMENT
 closely in line with measured values. The wide fluctua-
 tions in measured data were also predicted by the
 model even  though this was not exhibited for the
 verification dates.
    Ideally, flux information should be available for an/
 process for a particular variable at the level of resolu-
 tion used  by the model  (i.e., time step and dimen-
 sionality). This is often difficult given the constraints
 of the model  as well as the limnological sampling pro-
 gram. It is often expensive to run a model at the time
 step  used in measuring many  rates.  For example,
 many process rates are measured per hour and CEi-
 QUAL-R1  is run with a time step of 1 day. The tims
 step  used by  the  model will affect  the computed
 values. A model such as CE-QUAL-R1 is also at a dis-
 advantage because of its one-dimensional  structure.
                  DEGRAY LAKE 5m  1980
   5
   a
   "o
   *  5-»C4
   «
   2

   s
   +:
   I

   *
                                     • Station 4

                                     O prediction
                  DEQRAY LAKE 15m 1980
  The model predicts one value for the entire reservoir;
  station  to  station  differences  are  ignored.  Photo-
  synthesis,   respiration,  bacterial  mineralization,
  decomposition, nitrogen fixation, zooplankton filtra-
  tion, sedimentation, sediment oxygen utilization, and

                        DEQRAY LAKE 5m  1»»0
                                          • Stitlon 4

                                          O prediction
          -I	1	1	1	1	1	1   I    I   I
              FMAMJ   J   ASOND


                        DEGRAY LAKE 15m  1980
                                                                                                 • Station 4

                                                                                                 O prediction
                                                                 -I	1	1	1	1	I   I
                                                                     FMAMJ   J   ASOND
                                                                              DEGRAY LAKE 45m 1980
                  DEGRAY LAKE 45m 1980
                                       • Station 4

                                       Opredlctlon
                                                                                                • Station 4

                                                                                                O prediction
Figure 4 a-c.—Results of measured total dry weight settling
rate and model simulation values at (a) 5 m, (b) 15 m, (c) 45 m.
Figure 5 a-c.—Results of measured organic matter settling
rate and model simulation values at (a) 5 m, (b) 15 m, (c) 45 m.
                                                   210

-------
inflow and outflow rates are examples of important
process rates that can  be used as  a guide  in the
verification procedure.
  Increased attention to process rates will ensure the
continued use and improvement of models. The need
to  further test  models  with  these evaluation re-
quirements will depend on a coordinated data acquisi-
tion and modeling effort.

ACKNOWLEDGEMENT: Research support for C.D.  Collins
was provided  by a contract with the U.S. Army Engineer
Waterways Experiment Station  under the Environmental
Water Quality and Operational Studies Program.


REFERENCES

Bohan, J.P., and J.L Grace, Jr. 1973. Selective withdrawal
  from man-made lakes. U.S. Army Corps Eng. Waterways
  Exp. Sta. Tech. Rep. H-73-4. Vicksburg, Miss.
Environmental  Laboratory.  1982. CE-QUAL-R1: A numerical
  one-dimensional  model of reservoir water quality; Users
  Manual, Instruction rep. E-82-1 (revised ed.; supersedes IR
  E-82-1 dated April 1982). U.S. Army Corps Eng. Waterways
  Exp. Sta., Vicksburg, Miss.
James, W.F., and R.H. Kennedy. In prep. Patterns of sedi-
  mentation at DeGray Lake.
Johnson,  L.S.  In prep. Thermal stratification modeling in
  DeGray. In Proc. Arkansas Lakes Symp., Arkadelphia, Ark.
  Oct. 4-6, 1983.
                MODELING TECHNIQUES AND INNOVATIONS

Johnson, L.S.,  and D.E. Ford. 1981. Verification of a one-
  dimensional  reservoir thermal model. Am. Soc. Civil Eng.
  1981 Meet. St. Louis.
Kennedy, R. Unpubl. data. U.S. Army Corps Eng. Waterways
  Exp. Sta., Vicksburg, Miss.
Scavia, D. 1980. The need for innovative verification of eutro-
  phication models. In Workshop on Verification of Water
  Quality Models. EPA-600-9-80-016. U.S. Environ. Prot. Agen-
  cy, Washington, D.C.
Spangler, F.L 1969. Chlorophyll and carotenoid distribution
  and phytoplankton ecology in Keystone Reservoir, Tulsa,
  Okla. Ph.D. dissertation. Okla. State Univ., Stillwater.
Thornton, K.W., J.F.  Nix, and J.D. Gragg. 1980. Coliforms
  and water quality: Use of data in project design and opera-
  tion. Water Resour. Bull. 16: 86-92.
Wlosinski, J.H. In press. Evaluation techniques for CE-QUAL-
  R1: a one- dimensional water quality model. Misc. Pap.
  U.S. Army Corps Eng. Waterways Exp. Sta., Vicksburg,
  Miss.
Wlosinski, J.H., and C.D. Collins. In prep. Application of a
  water  quality model (CE-QUAL-R1) to DeGray Lake, Ark.
  Proc. Arkansas Lakes Symp., Arkadelphia, Ark. Oct. 4-6,
  1983.
Wlosinski, J.H., K.W. Thornton and D.E. Ford. In Prep. The
  use of fluxes in the verification of water quality models.
                                                      211

-------
                                                        Case  Study:
                              The   Bear  Lake  Project
A HISTORICAL PERSPECTIVE AND PRESENT WATER QUALITY
CONDITIONS IN BEAR LAKE, UTAH-IDAHO
VINCENT A. LAMARRA
Ecosystem Research Institute
Logan, Utah

V. DEAN ADAMS
Utah Water Research Laboratory
Logan, Utah


CRAIG THOMAS
Bear  Lake Regional Commission
Fish Haven, Idaho
REX HERRON
PAUL BIRDSEY
Department of  Fisheries & Wildlife
Utah State University
Logan, Utah

VICTOR KOLLOCK
MARY PITTS
Utah Water Research Labs
Logan, Utah
          ABSTRACT

          In the 1975 National Eutrophication Survey Bear Lake had the best overall water quality of all Utah
          lakes sampled. However, this oligotrophic state would not be retained because of a mesotrophic level
          of loading. Because of the unique characteristics of the Bear Lake ecosystem and the present danger
          of cultural eutrophication, the objectives of this 314 Clean Lakes Study were to: (1) quantify the major
          nonpoint sources of nitrogen and phosphorus into Bear Lake; (2) quantify the major sources of nitrogen
          and phosphorus in the Bear River prior to its diversion into Dingle Marsh and Bear Lake; (3) deter-
          mine the nitrogen, phosphorus, and carbon budgets of Dingle Marsh and define the factors which
          may regulate the flux of these nutrients into Bear Lake, and (4) if necessary, develop a set of viable
          cost-effective alternatives for the reduction of the nitrogen and phosphorus loading into Bear Lake
          Nutrient loading was determined for each major tributary to Bear Lake and the trophic condition of
          the lake was determined over an 18-month period. Trophic state determinations were made using
          TSI values, areal oxygen deficits, and areal phosphorus loadings Differences in parameters predic-
          tions are explained. An historical perspective of the water quality trends is given for Bear Lake
          (1975-1983). Based upon the observed changes and associated land use alterations, a series of
          management plans is proposed for maintaining or improving the water quality in Bear Lake.
INTRODUCTION
Bear Lake, located on the border of Utah and Idaho is
a 282 km2 natural body of water. The lake has a con-
tinuous lacustrian history of at least 28,000 years (BP).
During most of this time, the lake has been isolated
from the major drainage networks, which has led to
the development of four endemic fish species that still
                                          213

-------
 LAKE AND RESERVOIR MANAGEMENT
inhabit  the lake in large  numbers,  and a  unique
macrochemistry with magnesium as the predominant
divalent cation (Kemmerer et al. 1923).
   Dingle Marsh (also referred to as Dingle Swamp) is a
61 km2 freshwater riverine marsh situated adjacent -;o
the north end of Bear Lake.  Historically, the marsh
was separated from the lake by a naturally occurring
sandbar; however, the Telluride and Utah Power and
Light Companies constructed three canals through
Dingle Marsh  and control gates and pumps at Lifton
Station, completing the work in 1915, to divert Bear
River water through the marsh and into the lake.
   Water enters the marsh from Bear River through the
Rainbow and Ream-Crockett (previously referred to as
Dingle Canal) Canals on the north end, from Bear Lake
at Lifton Station on  the south  end  in the summer, and
from three relatively small streams on its west side.
Surface water exits  the marsh at the Outlet Canal, the
Lifton causeway structure during high spring runoffs,
and at Lifton Station, depending on whether runoff is
being stored in Bear Lake or released for downstrea-n
needs. Some exchange of water between the canals
and the main body of the  marsh has resulted fro-n
leakage of the dikes, especially  where water leaks
from the Rainbow Canal and  follows Black Canal, a
natural meandering  channel, through  the marsh.
   Reeves (1954) summarized the types and amounts
of vegetation within the marsh system and found that
almost  71 percent of  the  area was covered  by
emergent vegetation, 78  percent of it hardstem bul-
rush (Scirpus acutis). Wiregrass  (Juncus spp.) and
sedge (Carex spp.) occurred to a much lesser extent.
Mud Lake accounted for most  of the remaining marsh.
Visual inspections have revealed that heavy growths
of  periphyton occur on  the  inundated portions  of
stems of hardstem bulrush during summer months.
   In 1975 the National Eutrophication Survey noted
that Bear Lake had the best overall water quality of all
Utah lakes sampled. However, this oligotrophic state
would not  be  retained because of mesotrophic-leval
loading. Because of the unique characteristics of the
Bear  Lake ecosystem,  and the  present danger  of
cultural eutrophication,  the  objectives of the 314
Clean Lakes Study were:
   1. Quantify the major nonpoint sources of nitrogen
and phosphorus into Bear Lake.
   2. Quantify the major sources of nitrogen and phos-
phorus in the Bear River prior to its diversion into
Dingle Marsh and Bear Lake.
   3. Determine the  nitrogen and phosphorus budgels
of  Dingle  Marsh and define the  factors that  may
regulate the flux of these nutrients into Bear Lake.
   4. If necessary,  develop  a set  of viable, cost-
effective alternatives for reducing  nitrogen and phos-
phorus loading into  Bear  Lake.

Location

The Bear Lake ecosystem and its associated water-
sheds cover approximately 8,250 km2, with 7,000 km2
in the upper Bear River basin and the  remaining 1,250
km2 within the natural Bear  Lake drainage. These m&-
jor watersheds are  within the States of Idaho, Utah,
and Wyoming.
  The Bear Lake valley is, in part at least, of structural
origin. The valley appears to be a graben (rift valley)
bound on both sides by active faults. According  to
Robertson  (1978), Bear  Lake  has undergone three
distinct  stages since its proposed formation 28,000
year B.P. About 8,000 years B.P., major faulting on the
east side of Bear Lake (Lifton Episode) resulted in the
 lake's present position, with a historical water level
 elevation (prior to 1912) of 1,808 m. The geologic activi-
 ty during the Lifton Episode isolated the lake from the
 major drainage networks, resulting in a closed basin
 lake with inflow approximately equal to evaporation.

 Historical Perspective

 The first reported limnological  investigation of Bear
 Lake was conducted in 1912 by Kemmerer, Bovand,
 and Boorman (1923). Since then numerous studies
 have been made of Bear Lake. A discussion of the lim-
 nological characteristics of Bear Lake will  be made
 under three general areas: physical,  chemical, and
 biological.
   Physical Characteristics: Bear Lake is oval shaped,
 about 34 km long and 14 km wide. It has an 81 km
 shoreline and a surface area of 284 km2.
   The six major tributaries to the lake excluding the
 Bear River drain a 1,250 km2 watershed. An average of
 8.1 x  107 m3 of water per year enters the lake from
 this watershed.  Historically,  most  of  this  water
 evaporated. At  the present time the 7,000 km2 Bear
 River watershed is diverted into the lake.
   Bear Lake has been described as dimictic with a
 distinct thermocline at  15-17  m. Summer surface
 temperatures range between 20 and 22°C, while hypo-
 limnetic temperatures are usually below 7°C. The max-
 imum temperature fluctuations of hypolimnetic water
 below 50 m have been  found to be 2°C to 7°C. Secchi
 disk readings are given in Table 1.
 Table 1.—Selected Secchi disk transparencies from studies
              conducted on Bear Lake.
Author (year)
Kemmerer et al. (1923)
Hazzard (1935)
Perry (1943)
McConnell et al. (1957)
U.S. EPA (1975)
Lamarra (1980)
This study (1981-1982)
Secchi disk (meters)
10
3.3-5.8
3-9
4.5
1.8-3.6
4.5-6.7
2.8-6.6
Table 2.—The physical and chemical characteristics of Bear
  Lake, Utah. Chemical characteristics were conducted by
	Werner et al. (1982).	

	Physical-Morphometric Characteristics
Surface area
Mean depth
Maximum depth
Volume
Mean hydraulic retention time
    282 km2
    27m
    63.4m
    7.98 x 109 m3
    92 years
           Chemical Characteristics 10-19-79
Alkalinity
Ca+ +
Mg+ +
K+
Na +
Cl-
S04
Suspended solids
Total dissolved solids
Volatile suspended solids
Total solids
265 mg as CaCOg/l
25 mg Ca++/l
75 mg Mg++/l
3.1 mg/l
39.1 mg/l
54.2 mg/l
19.7 mg/l
5.0 mg/l
457 mg/l
1.5 mg/l
475 mg/l
                                                214

-------
  Part of the north, northwest, and northeast shore of
the lake is sandy beaches. The remaining shoreline is
rocky. However, this rocky zone is not extensive, ex-
tending only 4 meters into the lake.
  Chemical Characteristics: The macro-chemical con-
stituents found in Bear Lake are rather unique in their
relative  abundance  (Table 2). Each  investigation on
Bear Lake has shown that Mg+ + > Ca+ + > Na+ >
K +  and  HCC>3>CI'>SO4= >CO3 = . Conductivities
range between 720 and 680 ^mohs/cm at 25°C and pH
between 8.3 and 9.0. Nunan (1972) developed an em-
pirical model  that predicted that the macrochemistry
of Bear Lake would be similar to the Bear River by the
year 2020. The relationships developed were based on
initial TDS levels of  1,060 mg/l (Kemmerer et al. 1923)
and subsequent reduced  concentrations  which
resulted from the diversion of the Bear River into Bear
Lake. Present levels are 475 mg/l.
  The surface oxygen concentrations during the sum-
mer in Bear Lake have been reported to be near satura-
tion, based on temperature  and pressure  (Lamarra,
1980). However, it  was  also noted that the  areal
hypolimnetic  oxygen deficits were at a mesotrophic
level between  1976 and 1980. These deficits (approx-
imately 40 percent saturation at the end of the sum-
mer) were correlated with the volume of Bear River
water entering the lake through Dingle Marsh.

  Biological Characteristics: Because of the uniform
shoreline in Bear Lake, rooted aquatic plants in the lit-
toral zone of the lake are  scarce. There are a few pat-
ches of cattail (Typha sp.) growing  along the  north-
west shore between  Fish Haven Creek and St. Charles
Creek. Scirpus has also been noted in this same area.
The north  and  south shores  are totally  lacking
emergent vegetation.  Potamogeton is the   major
submergent aquatic plant with beds occurring  along
the west shore from St. Charles Creek to Garden City.
  The  zooplankton  communities  reported  in the
literature for  Bear  Lake are extremely interesting
because of the noticeable lack of large cladocerans.
                  CASE STUDY: THE BEAR LAKE PROJECT

Kemmerer et al. (1923) noted two copepods (Epischura
and Canthocamptus) dominated the community. Perry
(1943), studying the food habits of the Bear Lake cisco,
noted 12 genera of zooplankton: three copepods, three
rotifers, and six cladocerans. The maximum number
of zooplankton found in Bear Lake have ranged bet-
ween 5 and 10 individuals/l (Lentz and Lamarra, 1981).
  The  phytoplankton populations  in Bear Lake  are
greatest in the spring and fall. The most abundant
algal genera  were Ankistrodesmus,  Oocystus,
Lyngbya,  Lagerheimia,  Dinobryon, and  Dictyo-
sphaerium (McConnell et al. 1957).  Diatoms were  not
numerous, and did  not exceed 5 percent by number of
the total cell counts. The mean annual Chi a levels
found in Bear Lake  between 1976 and 1981 ranged bet-
ween 0.45 and 1.03 /^g Chi a/I, with maximum levels
reaching 2.0 ^g Chi a/I.
  The trophic structure of the fish populations in Bear
Lake include two major salmonid predators, the Bear
lake cutthroat trout (Salmo clarki) and the lake trout
(Salvelinus namaycush). The major prey species in-
clude the  Bonneville cisco (Prosopium gemmiferum)
Bonneville  whitefish (Prosopium  spilonotus),  Bear
Lake whitefish (Prosopium abyss/cola), and the Bear
Lake sculpin  (Cotlus extensus).  Because of  Bear
Lake's  isolation for at least 8,000 years B.P. (Robert-
son, 1978) a community of fish  has  developed that in-
cludes  four endemic species: the Bear Lake sculpin,
Bonneville cisco, Bonneville whitefish,  and Bear Lake
whitefish.
   The  uniqueness of the Bear Lake fish community
 lies in  the adaptations of the organisms to each other
 and the importance of the cisco to the overall trophic
 structure. The cisco is a dominant food  item of the
 large predators and is, itself, a planktivore feeding ex-
 clusively on zooplankton within the metalimnion  dur-
 ing  summer stratification. In  turn, the  zooplankton
 community,  as previously noted, has  few  large
 cladocerans, with  its structure dominated by a large
 Epischura  sp. This  organism has adapted a swift
 predatory escape mechanism.
 Table 3.—The estimated TSI values (Carlson, 1977) for Secchi disk, total phosphorus, and chlorophyll a, the areal oxygen
deficits (mgOj/cntfday) and the total phosphorus areal loadings (gmP/mz/yr) for Bear Lake during the 2 study years 1981-82.
I TSI Parameters
Secchi disk transparency (m)
Total phosphorus (ug/l)
Chlorophyll a (ug/l)
Oligotrophic*
Mesotrophic
Eutrophic
II Areal Oxygen Deficits (mg02/cm2/day)
1981
1982
Oligotrophic**
Mesotrophic
Eutrophic
n
21
132
70




0.031
0.043
<.025
.025 - .055
>.055
Range TSI Values
6.6 - 2.8 33 - 45
2-50 14-61
.19-2.67 14-40
<30
30-60
>60





<
III Areal Phosphorus Loading (gm P/m^/yr)	

1981                                   0.049  ± .016
1982                                   0.094  ± .031
Oligotrophic***                          < .07
Mesotrophic                             .07 - .15
Eutrophic                               >.15
•Carlson (1977)
"Hutchinson(1957)
•••Vollenweider(1976)
                                                 215

-------
LAKE AND RESERVOIR MANAGEMENT
Current Limnological Data

During the 18 months of field investigations in  Bear
Lake, vertical profiles (seven depths) were taken at 20
different times. All water quality and biological para-
meters have been summarized by Birdsey et al. (1983).
  The concept of a trophic classification for lake eco-
systems has long been recognized. Early studies in-
vestigated the quantity and quality of plankton and
have been summarized by Rawson (1956). More recent-
ly, trophic state has been defined by nutrient loading
(Vollenweider, 1976), complex  ecosystem  models
(Simon and Lam, 1980; Ditoro and Matystik, 1980), aid
by the interrelationships of a variety of parameters
(Porcella, 1980). A coarse resolution technique ussd
by Carlson (1977) resulted  in using single but inter-
related parameters.
  Total  phosphorus, chlorophyll a, and  Secchi  disk
transparency  were  shown  to provide  an excellent
basis for a trophic state index (TSI). A comparison was
made between those three  parameters and their TSI
values with areal phosphorus loadings (g  P/m2/year)
and areal hypolimnetic  oxygen deficits  (mg  02/cm2/-
day). These comparisons can be seen in Table  3. In
each  case the static (TSI), dynamic (areal  oxygon
deficits), and  predictive (areal phosphorus loadings)
trophic state classifications indicated that Bear Lake
was upper oligotrophic to strongly mesotrophic. Prac-
tical differences in  the  TSI values are  explained by
Birdsey  et al. (1983).
  Phosphorus and the  Bear Lake Ecosystem: During
1981 and 1982 a detailed phosphorus budget was con-
ducted for the Bear Lake watershed. The results of
this investigation are summarized in Table 4. The areal
phosphorus loading during  a dry year (1981)  from all
sources was 0.049 ± 0.016 g P/m2/yr, well within tie
acceptable loading  levels  for an oligotrophic lake.
However,  during  1982  (wet  year), the  phosphorjs
loading  levels more  than doubled  (0.094 ±  0.031 g
P/m2/yr), placing the  lake  into the  mesotrophic
category. During both years,  the single  most impor-
tant source was the Bear River connection,  accoun-
ting for  61 percent and 71 percent of the total phos-
phorus.  This  source has  been  diverted into  the
historical  Bear Lake watershed and represents a
substantial addition to the phosphorus budget.
  To place this study in historical perspective, these
data must be compared to previous investigations. Lit-
tle detailed limnology was  conducted on Bear La
-------
                                                                       CASE STUDY: THE BEAR LAKE PROJECT
0.26 HQ chl a/I). The cause of this increase is unknown.
A gradual increase has occurred between 1979 and
1980. During 1981, the chlorophyll a concentration ap-
peared  to decrease.  The reason is believed to be
related  to  the  Bear  River  inflow  and  watershed
loading.
  The phosphorus levels in Bear Lake, expressed as
the mean monthly (April  to  September) epilimnetic
total phosphorus  concentrations, can be  seen in
Figure  3. These data indicate a significant  relation-
ship (r2  = .70; N  = 28) between the increase in total
phosphorus  and  time (1976-82). Although  a linear
model was fit to these data, the last 2 years indicate a
substantial increase over previous concentrations.
  Because of the importance of the Bear River con-
nection, and its role in the phosphorus cycle (50 to 60
percent of the total phosphorus loading to Bear Lake),
a historical view of this source is needed.
  A comparison of the 1981  and 1982 areal oxygen
deficits (mg O2/cm2/day) with the annual  Bear River
    1.75
    1.50

 ^  1.25
 2  1.00

 •'  0.75
 §  0.50

    0.25-j
      0
           1976 1977 1978 1979 1980 1981 1982
                        YEARS
 Figure 2.—The mean annual chl a (MQ/I  ± S.E.) values for
 Bear Lake between 1976 and 1982. The data were obtained
 from the following sources: • Lamarra (1977); o Lamarra et al.
 (1982); D this study; 1981-April to December; 1982-January
 to July.
   16


   14
 1
 I10
 o
    8
 o
 »-  4
Y:.085X
N:28  r?.70
                   flow (m3 x 108) into Bear Lake indicates that a model
                   previously used (Lamarra, 1980) remains valid (Fig. 4).
                   These  data indicate that  this  source  of  water,
                   nutrients, and organics is directly related to the rate of
                   oxygen depletion in  the hypolimnion  of  Bear Lake.
                   Low water years  produce lower oxygen deficits while
                   high  flow results  in  high oxygen utilization in the
                   hypolimnion. In 6 of 8 years (since 1975) the lake has
                   had a eutrophic oxygen deficit. The cause of this ox-
                   ygen loss could  be stimulated production by nutrient
                   loading or degradation of allochthonous organic input
                   (92  x  104 kg TOG in 1981 and 252 x  10* kg TOC in
                   1982).
                     Effect  of Water Quality Changes on the Bear Lake
                   Ecosystem:  The Bear Lake ecosystem  is  unique.
                   Because of its  isolation  for over 8,000  years, the
                   biological community  evolved  into  a  simple,  co-
                   existing trophic  structure, with four endemic species
                   of fish. The study presented here has noted  that in-
                   creased  phosphorus loadings  have  resulted in in-
                   creases in algal biomass and  total phosphorus and
                   decreases in hypolimnetic oxygen levels. The results
                   of this cultural  eutrophication  can  be  manifested
                   through two mechanisms:  a physical and chemical
                   one and a biological one.
                     The physical and chemical changes in Bear Lake
                   that  could be  associated with increased eutrophica-
                   tion are the increases in turbidity and the reduction of
                   oxygen. Because of  the  projected increase in algal
                   biomass with  increases in  phosphorus loading,  light
                   penetration could  be expected to decrease. Lamarra
                   et al. (1982b)  noted  a  log-log  relationship  between
                   chlorophyll a and transparency for areas within Bear
                   Lake. A  doubling  in chlorophyll  a from  its existing
                   summer low of 0.20 pg/l to 0.40 f*g/l will  result in a max-
                   imum Secchi disk transparency of only 4.0 m.
                     Three of the four endemic fish species that exist in
                   the lake require cold  water. A reduction in the oxygen
                   content of the hypolimnion will reduce the available
                   habitat for these species. Anaerobic conditions within
                   the hypolimnion  of Bear Lake will result in the reduc-
                   tion of the Bonneville Cisco, Bonneville whitefish, Bear
                   Lake whitefish, the Bear Lake cutthroat trout, and the
                   lake trout.
                     Prior to the loss of hypolimnetic oxygen, certain
                   biological changes may occur within the lake, which
                   could alter the endemic fish community. At present,
w 0.10

o

£J 0.08 -
O)

~ 0.06 -

O
uj 0.04 -
O
                                                       0.02 H
       1976  1977  1978  1979  1980  1981  1982
                         YEARS
                    x
                    O
                    _l
                    UJ
                    cc
r2=.83 n=8
y:.016x + .01
                1.0       2.0       3.0
          ANNUAL BEAR RIVER FLOW (m3x108)
                                                                                                    I
                                                                                                   4.0
 Figure 3.—The  mean summer months (April-September)
 total phosphorus concentrations from the epilimnion of Bear
 Lake. Sources of  the data are: 1976-77  Lamarra  (1977);
 1978-81 Lamarra et al. (1982); 1981-82 this study.
                    Figure 4.—The areal oxygen deficits (mg O2/cm2/day) during
                    summer stratification as related to the Bear River inflow (m3
                    x 108). Sources of the data are: • Lamarra 1980; o this study.
                                                  217

-------
 LAKE AND RESERVOIR MANAGEMENT

 the cutthroat trout restoration program has markedly
 improved the population density of this species. An in-
 crease in the primary producers will initially increase
 biomass throughout all trophic levels including fish.
 However, if allowed  to  continue, changes  in species
 composition  within the lower trophic levels (primary
 producers and zooplankton) may result in changes of
 the intermediate  prey organisms (cisco). At present,
 the Cisco plays an important role in the structure and
 function of this ecosystem (Lentz and Lamarra, 1981).
 The potential removal of this population (by whatever
 means)  may  dramatically change  the Bear  Lake
 ecosystem.
   Based upon this study and previous investigations,
 it was recommended that the regulatory  agencies
 adopt a policy that will preserve the Bear Lake ecosys-
 tem as it exists. The integrity of the total community
 must  be maintained  and allowed to  evolve indepen-
 dent of man's activities. The management of a single
 species or trophic level would be contradictory to this
 end. A 5-year Bear Lake preservation plan is underway
 (Thomas et  al. 1983) to help preserve this unique
 resource.
REFERENCES

Birdsey, P., V.A. Lamarra, and V.D. Adams. This volume. The
  effect of coprecipitation of CaC03 and phosphorus on the
  trophic state of Bear Lake. In Proc. Int. Symp. Lake Reser-
  voir Manage., Knoxville, Tenn. N. Am. Lake Manage. Soc.
Carlson, R.E. 1977.  A trophic  state index for lakes. Limnol.
  Oceanogr. 22:361-9.

Ditoro, D.M., and W.F. Matystik. 1980.  Mathematical model
  of water quality  in large lakes, part 1: Lake Huron and
  Saginaw Bay. EPA-600/3-80-056. U.S. Environ.  Prot. Agen-
  cy, Duluth, Minn.

Hazzard, A.S. 1935.  A preliminary limnological study of Bear
  Lake, Utah-Idaho with particular reference to its fish pro-
  ducing possibilities. U.S. Bur. Fish. Unpubl. rep.

Hutchinson, G.E. 1957. A Treatise  on Limnology,  Vol.  I.
  Geography, physics, and chemistry. John Wiley and Soms,
  Inc., New York.

Kemmerer, G., J.F. Bovard, and W.R. Boorman. 1923. North-
  western  lakes of  the United States;  biological  and
  chemical studies  with reference to possibilities to produc-
  tion of fish. U.S. Bur. Fish. Bull. 39:51-140.
Lamarra, V. 1977. Limnology of springs and lakes in northern
  Utah. Unpubl. data.
	1980.  The  nitrogen  and  phosphorus  budgets of
 the Bear River and Dingle Marsh, and their impact on a
  large oligotrophic water storage reservoir. Pages 371-81 in
  H.G. Stefan, ed.  Proc.  Symp. Surface  Water Impound-
  ments. Am. Soc. Civil Eng.
  Lamarra, V.A., C. Thomas, and D. Lentz. 1982a. The .physical,
   chemical, and biological effects of large marinas on the lit-
   toral zone of  Bear Lake, part I. Limnological Conditions
   Bear Lake Region. Comm.

 	1982b.  The  physical,  cheimcal,  and  biological
  effects of large marinas on the littoral zone of Bear Lake,
  part II. Water Qual. Manage. Plan. Bear Lake Region. Comm!

 Lentz, D., and V.A. Lamarra. 1981. The effect of the Cisco on
   Bear Lake zooplankton. Paper presented at the Bonneville
   Chapter, Am. Fish. Soc. Salt Lake, Utah. Feb. 1-2.

 McConnell, W.J., W.J.  Clark, and W.F. Sigler. 1957. Bear
   Lake: its fish and fishing. Joint Publ. Utah Dep. Fish Game,
   Idaho Dep. Fish Game, and Wildl. Manage. Dep. Utah State
   Agric. College.

 Nunan, J. 1972. Effect of Bear  River storage on water quality
   of Bear Lake. M.S. Thesis. Utah State Univ.,  Logan.
 Perry, LE. 1943. Biological and economic significance of the
   peaknose Cisco of Bear  Lake, Idaho and Utah. Unpubl
   Ph.D. Diss. Univ. Michigan.

 Porcella,  D.B. 1980. Index to  evaluate lake restoration. J.
   Environ. Eng. Div.  Proc. Am. Soc. Civil Eng. 106:1151.
 Rawson,  D.S. 1956. Algal indicators of trophic state types.
   Limnol. Oceanogr. 1:18-25.

 Reeves, H.M. 1954. Muskrat and waterfowl production  and
   harvest on Dingle  Swamp, Bear Lake County, Idaho. M.S.
  Thesis. Utah State Univ., Logan.

 Robertson, G.C.  1978. Surficial deposits and  geologic  his-
  tory, Northern  Bear  Lake  Valley,  Idaho. Unpubl  MS
  Thesis. Utah State Univ.

 Simon, T.J., and D.C. Lam. 1980. Some limitations on water
  quality models for large lakes: A case study of Lake  On-
  tario. Water Resour. Bull. 16:105.

 Thomas,  C, V.A. Lamarra,  and V.D. Adams.  1984. Socio-
  economic and  political problems associated with the im-
  plementation  phase of the  Bear Lake 314  Clean Lakes
  Study. In Proc. Symp. Lake  Reservoir Manage. Knoxville,
  Tenn. N. Am. Lake Manage. Soc.

 U.S.  Environmental  Protection Agency.  1975.  National
  Eutrophication Survey Methods 1973-76. Working Paper
  No. 175. NTIS: PB-248 886.

 Utah State Water Research  Laboratory. 1974.  Planning for
  water quality in the Bear River system in the State of Utah
  Bur. Environ. Health. PRWG-142-1.

Vollenweider, R.A. 1976. Advances in defining critical loading
  levels for phosphorus in lake  eutrophication. Mem. 1st Ital
  Idrobiol. 33:53-83.
                                                     218

-------
SOCIOECONOMIC AND POLITICAL ISSUES ASSOCIATED WITH
THE IMPLEMENTATION PHASE OF THE
BEAR LAKE 314 CLEAN LAKES STUDY
CRAIG THOMAS
Bear Lake Regional Commission
Fish Haven, Idaho

VINCENT LAMARRA
Ecosystem Research Institute
Logan, Utah

V. DEAN ADAMS
Utah Water Research Laboratory
Utah State University
Logan, Utah
            ABSTRACT

            As a result of the diagnostic portion of the 314 Clean Lakes Study on Bear lake, three alternatives
            were selected because of their expected water quality improvement potential, technical feasibility,
            environmental impact, and estimated cost: (1) site specific recommendations—Upper Bear River water-
            shed management, Dingle Marsh modifications, and the use of BMP's in two small problem water-
            sheds within the Bear Lake drainage; (2) a site specific basinwide water quality management plan;
            and (3) an environment information and education program. Because of the limitation in Phase II Clean
            Lakes moneys, the implementation of the diagnostic study was difficult A case history is presented
            on how the project was continued Because of the location of Bear Lake, four Federal agencies and
            three States became involved in the existing Bear Lake Preservation Project. An example of the value
            of a Regional Commission and the important coordinating role played by this type of agency will be
            discussed
INTRODUCTION

Bear Lake stradles the Utah-Idaho  State line near
Wyoming;  (Fig. 1) this along  with other unique cir-
cumstances and characteristics makes the long-term
management and protection of the Lake complicated
and challenging. Regulatory responsibilities not only
reside in three States, but also two regions of the U.S.
Environmental Protection Agency (VIII and X).

MULTI-USE RESOURCE

Uses of Bear Lake are multifaceted. In the last 15 to 20
years, Beat Lake  has developed into a prime recrea-
tional area. Small resorts have given way to  large
recreational  development companies  offering ex-
clusive recreational  packages. Condominiums and
time share sales, guaranteed country club rights, golf
courses, tennis, swimming pools, and  marinas are
now the major attractions. Three of the major develop-
ments are Bear Lake West, Swan Creek Village, and
Sweetwater Park.
  Small farms around  Bear Lake have been sold to
large recreational development companies or have
been subdivided for second home sites. The number
of second home or summer home developments in the
Bear Lake Basin can be seen in Table 1.
  Bear Lake has a very significant seasonal user pop-
ulation. This user population is made up of State park
users, private recreation development users, and se-
cond home owners.
  The day usage at North Beach State Park, Idaho
was nearly 78,000 from April to September of 1981.
During  this period (April-September) nearly  245,000
total visitations occurred for all State parks within the
basin (Utah & Idaho State Dep. Parks Peer., 1982).
  A random sample survey of the visitors to the Utah
park areas was  conducted during the 1981  user
season. The  survey  indicated that  over 70  percent
were residents of Utah, 15 percent were residents of
Idaho, and approximately 5 percent were from Wyom-
ing (Utah State Div. Parks Recr.,  1982).
  Population densities in the basin  during the prime
recreational season and on long weekends are as  high
as 300,000 people within the Bear Lake  Valley. (Bear
Lake Region. Comm., 1980) Since the permanent popu-
lation (1980 census), is approximately 10,000 in Bear
Lake County and Rich County combined, the impact
of this  transitory increase is tremendous.
  State operated  parks  (Rendezvous  Beach, Utah
State Boat Park, Eastshore, and  North Beach) offer a
variety  of  recreational  opportunities.  Launching
facilities and mooring capabilities are available at the
State Boat Park. Picnicking and camping are available
at Rendezvous  Beach. North Beach offers launching
facilities and an extensive stretch of swimming beach.
Figure  2 gives usage totals for yearly visitations for
the last 5 years. The constant and steady rise in use of
the facilities indicates the popularity of Bear Lake as a
recreational area.
                                              219

-------
LAKE AND RESERVOIR MANAGEMENT
   A National Wildlife Refuge is located adjacent to
 the north end of Bear Lake. The Mud Lake and sur-
 rounding  marsh  area  provide bird  watching oppor-
 tunities and seasonal hunting for ducks and geese.
   Studies on  the Bear Lake fish  populations have
 classified  four species of fish  endemic to the lake.
 These species, found only in Bear Lake, are: the Bon-
 neville Cisco; Bonneville whitefish; Bear Lake white-
 fish; and the Bear Lake sculpin. A program to enhance
 the Bear Lake cutthroat trout population is presently
 underway by the Utah and Idaho Divisions of Wildl fe
 Resources.
   A primary use of Bear Lake is the diversion of the
 Bear River into  the lake and subsequent use of  the
 water for downstream power production and irriga-
 tion. The initial construction of facilities to divert Bear
 River water into Bear Lake was  begun in 1909 by the
 Telluride  Power Company. In 1912,  Utah Power and
 Light Company continued work on the  Bear Rivef-
 Bear Lake project with completion occurring  in 19'8.
 (Utah Power Light Co, 1975)
                °s-Jf,      s   r
               \fc    I:      i   (
   The key to the whole water development policy has
 been the use of Bear Lake as a storage reservoir. Con-
 verting Bear Lake into a  reservoir required digging of
 inlet and outlet canals connecting the river to the lake
 at the north end. This system of canals and control
 structures allowed the spring runoff to be diverted
 from the Bear River into  the Lake, and released back
 into the river at a later date.
   The coordination of this water use is the  respon-
 sibility of the Bear River  Commission, with water dis-
 tribution controlled by the water boards  of the States
 of Idaho, Utah, and Wyoming. A compact outlining the
 stipulations for the tri-State division of water is called
 the Bear River Compact.


 FORMATION AND ACTIVITIES OF THE BEAR
 LAKE REGIONAL COMMISSION

 As previously stated, in the early 1960's Bear Lake be-
 came a major recreation area. By the late 1960's an in-
 crease in recreational  demands  on  the  lake environ-
  Table 1.—The recreational developments (proposed and
     platted lots) through 1982 for Bear Lake and Rich
                     Counties.

   Location                                Number

 Idaho
   platted
   proposed
   total
 Utah
   platted
   proposed
   total
 Grant Total:
 Source Bear Lake Region Comm
 2384
 4551
 6935


 3962
 9514
13476
20411
   35O--


   300"


~ 250-


   200--


   150-


   100--
                                                    X

                                                    CO

                                                    O
                                                   O
                                                   OL
                                                                                            TOTAL
                                                                                            UTAH
                                                                                             IDAHO

                                                                                         .-O---O
                                                                 77
—I-	f—
 78    79
  YEARS
                                80    81
   82
Figure 1.—A location map for Bear Lake and its watershed.
Figure 2.—The total number of visits (person nights) at the
major Bear  Lake recreational areas between 1977-1982.
Source: Utah Division of Parks & Rec. (1977-1982); Idaho
Dept. of Parks and Rec. (1977-1982).
                                                220

-------
                                                                    CASE STUDY: THE BEAR LAKE PROJECT
ment was evident and a number of large development
corporations began focusing attention on Bear Lake
and its basin.
  Local elected officials and  citizens  became con-
cerned as the character of the basin and the quality of
Bear  Lake began  to  change. As  a  result,  public
meetings were held on Aug. 12, 1971, at an informal
congressional  hearing.  Local representatives  ex-
pressed their concern regarding the rate of pollution
into the lake, due to the large numbers of tourists and
developers, in  combination with a lack of adequate
sanitary facilities.
  It was evident that Bear Lake citizens favored  the
formation of a bi-State, bi-county organization to coor-
dinate planning  efforts  in both counties and sup-
ported construction of adequate sewage facilities.
  Early in 1973, through the efforts of Governors Cecil
Andrus of Idaho and Calvin Rampton of Utah, the Bear
Lake  Regional Commission was formed  with repre-
sentatives from local and State governments. Goals
and objectives were formulated by the Commissioners
to provide  long-term  direction  and guidance  in  ad-
dressing the needs and  problems of the Bear Lake
area,  and  to preserve and promote Bear Lake's  en-
vironment  and  the Bear Lake Basin resources.
  Five of the major goals of the Regional Commission
were (Bear Lake Region. Comm., 1975):
  1.  To acknowledge the Bear  Lake Valley as a
resource of regional significance, and provide for its
continued  utilization and preservation.
  2.  To provide  for the  maximum public benefit of
valley  resources at minimum public cost along with
preservation of natural resources.
  3.  To coordinate public and private interests, in-
cluding the local, regional, State and Federal govern-
ments in the long-term management of the Bear Lake
Valley environment.
  4.  To maintain technical staff competent to advise
or assist local  government.
  5.  To develop and assist in the implementation of a
total planning  program based on the natural environ-
mental constraints of the air, land, and water.
  These five goals set the framework  for five major
functions  that encompass the Bear Lake Regional
Commission's  activities. These functions are (Bear
Lake  Regi9n. Comm., 1975):
  1.  Bear Lake protection and development
  2.  Natural resources planning
  3.  Coordination and cooperation
  4.  Local planning and  grantsmanship assistance
  5.  Public involvement and education
  Underlying all of  these  functions has been the Com-
mission's  policy to embark only  upon projects  and
programs that  ultimately lead to implementation.
  In keeping with these  stated goals and policy,  the
Commission has been pursuing construction of sewer
facilities around Bear Lake. The Utah portion, from the
State line to Sweetwater, has been designed with con-
struction starting by late fall 1983.
  The Regional  Commission has also been  investi-
gating other factors affecting the water quality  of Bear
Lake. These include studies related to marina design
and operation in Bear Lake, possible effects of oil  and
gas spills into the Lake, and the just completed Phase
I 314  Clean Lakes Study.
  The successful results of these projects are due to
a proper natural  resource data  base, sound scientific
techniques, and the cooperation and coordination of
all parties  involved. Coordination  and cooperation is
the very essence of the Bear Lake Regional Commis-
sion.  Beginning with  its formation, the linking of  two
counties and two States facilitates problem solving of
regionally important issues.
  In the case of the 314 Study and its subsequent im-
plementation program, "The Bear Lake Preservation
Project," the coordination and cooperation  function
was called upon for its ultimate test.
COMMISSION'S COORDINATION ROLE

An example of this key function can be seen in the 314
Phase I  project. The State of Idaho was the funding
agency for the 314 monies and the State of Utah pro-
vided in-kind match to the project through its sam-
pling and laboratory services on the upper Bear River.
Some match monies were also  provided by the main
subcontractor, the Utah Water Research Laboratory,
Utah State University. This was all coordinated by the
Bear Lake Regional Commission, the contractor for
the project.


THE BEAR LAKE PRESERVATION PROJECT

The results of the 314 Phase I project were alarming to
say the  least. It pointed out that  unless the nutrient
loading  to Bear Lake is reduced, cultural eutrophica-
tion will continue unchecked. Any deterioration from
the current status will have adverse  implications on
the unique ecosystem and multiple use of Bear Lake.
  At the time the 314 Phase I Study ended, the future
of Phase II monies was in doubt. Realizing the signifi-
cance of these nutrient loading trends and the cost ef-
fectiveness of  addressing the problems now, the
Regional Commission  immediately   developed  an
implementation work plan and began seeking funding.
  This 5-year plan entitled the Bear Lake Preservation
Project includes five major tasks as outlined in Figure
3 (Bear Lake Region. Comm., 1983a, b).
  This plan was then submitted to the States of Utah
and Idaho for funding. The State of Utah led by Gover-
nor Matheson and the State legislature, passed an ap-
propriation for the first year's funding. Before any fur-
ther funding will be considered by the State of Utah,
the support and cooperation of the States of Idaho
and Wyoming, and Utah Power and  Light Company
must be obtained. The support of these entities has
been made essential by the natural course of the Bear
River through the three States and its manmade con-
nection  to Bear Lake, which is operated by Utah Power
and Light Company.
        BEAR LAKE PRESERVATION
                   PROJECT


 • ADMINISTRATION AND COORDINATION

 • WATER QUALITY MANAGEMENT PLAN

 • INFORMATION AND EDUCATION PROGRAM

 • SITE SPECIFIC ALTERNATIVES

     UPPER BEAR RIVER WATERSHED
     MARSH MODIFICATION
     B M P's: BIG CREED SWAN CREEK
 • BASE LEVEL MONITORING PROGRAM

Figure 3.—Major tasks of the Bear Lake Preservation Pro-
ject.
                                                221

-------
 LAKE AND RESERVOIR MANAGEMENT
                                 BEAR LAKE RI-GIONAL COMMISSION
                             EXECUTIVE  COORDINATING COMMITTEE
 Figure 4.—Interdisciplinary Task Force organizational chart.
   Built into the original Bear Lake Preservation Pro-
ject work plan is the formation  of  interdisciplinary
task forces  from all  parties with key interests in the
lake and its pollution problems. (Fig. 4). The overall
purposes of these task forces are to  insure coordina-
tion of the  project, to identify  funding  sources and
services and to insure that all parties have an oppor-
tunity  to assist in refinement and implementation of
the solutions. The conditions placed on Utah funding
made this task force element vital to the Preservation
Project's overall success.
   At this stage of the project, 6 months into the first
year, all of the individual task forces have been formed
and are functioning.
   As an example, at the time the Bear Lake Preserva-
tion Project  was started, a separate effort had already
begun  to address a problem of  indiscriminant camp-
ing with lack of  adequate sewage  and solid waste
facilities along the shores of Bear  Lake,  mainly in
Utah.
   The  Bear  Lake Regional Commission participated
in this effort by coordinating it with the Bear  Lake
Preservation Project and  obtaining  assistance  from
members of  the Utah task force and others in controll-
ing the camping problem.
   The  details of this effort, under the direction of tie
Bear Lake Regional Commission, included the writing
and adoption of  a camping ordinance  by the  local
county, followed by a coordinated  information and
education program consisting of a pamphlet, news
releases and highway signage, participated in by the
State  Highway  Departments, State  Health Depart-
ments, Parks and Recreation Departments, and Com-
munity Affairs Departments.
  The coordinated effort and success described  is an
excellent example of what is hoped to be accom-
plished by the entire Bear  Lake Preservation Project.
  The current effort by the Bear Lake Regional Com-
mission is now concentrated on meeting the condi-
tions of the Utah funding. With the task forces in place
and functioning, the obtaining of future funding is in
progress.  Potential funding sources include  private
funds from  Utah Power and Light Company, State
funds from Utah, Idaho, and Wyoming and 314 funding
through EPA.
  The  State of  Idaho  in behalf of the Bear  Lake
Regional Commission and the future of Bear Lake, is
applying  for  Phase II 314 monies under  the  current
funding allocation. The decision on which  elements of
the Bear Lake Preservation Project to be funded under
this Phase II  application has not been made yet.
  With sound goals and a forward direction, the Bear
Lake Regional Commission looks to the future  of Bear
Lake and its preservation with optimism.
REFERENCES

Bear Lake Regional Commission. 1975. History, Goals and
  Objectives of the Bear Lake Regional Commission. Work.
  File. Fish Haven, Idaho.
      „. 1980. Recreation use levels. 1980 file. Fish Haven,
  Idaho.

	. 1983a. The Bear Lake 314 Clean Lakes Study. Fish
  Haven, Idaho.

      _. 1983b. Bear Lake Preservation  Project, Five Year
  Workplan. Work. File. Fish Haven, Idaho.

Idaho Department of Parks and Recreation. 1982. Parks visi-
  tation data file. Boise.

Utah Division of Parks and Recreation. 1982. 1982 visitation
  data file. Salt Lake City.

Utah Power and Light Company.  1975. Bear River System
  brochure. Salt Lake City.
                                                 222

-------
 THE NITROGEN, PHOSPHORUS, AND CARBON  BUDGETS OF A

 LARGE RIVERINE  MARSH, AND THEIR IMPACT

 ON  THE BEAR  LAKE ECOSYSTEM


 REX C. HERRON
 Department of Fisheries and Wildlife
 Utah State University
 Logan, Utah

 VINCENT A. LAMARRA
 Ecosystem  Research Institute
 Logan, Utah

 V.  DEAN ADAMS
 Utah Water Research Laboratory
 Logan, Utah



           ABSTRACT

           Adjacent to the north end of Bear Lake is a large (65 km2) freshwater marsh. Prior to 1912, Dingle
           Marsh was separated  from Bear Lake by a naturally occurring sandbar and covered approximately
           100 km2. Seventy years ago, Utah Power and Light constructed a canal system which effectively diverted
           the Bear River into Dingle Marsh The present water system operates by diverting spring water from
           the Bear River through the marsh and into Bear Lake. During late summer when irrigation demand
           is high, water is removed from the lake, passed through the marsh, and released into the rwer. The
           major objective of this portion of the Clean Lakes Study was to determine the nitrogen, phosphorus,
           and carbon budgets of Dingle Marsh and define the factors that may regulate the flux of these nutrients
           into Bear Lake. Sixteen sites within the marsh and all major inflows and outflows were sampled over
           an 18-month period. The data indicated specific seasonal trends within the marsh (sources or sinks)
           of the target nutrients. Furthermore, the marsh mass balances indicated  that on an annual basis, the
           marsh acted as a  sink. However, during specific periods of time, nitrogen, phosphorus, and carbon
           were produced within the marsh system and exported (the marsh was a  source). Management alter-
           natives were generated as a result of this portion of the project to maximize the marsh as a nutrient
           sink for inflowing  Bear River water.
INTRODUCTION

Since water from  Bear River flowed through Dingle
Marsh  immediately prior to its entrance  into  Bear
Lake, an assessment of the marsh's impact on river
water  quality  was  desirable.  The  marsh  was
suspected  of causing changes  in  loading values
although  the magnitude and direction  of changes
were unknown. More specifically, the objective of this
study was  to determine  mass balances for Dingle
Marsh  for  total  suspended  solids,  phosphorus,
nitrogen, and total organic carbon so that an estimate
could be made of the increase or reduction of nutrient
loadings into Bear Lake from Dingle Marsh.
SAMPLING  PROCEDURES

Eight  major  inflows  and outflows  were sampled
around the perimeter of Dingle Marsh: Lifton Station,
Lifton  causeway  structure,  Ream-Crockett  Canal,
Rainbow Canal at Stewart Dam, Outlet Canal at Dam,
Bloomington Creek, Spring Creek, and an upper arm of
St.  Charles Creek (Fig.  1). A total  of 22 surface
samples were collected twice a month  from April
through August of each year and once a month the re-
mainder of the study period. Samples were returned to
the Utah State Water Research Laboratory and analyz-
ed for total alkalinity, pH,  total suspended solids
(TSS), total phosphorus (TP), orthophosphates (P04-P),
total nitrogen (TN),  total  Kjeldahl nitrogen (TKN),
                            Stewart
                              Dam
                              ••

                              Ream-Crockett
                                 Canal
                    Lifton
                   Station
Figure 1.—Map of Dingle Marsh, illustrating sampling loca-
tions. ML is Mud Lake.
                                              223

-------
LAKE AND RESERVOIR MANAGEMENT
nitrates (NO3-N), nitrites (NO2-N), ammonia (NH3-N),
total organic carbon (TOG), and chlorophyll a (Table 1).
  Total soluble inorganic nitrogen (TSIN) was deter-
mined  by summing nitrates, nitrites, and ammonia
while a peroxydisulfate technique, generally evaluated
to be equal  to or better than the standard Kjeldahl
method, was used to determine TKN. Temperature,
conductivity and dissolved oxygen were measured in
the field using a YSI dissolved oxygen or conductivity
probe.
  Flow measurements for the major inlet and outlet
canals were collected from Utah Power and Light per-
sonnel while flow measurements for the three streams
were measured  each sampling  trip with a Marsh-
McBirney Model 201 water-current meter. Constituent
concentrations for each sample were multiplied by the
respective flow rates to determine loadings, and mass
balances were determined by summing all inputs and
outputs over time and taking the difference between
the two sums to obtain net uptake/release by the
marsh.
  The  sampling  period (April 24, 1981, to June  23,
1982) was divided into 2 sampling years. Sample year I
extended from April 24, 1981, to April 24, 1982, and
sample year II ran from June 23,1981, to June 23,1982.
Since spring is the peak runoff time for snowmelt, and
since 1981 was a very dry year and 1982 was a very wet
year, each runoff  period was incorporated  into a
separate sampling year. The overlap encompassed
late summer and winter efforts only, which are assum-
ed to be typical for southeastern Idaho. This method
of examining the data allowed comparisons between
 wet and  dry  years  and their  impacts  on nutrient
 dynamics within Dingle Marsh.


 RESULTS AND DISCUSSION

 General Characteristics

 Many general  physical and chemical characteristics
 (temperature, conductivity, pH, and total alkalinity) of
 the marsh were similar to the Bear River or Bear Lake,
 depending  upon the direction of water movement.
 Temperature varied from 20 to 25°C during the sum-
 mer, but decreased to 0°C in the winter with ice cover-
 ing the entire marsh. Conductivity ranged primarily
 between 400 and 900 ^mho/cm while pH ranged from
 7.7 to 8.9, but was generally greater than  8.0. Total
 alkalinity fluctuated between 170-340 mg/l CaCO3 at
 all  sites during the study period. Dissolved oxygen
 concentrations were relatively high throughout  the
 study period, ranging from lows of approximately 4
 mg/l during ice cover to a high of almost 12 mg/l in the
 spring. Generally, dissolved oxygen  concentrations
 were greater than 6 mg/l.

 Loadings

 Bear River supplied most of the mass of nutrients to
 Dingle   Marsh,  however,  Bear Lake  contributed
substantial amounts  at Lifton Station, especially in
sample year I when less water came down the river
(Table 2). Bloomington Creek contributed a sizeable
portion of the orthophosphate loading to the marsh in
                          Table 1.—Parameters measured and methods of analyses.



Parameter
PH
Total alkalinity

Suspended solids
Orthophosphates

Total phosphorus


Ammonia
Nitrate


Nitrite

TKN



Total organic
carbon
Dissolved oxygen
(at laboratory)
Chlorophyll a
1Solorzano (1969)
2D'Eha et al (1977)
3Nydahl (1978)
4Solorzano and Sharp (1980)
5Smart et al (1981)
6Adams et al. (1981)



Method
Potentiometric, electrode1
Titrimetric, manual
colorimetric
Gravimetric
Colorimetric, automated
or manual
Manual acidic digestion,
colorimetric, automated
or manual
Manual phenate
Automated cadmium reduction
plus diazotization
colorimetric
Automated diazotization
colorimetric
Peroxydisulfate: manual
digestion, automated
nitrate via cadmium
reduction
Oxidation

Modified Winkler

Fluorometric







U.S.
EPA
(1979)
150.1
310.2

160.2
365.1
365.3
365.1
365.3


353.2


353.2

(353.2)



415.1

360.2








Reference
Stand.
Methods
(1980) Other
p. 402
p. 253

p. 94
p. 420

p. 413
p. 420

1
p. 376


p. 376

(p. 376) 2,3,4,5,6





p. 390

p. 952






                                               224

-------
                                            Table 2.— Contributions of constituents from different sources to loadings in Dingle Marsh, Idaho in Sample Years I
                                                                    (April 24, 1981 to April 24, 1982) and II (June 23, 1981 to June 23, 1982).
                                                                          Figures in parentheses are percents of total contributions.
                                                                                                   Constituent
ro
ro
en
Sample
Site Year
Rainbow Canal 1
at Stewart Dam II
Ream-Crockett 1
Canal II
Bloomington 1
Creek II
Spring 1
Creek II
St. Charles 1
Creek (upper arm) II
Lifton 1
Station3 II

Lifton 1
Station II
Lifton 1
Causeway II
Outlet 1
Canal II
Total entering 1
Marsh (kg/yr) II
Total leaving 1
Marsh (kg/yr) II
Net difference13 1
(kg/yr) II
Percent (retained 1
or released)0 II
Suspended
Solids Ortho- Total
( x 10') Phosphates Phosphorus
12.28(83.6)
49.14(93.5)
0.21(1.4)
0.51(1.0)
0.07(0.5)
0.38(0.7)
0.21(1.4)
0.25(0.5)
0.02(0.1)
0.37(0.7)
1.91(13.0)
1.91(3.6)

3.11(21.0)
7.75(33.3)
0
5.23(22.4)
11.72(79.0)
1031(44.3)
14.70
52.56
14.84
23.29
+ 0.13
- 29.27
+ 0.9
-55.7
1123(64.9)
3464(75.8)
69(4.0)
79(1.7)
158(9.1)
416(9.1)
23(1.3)
34(0.7)
28(1.6)
245(5.4)
329(19.0)
329(7.2)

669(42.2)
909(64.3)
0
136(9.6)
916(57.8)
369(26.1)
1730
4567
1585
1414
-145
-3153
-8.4
-69.0
21965(82.6)
65200(90.9)
442(1.7)
598(0.8)
441(1.7)
1552(2.2)
194(0.7)
255(3.6)
179(0.7)
734(1.0)
3367(12.7)
3367(4.7)

8336(33.0)
13178(39.9)
0
5732(17.4)
16924(67.0)
14114(42.7)
26588
71706
25260
33024
-1328
-38682
-5.0
-33.9
Ammonia
Nitrogen
6356(48.1)
10240(57.3)
358(2.7)
383(2.1)
431(3.3)
786(4.4)
94(0.7)
123(0.7)
70(0.5)
446(2.5)
5904(44.7)
5904(33.0)

6500(36.7)
7936(40.1)
0
1519(7.6)
11227(63.3)
10347(52.3)
13213
17882
17727
19802
+ 4514
+ 1920
+ 34.2
+ 10.7
Nitrate Nitrite
Nitrogen Nitrogen
25560(78.2)
47500(75.6)
1330(4.1)
1480(2.4)
1240(3.8)
5620(8.9)
730(2.2)
1290(2.1)
720(2.2)
3870(6.2)
3100(9.5)
3100(4.9)
Leaving Marsh
17420(61.5)
22490(61.7)
0
3710(10.2)
10920(38.5)
10246(28.1)
32680
62860
28340
36446
-4340
-26414
-13.3
-42.0
676(72.5)
1780(83.4)
27(2.9)
33(1.5)
26(2.8)
72(3.4)
5(0.5)
8(0.4)
6(0.6)
48(2.2)
193(20.7)
193(9.0)
(kg/yr)
587(55.6)
833(58.7)
0
136(9.6)
468(44.4)
449(31.7)
933
2134
1055
1418
+ 122
-716
+ 13.1
-33.6
Total
Soluble
Inorganic
Nitrogen
32600(69.6)
59520(71.8)
1710<3.7)
1900(2.3)
1700(3.6)
6480(7.8)
830(1.8)
1420(1.7)
790(1.7)
4370(5.3)
9200(19.6)
9200(11.1)

24510(52.0)
31260(54.2)
0
5370(9.3)
22620(48.0)
21040(36.5)
46830
82890
47130
57670
+ 300
- 25220
+ 0.6
-30.4
TKN
x10*
6.57(56.8)
21.25(78.3)
0.22(1.9)
0.31(1.1)
0.24(2.1)
0.64(2.4)
0.08(0.7)
0.11(0.4)
0.03(0.3)
0.42(1.6)
4.42(38.3)
4.42(16.3)

5.96(23.1)
9.98(26.8)
0
7.38(19.9)
19.85(76.9)
19.83(53.3)
11.55
27.15
25.82
37.19
+ 14.26
+ 10.05
+ 123.5
+ 37.0
Total
Nitrogen
xlO4
9.18(61.6)
26.18(77.8
0.35(2.4)
0.46(1.4)
0.37(2.5)
1.21(3.6)
0.16(1.0)
0.24(0.7)
0.10(0.7)
0.82(2.4)
4.75(31.9)
4.75(14.1)

7.77(27.0)
12.32(30.9)
0
7.77(19.5)
20.99(73.0)
19.83(49.7)
14.91
33.65
28.76
39.91
+ 13.84
+ 6.26
+ 92.8
+ 18.6
Total
Organic
Carbon
x10*
103.62(65.2)
289.80(81.1)
2.87(1.8)
4.02(1.1)
2.46(1.5)
7.58(2.1)
0.91(0.6)
1.30(0.4)
0.66(0.4)
6.23(1.7)
48.48(30.5)
48.48(13.6)

92.74(38.6)
150.63(39.2)
* 0
102.57(26.0)
147.55(61.4)
131.33(34.2)
159.00
357.41
240.29
384.53
+ 81.30
+ 27.12
+ 51.1
+ 7.6
                     aThe amount contributed to Dingle Marsh from Bear Lake is identical for both sample years because water was flowing from Bear Lake and into the marsh only
                     during the period of overlap of the sample years.

                     "Positive indicates that more left the marsh than entered while negative indicates that more entered than left

                     °Positive indicates that a greater percent left the marsh than entered while negative indicates the percentage of entering constituent retained by the marsh.
                                                                                                                                                                                                                O
                                                                                                                                                                                                                GO
                                                                                                                                                                                                                m
                                                                                                                                                                                                                I
                                                                                                                                                                                                                m
                                                                                                                                                                                                                CD
                                                                                                                                                                                                                -a
                                                                                                                                                                                                                3D
                                                                                                                                                                                                                O

                                                                                                                                                                                                                m
                                                                                                                                                                                                                O

-------
LAKE AND RESERVOIR MANAGEMENT
 both sample years (9 percent) and nitrate loading in
 sample year II (9 percent). The upper arm of St. Charles
 Creek contributed another 5 percent of the orthophos-
 phates and 6 percent of the nitrates in sample year II.
 Bear  River  and  Bear  Lake,  however,  contributed
 greater than 90 percent of the mass of all constituents
 to the marsh in  both years, except  for orthophos-
 phates, nitrates, and total  soluble inorganic nitrogen,
 which reflects nitrate loadings.
  Generally,  less than half  the mass of the consti-
 tuent leaving the marsh entered Bear Lake in sample
 year I,  while in sample year  II the opposite was true;
 usually more than half entered Bear Lake. For exam-
 ple, in sample year I, only  21  percent of the  total
 suspended solids  leaving the  marsh  entered  Bear
 lake; the remainder left the marsh through the Out et
 Canal and flowed down the Bear River. In sample yeiar
 II, about 56 percent of the TSS being exported from the
 marsh entered Bear Lake. This can be attributed to t fie
 greater volume of water entering Bear Lake in sample
 year II, especially with the  utilization of the causeway
 structure.

 Mass  Balances

 The  spatial  and temporal  patterns  of  the  mass
 balance  parameters  measured  within  the  marsh
 system fell into three general patterns for the study
 period. They were: (1) the marsh acted as a net sourse
 in both sample years, (2) the marsh acted as a net shk
 in both sample years, and (3) the marsh acted as a net
 source in sample year I and a net sink in sample year
 II. TOG, TN, TKN, and NH3-N  fell into the first pattern
 since the marsh acted as a net source for these con-
 stituents in both sample years (Table 2). Although the
 amounts entering the marsh were greater in sample
 year II, the percent exported by the marsh was greater
 (two to three times  greater in magnitude) in the dry
year (sample year I) than in the wet year (sample year
 II).
  During early spring  of 1981 (April and May), Dingle
 Marsh alternately acted as  a net sink and a net source
for TOC, TN, TKN, and NH3-N; however, the period of
greatest net export from the marsh occurred during
 late spring and early  summer, especially during the
                              months of June, July, and August of 1981 (Fig. 2, 3).
                              The marsh continued to act as a net source for these
                              constituents throughout the fall and winter of sample
                              year I, although not to the extent  it did  during late
                              spring and summer. The period of greatest retention
                              of these constituents by  the  marsh occurred from
                              April through June of 1982, although the marsh had
                              again been alternately acting as a net sink or source
                              just prior to this period.
                                The second pattern was exemplified  by TP, P04-P
                              and NO3-N. During both sample years  these consti-
                              tuents were removed by the marsh with 20 to 30 times
                              more  phosphorus being removed in the second year
                              (wet year) than in the first. The percentage of these
                              constituents retained by the marsh was much higher
                              in sample year II, probably because of the greater
                              amounts entering the marsh.
                                Dingle Marsh alternately acted  as  a  weak  net
                              source and sink for TP, PO4-P, and NO3-N in early
                                                                  SOURCE
                                      AMJ  JASON  DM
                                              1981     TIME
                                                     f 'M  AM  J
                                                      1982
                              Figure 3.—Mass balance diagram for total nitrogen showing
                              net difference between input and output in kg/day for Dingle
                              Marsh. Source  indicates more TN is released from than
                              enters the marsh, and sink indicates that more TN enters
                              than leaves the marsh.
                                    SOURCE
 100,000
        AMJ
                J  A  S
                 1981
'O'N ' D I J
   TIME
F M  A
 1982
                                           M  J
Figure 2.—Mass balance diagram for total organic carbon
showing net difference between input and output in kg/day
for Dingle Marsh. Source indicates more TOC is  released
from  than enters the marsh, and sink indicates that more
TOC enters than leaves the marsh.
                                                                                       SOURCE
                                                        1000-
A'M 'J'J'A'S'O N  D I J
         1981     TIME
F M  A ' M' J
 1982
                             Figure  4.—Mass  balance diagram for total phosphorus
                             showing net difference between input and output in kg/day
                             for Dingle Marsh. Source indicates more TP is released from
                             than enters the  marsh, and sink  indicates that more TP
                             enters than leaves the marsh.
                                                 226

-------
                                                                      CASE STUDY: THE BEAR LAKE PROJECT
 spring of 1981; however, it acted as a net source for TP
 in late spring and summer (Fig. 4). The mass of PO4-P
 entering the marsh roughly equaled the amount  leav-
 ing during the same time period (Fig. 5). The marsh
 acted as a net sink for these three constituents in the
 spring of  1982, beginning in February and continuing
 through the duration of the study period. It was a net
 source for NO3-N from June through August of 1981,
 but was a strong net sink from March through June of
 1982 and  was roughly balanced during late summer
 and winter.
   The third pattern observed in the  marsh was follow-
 ed by TSIN, N02-N, and TSS. The marsh acted as a net
 source for these constituents in sample year I, but
 was a net sink for them in sample year II. Although the
 calculations  indicate that less than 1 percent more
 TSIN and TSS left the marsh than  entered in sample
 year I, the figures  are well within the range of experi-
 mental error.  The marsh did appear to be a net source
 for nitrates in sample year I, however, releasing 13 per-
 cent more nitrates than it received. In  sample year II,
 the marsh clearly acted as a net sink for all three con-
 stituents,  retaining  about  31 percent of the TSIN.
 Nitrates made up  more than 70 percent of the TSIN
 entering the marsh in both sample years, but decreas-
 ed to about 60 percent of the TSIN  leaving the marsh.
 Ammonia, which made up more than 22 percent of the
 TSIN entering the marsh, increased to more than 35
 percent of the outgoing TSIN. Nitrites made up less
 than 3 percent of the entering and leaving TSIN. It ap-
 peared that the  pattern displayed  by TSIN in Dingle
 Marsh  was  determined  by NH3-N and  NO3-N
 dynamics. Since the marsh was a  strong net source
 for ammonia  in sample year II and a strong net sink for
 nitrates in sample  year II,  TSIN mass balances
 reflected these patterns (Fig. 6).
   TSS  were  flushed  from  the marsh from June
 through August of 1981, but were settling in the marsh
 from September 1981 until the end of the study period
 (Fig. 7). When the heavily silt ladened Bear River water
 was allowed to flow through the marsh and into  Bear
 Lake, much of the total suspended solids settled in
 the marsh, however, when  Bear Lake  water, with its
 low suspended solids load, was flushed through the
                                             marsh  and into the Bear  River below the diversion
                                             dam, sediments were incorporated into the water col-
                                             umn and carried out of the marsh. In the former case,
                                             the marsh acted as a net sink for TSS while it acted as
                                             a net source in the latter case. The TSS mass balance
                                             adhered  to this  pattern very well: acting as a net
                                             source during most of the summer of 1981 when Bear
                                             Lake water was needed downstream and acting as a
                                             net sink when Bear River water flowed into the lake.
                                             CONCLUSIONS

                                             Dingle Marsh acted as a net sink for TP,  PO4-P, and
                                             N03-N in both sample years and for TSIN, N02-N, and
                                             TSS in sample year II. The marsh was a net source for
                                             TOC, TN, TKN, and  NH3-N in both sample years and
                                             100,000
                                             <
                                             Q
                                                                                 SOURCE
                                             100,000-
                                                     A'M'J'J'A'S  O  N  D I J  F 'M'A 'M' J '
                                                             1981     TIME      1982

                                            Figure 6.—Mass balance diagram for total soluble inorganic
                                            nitrogen showing net difference between input and output in
                                            kg/day  for Dingle Marsh. Source indicates more TSIN is
                                            released from than enters the marsh, and sink indicates that
                                            most TSIN enters than leaves the marsh.
                                   SOURCE
AM  J
               J  A  S  O  N D J J
                1981     TIME
Figure 5.—Mass balance diagram for total orthophosphates
showing net difference between input and output in kg/day
for Dingle Marsh. Source indicates more PO4-P is released
from than enters the marsh, and sink indicates that more
PO4-P enters than leaves the marsh.
                                                                                SOURCE
                                                            A M
J'A'S'O'NDIJ
 1981    TIME
F 'M  A
 1982
                                                                                              M  J
                                            Figure 7.—Mass balance diagram for total suspended solids
                                            showing net difference between input and output in kg/day
                                            for Dingle Marsh. Source indicates more TSS is  released
                                            from than enters the marsh, and sink indicates that more
                                            TSS enters than leaves the marsh.
                                                 227

-------
LAKE AND RESERVOIR MANAGEMENT
for TSIN, NO2-N, and TSS in sample year I. Since the
difference in input and output for TSIN and TSS was
so small in sample year I, one can only conclude that,
on an annual basis, these constituents were roughly
in equilibrium. They  were not in equilibrium  on a
seasonal basis, however, since huge  differences ex-
isted between inputs and outputs. Indeed, the marsh
alternately acted as a net source and sink for all con-
stituents measured depending on the season, amount
of flow, and direction of  flow (i.e.,  whether waler
entered the marsh from Bear Lake or vice versa).
  Several  processes for altering the concentration of
these constituents in the water column as the waler
moved through the marsh could have been at  wo'k.
One mode of action may have been the physical pro-
perties  associated with  reduced flows.  As  waler
entered the  marsh  and  flowed  through  emergent
vegetation,  velocities  were  reduced. This allowed
much of the TSS  and their associated nutrients (i.e.,
those nutrients bound to particulates in some manner)
to settle and become  part of the marsh sediments.
This occurred in the upper portion  of the marsh. When
flow  velocities were sufficient, particulate mater al
was transported to Mud Lake where it  has been  forT-
ing a delta as it settles in the slow  moving  water of t ie
lake.
   Biological  processes  were also at work.  Dense
periphyton  attached  to the  submerged  stems of
emergent  vegetation could have filtered much of the
dissolved  nutrients from the water column and  incor-
porated them into their cellular  components.  Also,
reduced flows and open water in the area of Mud Lake
allowed phytoplankton to flourish  which was probably
using PO4-P and  TSIN for growth and reproduction.
Furthermore, chlorophyll a, TN, and TOC concentra-
tions appeared to increase within the marsh, especial-
ly in the Mud Lake area. These increases were more
dramatic in sample year I  than II. It is believed that
these compounds may have been associated with
decaying  mats  of vegetation or  floating detritus
(phaeophytin and chlorophyll a) rather  than  viable
algal  cells  in 1981, whereas, they may have been
associated  with  a  plankton bloom  in  Mud  Lake
(chlorophyll a alone) in 1982.
   Many other physical, chemical, and biological pro-
cesses  could be operating in Dingle Marsh to  either
remove nutrients  from or add them to the water col-
umn. Decomposition of litter releases stored nutrients
into marsh waters in a relatively short period of time
although few conclusions can be  made, based on the
current  available data, concerning this  pathway in
Dingle Marsh. TOC and TN may have increased, how-
ever, as a result of this activity.
   It should be noted that, although Dingle Marsh acts
as  a seasonal and annual  net  source  for  some
nutrients, much of the mass of these nutrients left the
marsh through the Outlet Canal and flowed down the
Bear  River and away from Bear Lake. The amounts
leaving Dingle  Marsh at Lifton  Station  and  Lifton
Causeway and entering  Bear Lake were always less
than the amounts entering the marsh at Stewart Dam.
Thus, Dingle Marsh acted to  reduce the amount of
loading into Bear Lake regardless of its status as a net
source or sink. Based on the current data set,  it was
decided that Dingle Marsh could be managed to in-
crease its ability to reduce nutrient loading into Bear
Lake.
   One management strategy is to control the direc-
tion of water flowing through the marsh so that most
nutrients are carried away from Bear Lake when the
marsh is acting as a strong net source.  Another
possibility calls for diking off a unit within the marsh
to act as a settling basin for TSS and a trap for dissolv-
ed nutrients. The unit would reduce flow velocities and
retain water long enough for biological  processes to
reduce nutrient concentrations.
REFERENCES

Adams, V.O., et al. 1981. Analytical procedures for selected
  water quality parameters. Utah Water Res. Lab., Logan.
D'Elia, C.F., P.A. Steudler, and N. Corwin. 1977. Determination of
  total nitrogen in aqueous samples using persulfate digestion.
  Limnol. Oceanogr. 22(4):760-4.
Nydahl, F.  1978. On the  peroxodisulfate oxidation of total
  nitrogen in water to nitrate. Water Res. 12:1123-30.
Smart, M.M., F.A. Reid, and J.R. Jones. 1981. A comparison of a
  persulfate digestion  and the Kjeldahl  procedure for deter-
  mination of total nitrogen in freshwater samples. Water Res.
  15:919-21.
Solorzano, L 1969. Determination of ammonia in natural water
  by the  phenolypochlorite  method.  Limnol.  Oceanogr.
  14:799^)1.
Solorzano, L, and J.H. Sharp. 1980.  Determination of total
  nitrogen in natural waters. Limnol. Oceanogr. 25(4):751-4.
Standard Methods for Examination of Water and Wastewater.
  1980. 15th ed. Arn. Pub. Health Ass., New York.
U.S.  Environmental Protection Agency.  1979.  Methods for
  chemical analysis of water and waste. EPA-600/4-79-019. En-
  viron. Monitor. Support Lab., Cincinnati, Ohio.
                                                  228

-------
 THE EFFECT OF COPRECIPITATION OF CaCO3 AND
 PHOSPHORUS ON THE TROPHIC STATE  OF BEAR  LAKE
 PAUL BIRDSEY
 Fisheries and Wildlife Department
 Utah State University
 Logan, Utah

 VINCENT LAMARRA
 Ecosystem  Research Institute
 Logan, Utah

 V. DEAN  ADAMS
 Division of Environmental Engineering
 Utah State University
 Logan, Utah
             ABSTRACT

             Bear Lake is a hardwater lake located in a limestone basin on the border of Utah and Idaho with a
             surface area of 282 km*, maximum depth of 63 m, and a mean depth of 10 m. The lake was formed
             by tectonic activity approximately 28,000 years B.P. with no natural outfall. Inflow mainly from small
             tributaries probably did not equal evaporation. These conditions resulted in a concentration of car-
             bonate salts and a unique macrochemistry, with magnesium as the predominant divalent cation. The
             isolation from any major drainages also led to the development of endemic fish species. In 1912, Utah
             Power and Light Company completed a series of canals diverting water from the Bear River into the
             lake during the spring and later released for downstream irrigation and power needs. Diversion of
             the Bear River into the lake increased water flow and presumably loadings by as much as 70 percent
             above historic conditions. A 2-year study recently completed found nutrient loadings into Bear Lake
             at meso-eutrophic levels, but the lake limnologically oligo-mesotrophic. Phosphorus was the principal
             limiting nutrient. Because of this apparent anomaly in trophic status and the known relationship bet-
             ween calcium carbonate and phosphorus in marl  lakes, this study was undertaken  to quantify the
             reduction of potential algal biomass through coprecipitation of phosphorus. Initially, three different
             phosphorus levels were added to synthetic Bear Lake medium without algae to determine if coprecipita-
             tion would occur under ideal conditions. The pH of the medium was raised artificially with NaOH to
             8.5, a value not uncommon in Bear Lake. After 4 days 100 percent of the phosphorus had precipitated
             in the 10 /4j P/l treatment. Bioassays were then conducted in softwater and Bear Lake media with
             Selenastrum capricornutum at 10 different phosphorus levels. At similar nutrient levels the maximum
             biomass reached twice that of the biomass in the Bear Lake medium. These may explain the low
             primary production experienced in many hardwater lakes and Bear Lake in particular. It may also
             be inferred that a potential self-cleansing mechanism exists within Bear Lake that would allow a rapid
             reversion to historic water quality if  nutrient loadings were reduced.
INTRODUCTION

A  recently completed 2 year study, has noted that
nutrient loadings into Bear Lake were at mesotrophic
levels, but the lake was limnologically described as
oligo-mesotrpphic. Phosphorus was identified as the
principal limiting  nutrient in  that  investigation. Be-
cause of the apparent anomaly in trophic status and
the known relationship  between calcium carbonate
and phosphorus in marl lakes, this study was under-
taken to quantify the reduction in potential algal bio-
mass through coprecipitation  of phosphorus.
MATERIALS & METHODS

Water samples were collected from one limnetic sta-
tion and two littoral sites from April 1981 through June
1982.  Samples  were taken  biweekly  April through
August of each year and monthly during the remainder
of the study period. Originally, four other littoral sites
were sampled at similar time intervals, but were dis-
continued in March  1982 when  statistical  analysis
showed  no significant  difference between littoral
sites.
   Limnetic samples were collected at seven depths:
surface, 10m, 20m, 30m, 40m, 50m, and bottom with an
8 liter PVC Van Dorn water bottle. The samples were
chilled and transported to the Utah Water Research
Laboratory for analysis.  Field measurements  were
made for temperature and conductivity using a Yellow
Springs Instrument  Model 33 Conductivity-Salinity-
Temperature Meter. All laboratory analyses were per-
formed using Standard Methods (1980).
   Bioassays were conducted according to the guide-
lines provided by Miller et al. (1978).  The Bear Lake
medium  used was  slightly  modified from  Werner
(1982).  A softwater medium  was used to determine
                                                229

-------
LAKE AND RESERVOIR MANAGEMENT
potential biomass at each nutrient concentration and
constituent concentrations were  reduced  by 80 per-
cent of those in the Bear Lake medium.
  Algal  assays were conducted at 10 different orthc-
phosphate levels from 9 to 85 ^g/l  with five replica-
tions at each level. The nitrogen to  phosphorus ratio
was kept constant at 25:1 at all phosphorus levels. The
assays  used Selenastrum capricornutum obtained
from a sterile culture and were run until the fluoresc-
ence of  each flask changed less than 10 percent in a
24 hour period. This was considered  maximum bic-
mass.
  Assays without algae were conducted using Bear
Lake medium at five orthophosphate levels  and two
controls with three replicates for each treatment and
control.  The five levels were 15, 30, 50, 70, and 80 ng/l;
the controls were run at 15 and 80 ^g/l. Treatment:;
consisted of artificially raising the pH in the flasks by
suspending a 50 ml beaker of CaO  over the medium
and then  sealing the flask to the atmosphere. This
method  raised the pH to 8.7 to 9.0, values  not un-
common in Bear Lake.
  Following termination of the assays  using Bear
Lake  medium, the  flasks  were rinsed  with doubly
distilled  water (DDW) and then returned  to their
original  volume with DDW, then 1 ml  of  12 NH2SC>4
was added. This procedure dissolved the  precipitate
and allowed determination of a percent recovery of the
precipitated phosphorus.

RESULTS

Current  Limnology of Bear Lake: A summary of all
water quality and biological parameters can be seen
in Table 1 for the 2 years of the study (1981-82).
  Physical Data: Bear Lake appeared to be dimictic,
with a spring circulation occurring from April to June
(Fig. 1). The lake remained stratified from July through
November. Fall circulation started during October and
was completed  by  the middle  of December. Ice
covered the lake for approximately 6 weeks  in tho
winter of 1981-82. Maximum epilimnetic and  hypolim-
netic temperatures were 21.5°C and 4°C respectively.
Secchi disk transparency ranged from  2.8 to 6.6  m
with a mean value of 4.7 m. Variations corresponded
to the wax and wane of algal biomass and total sus-
pended solids. The vertical distribution of PAR (photo-
synthetically active radiation) indicated that the 2 per-
cent  light  level  was characteristically  at the 20  m
depth. These data indicated that the Secchi disk trans-
parency did not correspond to the 1 percent light level.
Light scattering due to suspended calcium carbonate
particles may have been the major cause.
  Chemical and  Biological Data: Surface  dissolved
oxygen concentrations for the  most part were at  or
above 100 percent saturation.  However, throughout
both summer stratification periods, the hypolimnion
of the lake experienced a marked decrease in oxygen

          	  OXYGEN mg/L
  10-

  20-

 iso-
  50-


 BOT
SURF.
           j  ' j  » TS'O'N'D'J'F'II'A'M'J
             ioa<         MONTHS         ««n.»
                    TEMPERATURE "C
  10-


  20-


 £30-


 |40-


  50-


 BOT.
                          N ' D
                        MONTHS
M ' A ' M

1982
Figure  1.—The temperature  (°C) and oxygen  (mg
isopleths for Bear Lake during 1981 and 1982.
  Table 1.—The summary of the results from the physical, chemical, and biological parameters at the limnetic station for
                                             1981 and 1982.
                                                                 Year
                                     1981 (4-24-81 to 12-3-81)
                1982 (1-19-82 to 6-22-82)
Parameter
Temperature (°C)
Secchi disk (m)
Oxygen (mg/l)
pH (units)
Total alkalinity (mgCaCOs/l)
Turbidity (NTU)
Suspended solids (mg/l)
P04-P (Mg P/l)
Total P (HQ P/l)
NH3-N (Mg N/l)
NO3-N (mg N/l)
NO2-N (^g N/l)
Total nitrogen (mg N/l)
Conductivity (unhos/cm)
TOC (mg/l)
Chlorophyll a* fcg/l)
N
75
13
70
84
97
72
74
85
77
78
82
77
75
27
54
34
x ± S.E.
9.4 ±0.5
4.91 ±0.34
8.1 ±0.2
8.5 + 0.1
271 ±1
2.8 ±0.4
7.3+1.5
3.1+0.3
11 +.6
35 + 3
0.10 ±.04
3 ±0.1
0.51 + .08
593 + 15
4.18 + .24
0.59 ± .09
Range
2.8-21.5
2.9-6.6
3.8-12.2
8.4-8.8
243-338
0.7-18.0
0.0-56.0
0.0-17
2.0-30
6-146
<.04 + .360
0-104
.09-2.63
490-703
1.9-11.80
0 18-2.67
N
48
8
48
55
56
56
55
55
55
55
56
56
56
45
54
36
x±S.E.
5.2 ± 0.5
4.35 ± 0.36
10.8 ±0.2
8.68 ±0.1
262 + 2
1.9 + 0.1
3.07 ± .83
2 ± 0.2
11 +1.3
16±1
.031 ± .003
2 ±.01
.23 ±.007
420 + 8
3.91 ±0.11
0.7 + .04
Range
0-15.7
2.8-6.1
9.3-14.4
8.3-9.0
244-348
0.1-5.7
0-6.4
0-5
2-50
5-33
0-.097
0-4
0.13-0.43
367-600
2.1-6.90
0.18-1.61
 •Surface, 10, 20, and 30 meter depth stations only
                                                 230

-------
                                                                         CASE STUDY: THE BEAR LAKE PROJECT
 (Fig. 1). In December 1981 the rate of oxygen loss re-
 sulted in an oxygen concentration of 3.8 mg/l near the
 bottom of the hypolimnion. Also during this period of
 stratification, dissolved oxygen concentrations reach-
 ed 110-125  percent of saturation at approximately 20
 m in depth. This metalimnetic maximum correspond-
 ed to high levels of chlorophyll a (Fig. 2) and the 2 per-
 cent light level. Algal biomass  was highest at the 30
 meter depth in 1981  (2.5 ±0.40 ^g chla/1).
    Surface concentrations of orthophosphate, nitrate,
 and ammonia corresponded to turnover and the tem-
 poral distribution  of chlorophyll a.  Maximum  con-
 centrations  within the epilimnion (Fig. 3 and 4) during
 July and August were  reduced to  below detectable
 limits while algal biomass reached a maximum (Fig.
 5). Chlorophyll a values had typical spring and fall
 peaks  with  the fall peak in 1981 significantly higher
 than the following spring peak (1982). One possible
 cause  for the fall peaks being higher than the spring
 chlorophylls may  be the accumulation of nutrients
 during the  summer stratification. This accumulation
 of nutrients corresponded to the low oxygen concen-
 SURFACE


       10


       20


       30


      40


      SO-


 BOTTOM-
                CHIS IIQ/L                TEMP *C

         0  0.50  140 1.50 2.00 2.50 3.00 2015 10 5 0
7-21-81
                        1981
 Figure 2.—The vertical concentrations of chlorophyll a
 ± S.E.) for Bear Lake during 1981 and 1982. Data for 1981
 was from April to December. Data for 1982 was from January
 to July.
                  ORTHO PHOSPHATE UQ/L
  10-

  2O-


 £30-


 f-H

  50-
     A ' H ' J'J'A'S'O'N'D'J'F'M'A'M'J
                  TOTAL PHOSPHORUS ug/L
  10-


  20-


 £30-
 X
     A'M'J'J'A'S'O'N'D'J'F'M'A'MJ'J
             iui         MONTHS         ,M2

Figure 3.—The orthophosphate and total phosphorus (^g P/l)
isopleths for Bear Lake in 1981 and 1982.
 trations.  Winter oxygen  levels during stratification
 were not  reduced to this extent.
   The total  phosphorus  concentration  within  Bear
 Lake had both spatial and temporal differences. Sur-
 face concentrations  were lower during  1981 (10-14
 ^g/l) but higher within the hypolimnion (40-100 pg/l).
 However, the opposite pattern was observed during
 1982 with the surface concentrations (10-50 /^g/l) be-
 ing higher than the hypolimnion (15-30 ^g/0 concentra-
 tions. Sampling did not continue through fall turnover.
   Bioassays: Four sets of bioassays were conducted,
 both with and without Selenastrum capricornutum in
 an effort to define calcium carbonate coprecipitation
 with phosphorus under ideal physical conditions and
 in competition  with algae.  Furthermore,  the assays
 performed with  algae were designed to measure the
 reduction in algal  biomass  due to precipitation  of
 phosphate with  CaCO3.
   Initially, three different phosphorus levels were add-
 ed to Bear Lake medium without algae and the pH was
 adjusted with NhOH to 8.5. After 4 days, 100 percent of
 the phosphorus had precipitated in the 10 ^g/l treat-
 ment. To  determine the rate of phosphorus removal
 the experiment  was repeated using  five  phosphorus
 levels; CaO  was used to withdraw C02 from the
 system to elevate  the pH and eliminate any inter-
 ference from NaOH  addition.  This experiment re-
 vealed an average precipitation rate of 0.52 ^g P/l/hr
 with a range of  0.28 to 0.7 ^g P/l/hr for the 15 and 80
 ng/l treatments respectively.
   Algae grown in the  softwater medium achieved ap-
 proximately twice the  biomass of the algae in the Bear
 Lake medium at all phosphorus levels (Fig. 6). Total
 suspended solids were also measured and an average
 difference of 5 mg/l was observed between the total
 and  volatile  suspended  solids  in  the   Bear  Lake
 medium. No difference in total and volatile suspended
 solids was observed in the softwater medium.

 DISCUSSION

 Historically,  Bear Lake  has been described  as  a
typical oligotrophic system with a well-oxygenated
hypolimnion,  low nutrient concentrations,  and low pri-
mary production (Nyquist, 1967). However, this study
                                                                  TOTAL SOLUBLE INORGANIC NITROGEN Ug/L
                                                          A'M'J'J'A'S'O'N'O'J '  r '  u ' A ' M '  J
         Figure 4.—The total soluble inorganic nitrogen (NO3 + NO, +
         NHj) and total nitrogen Ojg N/l) isopleths for Bear Lake durina
         1981 and 1982.
                                                  231

-------
LAKE AND RESERVOIR MANAGEMENT
found an apparent change in the trophic status to an
oligo-mesotrophic  level with many characteristics of
an eutrophic system. Total  phosphorus levels often
exceeded the 10 ^g/l level characteristically used as
the breakpoint between  oligo and  mesotrophy.  The
average summer total phosphorus concentration has
been used as an indicator of trophic status by several
authors (e.g. Carlson, 1977) and in Bear Lake this value
was 14-21 ngl\  which was within  the  mesotrophic
range. Total  phosphorus loading values are also at
mesotrophic  levels (Vollenweider, 1976)  especially in
the high water year of 1982.
  Another indicator of eutrophication was  the hypo-
limnetic oxygen deficit values of 0.041 mg C>2/cm2/dc,y
in 1981, and 0.043 mg O2/cm2/day in 1982, indicating
mesotrophic  conditions.  These values were believed
to be  more  of  a response to high allochthonois
organic  carbon  loading  (Lamarra et al. 1983) than
autochthonous production, but,  according to Lind
(1971)  can  still  be used   as  indicators  of  eutro-
phication.
  Lamarra et al. (1982) noted that the trophic state in-
dex (TSI) values calculated for seven areas  in Bear
Lake over a 4-year period were consistently different
for total phosphorus, Secchi disk, and chlorophyll a.
Total  phosphorus  and Secchi disk transparency gave
TSI values significantly higher  than those  values
calculated for chlorophyll a. This apparent discrepan-
   2.0

   1.3

 f,.oJ
 •I
    .5
19fll
                                 1982
          J  FMAMJJASON  D

                      MONTHS

Figure 5.—The temporal distribution of chl a (^g/l ± S.EL;
N= 3) for the photic zone in Bear Lake during 1981 and 1982.
Values are means for the surface, 10, and 20 meter limnetic
stations.
 - 20


 I
                      40    50    u    n
                     initial n^-f (u r/i>
 Figure 6.—The biomass of algae (Volatile Suspended Solids)
 produced in a hard water Bear Lake media and a soft water
 media from different initial levels of phosphate. Bars repre-
 sent 95 percent confidence interval with N = 5.
        cy between the chemical and biological data indicated
        a possible interference with algal nutrient uptake,
        specifically the limiting nutrient phosphorus.
          Other workers who have studied the coprecipitation
        of calcium carbonate and phosphorus (Ferguson et al.
        1973;  Rossknecht,  1980; Murphy et al. 1983) have all
        dealt with total  phosphorus concentrations of greater
        than 100 \IQ!\. It was noted that the major precipitate in
        those systems  was  calcite; however, in Bear Lake
        aragonite (Davidson, 1969) has  been the dominant
        form.  It was  not  known, therefore,  if calcium  car-
        bonate precipitation was having a major role in  reduc-
        ing  the potential  maximum production of algae in
        Bear Lake. Figure 6 suggests that the coprecipitation
        of phosphorus may have been a factor in determining
        the  level of algal productivity in hardwater, low phos-
        phorus systems such as Bear Lake. The relationship
        between phosphorus removal and algal biomass was
        even more apparent when the data from the dissolved
        precipitate were investigated. For example, 32 ^g P/l
        was precipitated from the Bear Lake medium with the
        highest phosphate  level  leaving 52 ^g/l available phos-
        phorus. The resulting biomass of 30 mg/l volatile sus-
        pended solids approximately equaled the 40 mg/l bio-
        mass obtained  in the 46 ^g P/l phosphate addition in
        the  softwater medium.
CONCLUSIONS

Bear  Lake has  apparently changed from an oligo-
trophic to an oligo-mesotrophic system within the last
15 years. This change has been the result of increased
recreational use within the basin and changing land
use patterns along the Bear River drainage. The lake
now has an average summer total  phosphorus con-
centration of 14-21 /ug P/l and meso-eutrophic hypo-
limnetic  oxygen deficits.  However, the  chemical
changes within the lake have only partially increased
phytoplankton production thus far. This lack of direct
response to increased nutrient concentrations may be
partially  due to coprecipitation  of  phosphorus with
calcium  carbonate.  Bioassays   conducted  without
algae have resulted in an average rate of phosphorus
loss  of  0.52  /^g/l/hr.  Assays  conducted  with
Selenastrum  capricornutum revealed that  algal  bio-
mass was reduced by approximately 50 percent in
hardwater versus softwater media and that this reduc-
tion may have been related to the loss of phosphorus
through precipitation.
  These results present a possible mechanism to ex-
plain  the low primary productivity in many  hardwater
lakes and Bear  Lake in particular. It may also be in-
ferred from  these  results  that a  potential self-
cleansing mechanism exists within Bear  Lake  that
would allow rapid reversion to historic water quality if
nutrient loadings were reduced.

REFERENCES

Carlson, R.E. 1977. A trophic  state index for  lakes. Limnol.
  Oceanogr. 22:361-9.
Davidson, D.F. 1969. Some aspects  of geochemistry and
  mineralogy of Bear  Lake  sediments, Utah-Idaho. M.S.
  Thesis. Utah State Univ., Logan.
Ferguson, J.F., D. Jenkins, and J. Eastman. 1973.  Calcium
  phosphate precipitation at  slightly  alkaline  pH values. J.
  Water Pollut. Control Fed. 45:620-31.
Lamarra, V.A., D. Lentz, and C. Thomas.  1982. The physical,
  chemical, and biological effects of large marinas on the lit-
  toral zone of Bear Lake, Part I. Limnological Conditions.
  Bear Lake Region. Comm.
                                                  232

-------
Lamarra, V.A., et al. 1983. A diagnostic-feasibility 314 Clean
  Lakes Study for Bear Lake and its watersheds. Final  Rep.
  to Bear Lake  Region. Comm.
Lind, O.T. 1971. The organic matter budget of a central Texas
  reservoir.  Pages  193-202  in  G.E.  Hall,  ed. Reservoir
  Fisheries and Limnology. Am. Fish. Soc. Spec. Publ. No. 8,
  Washington, D.C.
Miller,  W.E.,  J.C. Greene, and  T.  Shiroyama.  1978.  The
  Selenastrum capricornutum Printz algal assay bottle  test.
  Experimental design, application, and data interpretation
  protocol. EPA-600/9-78-018. U.S. Environ.  Prot. Agency,
  Corvallis, Ore.
Murphy, T.P., K.J. Hall, and I. Yesaki. 1983. Coprecipitation
  of phosphate with calcite in a naturally eutrophic  lake.
  Limnol. Oceanogr. 28:58-69.

Nyquist, D. 1967. Eutrophication trends of  Bear Lake, Idaho-
  Utah and their effect on the distribution and biological pro-
  ductivity  of zooplankton.  Ph.D. Diss.  Utah  State Univ.,
  Logan.
                    CASE STUDY: THE BEAR LAKE PROJECT

Rossknecht, V.H. 1980.  Phosphate limination durch autch-
  hone calcitfallung in Bodensee-Obersee. Arch. Hydrobiol.
  88:328-44.

Standard Methods for the Examination of Water and Waste-
  water.  1980. 15th ed. Am. Pub. Health Ass., Washington,
  D.C.

Vollenweider, R.A. 1976. Advances in defining critical loading
  levels for phosphorus in lake eutrophication. Mem. 1st.
  Ital. Idrobiol. 33:53-83.

Werner, M.  1982. Reponses of freshwater ecosystems  to
  crude oil impaction. Ph.D. Diss. Utah State Univ., Logan.
                                                      233

-------
                                             Sediment  Analysis
SEDIMENT  METALS ACCUMULATION IN A
SUBURBAN LAKE
JOHN  D. KOPPEN
STEPHEN J. SOUZA
Princeton Aqua Science
New Brunswick,  New Jersey
           ABSTRACT
           Lake Hopatcong is a 1,087 ha lake located in the suburban New York metropolitan area in Sussex
           and Morris Counties, N.J., and is the head waters of the Musconetcong River in the Delaware River
           Drainage Basin. The lake watershed is 5,483 ha in area with 71 percent forested and 25 percent in
           high-density residential development that is clustered around the immediate shoreline of the lake.
           The recreational use of the lake is extremely heavy, with an excessive number of large motor boats.
           Also, stormwater from the residential areas empties directly into the lake via storm sewers and direct
           runoff. Preliminary analysis of selected sediment samples indicated substantial levels of metals (especial-
           ly lead) in the surficial sediments. As part of a Sec. 314 Lake Restoration study a series of 30 shallow
           sediment core samples were taken at various locations throughout the lake. These were analyzed
           for lead, aluminum, iron, zinc, mercury, cadmium and percent of solids. The results indicated signifi-
           cant concentrations of lead and zinc in the most recent sediments as compared to the background
           levels in the older and deeper sediments Also, the spatial distribution of sediment metals within the
           lake was investigated to attempt to identify their sources. The implications of these findings and manage-
           ment implications based on the information are discussed.
INTRODUCTION

Several authors have investigated the origin of lead
and zinc in surface waters and sediments. Lead in sur-
face water is assumed to originate primarily from the
use of leaded gasolines  in the internal combustion
engine (Kuzminski and Hogan, 1974). A major con-
tributor of zinc in the environment comes from auto-
mobile tires (Christensen and Guinn, 1979). The routes
by which high concentrations of  lead and zinc con-
taminate sediments of lakes and streams include sur-
face runoff from streets in urban-suburban areas, the
use of leaded gasoline in outboard motors, and atmo-
spheric fallout.
  Kuzminski and Hogan (1974) indicated that the use
of leaded gasolines in outboard motors was the major
source of lead in lake sediments.  However, Sartor et
al. (1974) and Whipple and Hunter (1977) found that the
most prevalent metals in street runoff were lead and
zinc. Also, in New Jersey, Wilbur and Hunter (1975) in-
vestigated the concentrations of heavy metals in  ur-
ban stormwater runoff and found lead and zinc to
predominate. Christensen and Guinn (1979) quantita-
tively related concentrations of lead and zinc in urban
runoff to the levels of zinc in automobile tires  and lead
in gasoline.
  Preliminary analysis of selected sediment  samples
from Lake Hopatcong, New Jersey indicated substan-
tial levels of lead and zinc in the surficial sediments.
As part of a Section 314 Lake Restoration Study, a
series of 35 shallow sediment cores was taken at
various locations throughout the lake. Thirty of these
                                              235

-------
LAKE AND RESERVOIR MANAGEMENT
cores were stratified by depth. The top strata, at the
water-sediment interface, and the deepest strata a:
the bottom of each core were analyzed for lead, zinc,
aluminum,  iron, mercury,  and cadmium.  The results
for lead and zinc are reported here since  these could
be most easily related to man's activities in the water-
shed. The objectives of these analyses were to deter-
mine  overall concentration  of metals in  the  Lake
Hopatcong sediments, their spacial distribution, and
whether the lead and zinc accumulation  in the sedi-
ments could  be related, through  time,  with  the
development in the  watershed and the lake's recrea-
tional  use.  Also,  the suitability  for  disposal  o"
materials to be dredged was  evaluated.
MATERIALS AND METHODS

Lake sediment samples were collected from 35 loca-
tions within Lake Hopatcong (Fig. 1). These sites were
selected on the basis of their  proximity to marinas,
major sources of surface runoff, and areas of possible
future  dredging. Also, samples were  taken from the
open water areas away  from the intensive use areas
and away from shoreline influence.
  Core samples were taken by using a brass sediment
core sampler fitted with a plastic (cellulose acetate
butyrate) core tube liner. The depth  of the cores rang-
ed from 9 cm to 51 cm depending upon the firmness of
the sediments. The core samples were stored in an
upright position then returned  to the laboratory and
frozen  for preservation and ease of handling.
  Subsamples were taken by removing 2.54 cm from
the top and bottom of each core. These subsamples
were analyzed for  lead,  zinc, cadmium, iron, alumin-
um, and mercury as described in Standard Methods
(1980).  The results of the lead  and zinc analysis are
presented  here since lead and zinc  concentrations
were of particular  interest in this study. In addition,
four composite core samples were analyzed by the EP
toxicity procedure (Standard Methods, 1980) to deter-
mine the propensity for the lead to leach  into the
water from the sediment.
RESULTS

The results of the core strata analyses for lead and
zinc are given in Table 1 and summarized in Table 2. In
the surficial sediments a mean lead concentration
was  243 mg  kg-1 dry weight.  However,  the values
ranged f rom 12 mg kg -1 to 684 mg kg -1. (One val ue of
1,220 mg kg-1  was not included  in this average
because it came from a dredged area.) The mean lead
concentration of the bottom core strata was 19.4 mg
kg-1 dry weight. In relation to the value of this mean
the range (3-70 mg kg-1) is also large. The difference
between the mean surficial value of 243 mg kg-1 and
the mean  bottom  value  of  19.4  mg   kg-1   was
statistically significant (t  = 6.80).
  A similar pattern was observed in the zinc data. The
mean concentration of zinc in the surficial sediments
was 1,034 mg kg-1 dry weight with a range of  41  to
8,430 mg kg-1 compared to a mean concentration  in
the bottom core strata of 140.1 mg kg-1 with a range
of 15 to 544  mg kg-1. The difference between the
mean surficial value of  1,034 mg kg-1 and the mean
bottom  value of 140.1  mg kg-1  was statistically
significant (t =  5.07).
  It was expected that  sediment concentrations  of
lead would be the highest where considerable storm-
water enters the lake or where there was exceptional
motor boat traffic. Though this was generally the
case, the data was not definitive, as expected.  How-
ever, the general pattern is present.
  Four control stations (No.'s 12,15,16 and 23, Fig. 1)
were located  generally  away from the influences  of
stormwater discharges  and marina facilities.  The
mean surficial lead concentrations of these stations
were 52.2 mg kg -1 (range 11.6-95.3). The surficial lead
concentrations from the nearshore stations where the
influences of stormwater runoff  and excessive motor
boat  traffic would be felt were  314 mg kg-1 (range
43.1-1,220).
  The data on zinc distribution within the lake was
more definitive. The mean surficial zinc concentration
at the control stations  was 147.8 mg kg-1 (range
41-247 mg kg-1). The surficial zinc concentration  at
              ROXBURY
                                   MOUNT ARLINGTON
                                                                                 •EAVER BROOK
                                                           FEET

                                                           1000   200O
                                                        0   200  400

                                                          METERS
                                                        DEPTHS IN FEET
Figure 1.—Location of sediment core samples in Lake Hopatcong, N.J.
                                                236

-------
                                                                                     SEDIMENT ANALYSIS
the nearshore stations was 1,286.8 mg  kg-1 (range
124-8,430). As expected, the locations that received
stormwater runoff from  suburban land uses showed
considerably higher concentrations of zinc in the sur-
ficial sediments. It is interesting to note that core No.
30, taken in Ingram Cove, showed a surficial zinc con-
centration of 8,430  mg kg-1. This site receives two
direct stormwater discharges from Lakeside Drive, the
main  highway to Hopatcong Borough and the most
heavily travelled road in the lake basin.
  To determine the  leachability of lead into the water
column (leachate analysis for zinc was not done) an
EP toxicity analysis  of four composite sediment cores
was performed. These data are presented  in Table 3.
Under the conditions of the EP toxicity analysis (pH 5)
no significant amounts of lead appeared  to leach into
the liquid phase. The lead appears to be tightly bound
to the sediment particles and  is  relatively insoluble
under these conditions.
DISCUSSION

The accumulation of lead and zinc in the sediments of
Lake Hopatcong over time is evident from the data
presented. Thirty percent of Lake Hopatcong's 13,500
acre watershed is covered with  high and low density
residential  and  commercial development, most  of
which is in the immediate vicinity of the lake. Storm-
water from this developed area  is, for the  most part,
carried directly to the lake without detention or reten-
tion. The residential/commercial development has oc-
curred over the last 50 years. Comparison of the lead
and zinc concentrations from the top and bottom of
the 30 cores show a 12-fold increase in lead con-
centrations and a 7.5-fold increase in zinc concentra-
tions. The assumption that the lead and zinc concen-
trations at the bottom of the cores represent the pre-
development condition may not be the case, because
of the variation in the depth of the cores and the pro-
bable variation in rates of sedimentation from place to
place in the lake. However, the assumption that the
surficial strata were deposited later than the  deeper
strata is a reasonable  assumption  and  makes the
sediment surface to bottom comparison valid.
  The origin of the zinc in the surficial sediments can
be directly related to surface runoff from the streets in
the developed area. However, the origin of the lead in
the Lake Hopatcong sediments is not as clearly defin-
ed by the data. Lead probably comes from the use of
the internal combustion engine. However, it reaches
the lake sediments via three routes:  fallout from the
atmosphere, runoff from the streets, and the use of
leaded gasoline in outboard motors. Obviously, some
lead is contributed from all three sources, but the data
does allow for estimating the proportion from  each
route.
  The failure of the lead to leach into the liquid phase
during the EP toxicity extraction procedure indicates
that under the test conditions the lead either is bound
 Table 1.—Concentrations of lead and zinc in surficial and bottom strata from cores taken in Lake Hopatcong, N.J. under
       strata column (S) =  surficial, (B) = bottom and number is depth of bottom strata in core. All concentrations
                                         in mg kg -1 dry weight.
Core
No.
1
2
3
4
5
6
8
9
10
11
Strata
(cm)
S
B-30
S
B-27
S
B-27
S
B-24
S
B-29
S
B-30
S
B-24
S
B-26
S
B-32
S
B-29
Concentration
Lead
126
55
314
10
168
5
105
16
72
7
156
54
279
70
171
7
140
30
197
14
mg kg-1
Zinc
381
256
889
163
586
115
832
544
609
174
556
396
585
241
355
28
762
202
487
57
Core
No.
12
13
14
15
16
18
19
20
21
23
Strata
(cm)
S
B-23
S
B-30
S
B-30
S
B-28
S
B-20
S
B-43
S
B-48
S
B-24
S
B-19
S
B-9
Concentration
Lead
70
4
469
13
684
11
32
4
12
3
297
13
287
15
262
22
43
7
95
32
mg kg-1
Zinc
185
16
1730
52
2300
131
118
19
41
15
978
155
794
182
895
163
124
65
274
145
Core
No.
24
25
26
28
29
30
31
33
34
35
Strata
(cm)
S
B-17
S
B-28
S
B-33
S
B-34
S
B-29
S
B-32
S
B-51
S
B-29
S
B-20
S
B-22
Concentration
Lead
134
16
345
6
445
9
352
10
406
17
237
14
400
14
46
18
600
69
1220
1110
mgkg-1
Zinc
304
66
821
55
2010
130
1080
146
619
119
8430
131
984
36
833
78
1410
183
2450
1680
 Table 2.—Summary of lead and zinc data in core samples from Lake Hopatcong. All concentrations given in mg kg-1 dry
                                               weight.
Surficial Sediments
Statistic
N
Range
Mean
Standard Deviation
Depth
29
NA
NA
NA
Lead
29*
12-684
243
177
Zinc
29
41-8430
1034
1522
Depth
29
9-51 cm
28.9
8.6
Bottom of Core
Lead
29
3-70
19.4
18.8
Zinc
29
16-544
140.1
114.9
•Core No. 35 was eliminated from analysis.
                                                 237

-------
 LAKE AND RESERVOIR MANAGEMENT
  Table 3.—EP toxicity leachate analysis for lead in four composite cores compared to lead concentrations in sediments.
  Core Number
Lead Concentration
   In Sediment
 (mg Kg-1 dry wt.)
                                                                                         Lead Concentration
                                                                                             In Elutriate
        7

       22

       27

       32
       31.2
       9D.7

       2D.6
       13.0
  0.009

  0.026
  0.007

< 0.005
to the sediments or is insoluble. EP toxicity data for
other metals indicate that the dredged materials ars
suitable for disposal without any unusual precautions.
  The effect that the high concentrations of lead and
zinc have had on the biota of Lake Hopatcong or ths
Lake Hopatcong ecosystem is  not clear. Preliminary
fish tissue samples and water column samples have
not shown accumulations of lead and zinc. However,
the mobility  of these  metals in  the Lake Hopatcong
system is still being studied.
  The contribution of lead from outboard motors used
in Lake Hopatcong is a difficult problem to address;.
One of the major recreational activities on the lake is
motor  boating.  Approximately   15,000  boats  are
registered on the lake. The best answer may be  con-
verting outboard motors to unleaded gasolines.
  In managing  urbanizing watersheds the contribu-
tion of metals and other pollutants in nonpoint sur-
face runoff has to  be addressed. Currently consider-
able attention is being given to controlling the quality
of stormwater runoff. The work of Wanielista et a I.
(1982), Whipple and Hunter (1980), Whipple et al. (1981)
and Whipple (1981) have shown that passive treatment
of stormwater can substantially improve the quality of
stormwater.
  The Watershed  Management  Plan for  the  Lake
Hopatcong Basin (Lake Hop. Reg.  Plann. Board, 1983)
calls for stormwater quality management  as  an in-
tegral  part of the Plan.
             REFERENCES

             Christensen, E., and V. Guinn. 1979. Zinc from automobile tires
               in urban runoff. J. Environ. Eng. Div. Proc. Am. Soc. Civ. Eng.
               105: 165-9.

             Kuzminski, L, and W. Hogan. 1974. Heavy Metals in the Water
               and Sediment of Lakes in Western Massachusetts. III. Lead.
               Dep. Civ. Eng. Univ. Mass. Amherst.

             Lake Hopatcong Regional Planning Board. 1983. Lake Hopat-
               cong Management Study. Landing, New Jersey.

             Sartor, J., G. B. Boyd, and F. J. Agardy. 1974. Water pollution
               aspects of street surface contaminants. J. Water Pollut. Con-
               trol Fed. 46: 45&67.

             Standard Methods for the Examination of Water and  Waste-
               water. 1980.15th ed. Am. Pub. Health Ass., Washington, D.C.

             Wanielista, M.P., Y.A. Yousef, and J.S. Taylor. 1982. Stormwater
               management to improve lake water quality. EPA-600/2-82-048.
               U.S. Environ. Prot. Agency, Washington, D.C.

             Whipple, W., Jr. 1981. Dual purpose detention basins in storm-
               water management. Water Resour. Res. Bull. 17: 642-6.

             Whipple, W., Jr., and J.V. Hunter. 1980. Detention Basin Settle-
               ability of Urban  Runoff Pollution. Water Resour. Res.  Inst.
               Rutgers Univ., New Brunswick, N.J.

             Whipple, W., Jr., W.H. Clement, and S.D. Faust. 1981. Modeling
               of Alternative Criteria for Dual  Purpose Detention Basins.
               Water Resour. Res. Inst., Rutgers Univ., New Brunswick, N.J.

             Wilbur, W.G. and J.V. Hunter. 1975.  Pages 45-34 in Proc.  No. 20,
               Urbanization and Water Quality Control. Am. Water Resour.
               Ass. Washington, D.C.
                                                    238

-------
SEDIMENT INFLOWS AND WATER QUALITY IN AN
URBANIZING WATERSHED
DAVID F. BRAKKE
Institute for Watershed Studies
Western Washington University
Bellingham, Washington
             ABSTRACT

             Lake Whatcom is a large, deep, monomictic lake in the Puget lowlands of Washington State. Expan-
             ding development from the city of Bellingham and a large diversion of water from the Nooksack River
             to the lake are two recent impacts. Bellingham depends on the lake as its sole drinking and industrial
             water supply. Additionally, the lake is an important recreational resource. Monitoring of the lake began
             in 1962 and was continuous to 1972, and recent work began in  1979. These data and information
             from sediment cores have been used to determine trends in water quality. Historically, watershed
             uses were mainly logging and some coal mining. Sedimentation rates based on Pb-210 were about
             0.5 cm/yr. More recently, sedimentation rates have increased to 0.8-1.2 cm/yr., apparently due to in-
             creased runoff related to urbanization and the diversion of Nooksack River water containing glacial
             meltwater with high paniculate loads. Much of the paniculate material is sedimented in a 6.5 ha lake
             below the diversion tunnel (volume has decreased 20 percent since  1962), but very fine particulates
             are transported through this lake to Lake Whatcom. These silt  and clay-sized particles may sorb
             phosphorus descending through the water column. Nutrient concentrations and sediment metal con-
             centrations increase toward the urbanized portions of the watershed. The city water intake is located
             in the shallowest, most nutrient-rich basin, containing the greatest  development densities. Water level
             regulation is also an issue due to conflicting uses of shoreline development, water storage and use
             and downstream encroachment. Due to the flashiness of watershed streams, water and sediment
             transport can cause problems for lake level manipulation  and water quality. A Phase 1 lake restora-
             tion study funded by the State of Washington is underway, and  results will be discussed.
INTRODUCTION

Lake Whatcom is a large, coastal lake in the Puget
Sound Lowlands of Washington State. It  is relatively
deep and has three distinct sub-basins separated by
sills.  Coal  mining (until 1920's)  and logging (until
1950's)  have been major  uses  of  the  watershed,
although as  early as 1892 it  became the principal
water supply for the city of Bellingham. The growth of
the city (population 50,000) and increases in industrial
withdrawal of water caused water levels to fluctuate
markedly  during  the  year.  Lake   levels declined
substantially during  midsummer  when precipitation
was lowest.  A  diversion system  to  transport water
from  the Nooksack  River  was completed in 1962,
bringing large amounts  of water,  but also sediment,
into the Whatcom drainage basin  during the drier
months of the year.
  Expanding development from the city has increased
housing  density  principally  along  the  shallower
basins near the outlet. Water withdrawal for domestic
consumption and industrial usage is from one of the
shallow basins; consequently, lake water quality pro-
tection has been a major issue vis a vis watershed
development.
  Systematic sampling of the lake began in 1962 and
continued until 1973 (Sheaffer, 1978). Recent sampling
began in 1979 and is ongoing,  including sampling of
major  stream inlets.  In  addition, a series  of core
samples has been  taken  to  determine rates  of
sedimentation and sediment inflow  from the diver-
sion.
LAKE DESCRIPTION

Lake Whatcom is a 2,036 ha lake that dates from the
end of the Fraser glaciation as sea level retreated ex-
posing  the  basin 95  m  asl.  Major morphometric
features of the basin are given in Table 1. Most of the
narrow basin is eroded from Chuckanut Sandstone in
a southeast to northeast direction (Fig. 1). The lake
contains three major  sub-basins: the  deepest and
largest is at the southeast separated by two sills from
the other much shallower sub-basins. The  shallow
sub-basins are surrounded by development extending
from the city of Bellingham.
  Several major streams enter the lake. For about 6
months each year the Nooksack River diversion con-
tributes a large volume of water.and during the rainy
season  many intermittent streams flow because of
steep slopes  and high runoff ratios.


        Table 1.—Morphometry of Lake Whatcom.

   Parameter
Length, km
Surface area, ha
Volume, 108m3
Maximum depth, depression 1, m
Maximum depth, depression 2, m
Maximum depth, depression 3, m
Mean depth, m
Elevation, masl
  16.5
2036
   8.95
  30
  20
 100
  45
  95
                                                 239

-------
LAKE AND RESERVOIR MANAGEMENT
 RESULTS AND DISCUSSION

 Limnological Characteristics

 In all years of record, surface waters of the lake show
 pH maxima and Secchi disk minima during  summer.
 Representative summer and winter data are given in
 Table 2. Maximum lakewater transparency occurs dur-
 ing winter. The deepest sub-basin normally has lower
 nutrient concentrations and an oxygenated  hypolim-
 nion. Nutrient concentrations do not vary substantial-
 ly down-lake, but the shallower sub-basins approach
 anoxia during late summer whereas oxygen consump-
 tion in the deeper basin is slight.
   Over the period 1962-83 midsummer and maximum
 Secchi disk transparency has increased. This may be
 caused by a  decrease  in nutrient concentrations
 resulting from sewering  and the faster flushing rate
 caused by the Nooksack  River  diversion.
   Water withdrawal for the city of Bellingham is from
 sub-basin 2 at a depth of about 13  m. Approximately
 243,000 m3 of water is withdrawn daily. The Nooksack
 River diversion inputs 231,000 m3/day from April to Oc-
 tober into sub-basin 3. Neither  influence has affected
 thermal  structure,  but  has  accelerated  flushing
 (Loranger and Brakke, 1983).
   Watershed streams are flashy, and also somewhat
 varied in  basic chemistry. All major  stream inputs are
 to the larger, deeper basin, with essentially no surface
 drainage  into  the shallower   basins except during
 storm events. Storm events are  characterized by quick
 response and considerable sediment transport. Addi-
 tionally,  the  Nooksack  River  diversion contributes
 sediment-laden glacial runoff to basin 3.

 MIRROR LAKE

 The Nooksack River diversion empties  into Mirror
 Lake prior to passage down a stream channel to Lake
 Whatcom. The watershed is only 10.62 km2, and  water
 transport  has been  greatly accelerated by the  diver-
 sion. Turbid glacial  meltwater  reaches the lake, con-
 tributing  a substantial sediment load. The lake was
                                Wcitai Hwd oiid Sompfing
                                  Sit. Mop
 surveyed in 1946 and again in 1981 following the diver-
 sion season (Table 3). Assuming sedimentation rates
 were much  less prior to diversion ( <  1  cm from
 1946-61), in the last 20 years lake volume has declined
 by 20 percent
   A delta of coarse, but uniform-sized particles  has
 formed at the diversion inlet. Down-lake the sediment
 graded quickly  into silt-clay particulates that  con-
 tribute to relatively high (8-12  cm/yr) sedimentation
 rates throughout the remainder of the lake. Although
 this sedimentation was rapid, suspended sediments
 were lost in the downstream  contributory to Lake
 Whatcom.
   Both coarse and fine fractions were essentialy 100
 percent inorganic. The coarse fraction was comprised
 of felsic and  mafic minerals (quartz, plagioclases,
 feldspar, amphiboles, pyroxenes).  The fine fraction
 was composed of clay minerals. Sampling  of suspend-
 ed particulates  in the diversion water downstream
 from Mirror Lake and Mirror Lake sediment indicated
 that all were derived from the same source, with minor
 amounts of smectite, illite, and montmorillonite added
 from the Mirror Lake watershed.
                               DtPRt SSION 3o
                                      DtPRlSSlON ib
Figure 2.—Lake Whatcom morphometry.
                                                       Table 2.—Representative Lake Whatcom summer and
                                                            winter surface data, deep basin 1980-82.
Parameter
pH
Alkalinity (^eg/l)
Specific conductivity (^iS/cm)
SP fog/l)
N03-N (ng/l)
Secchi disk (m)
Winter
6.6-7.2
322
61.0
12-20
330-430
7.0-8.5
Summer
7.1-7.7
327
52.8
25-40
80-228
4.5-5.5
                                                          Table 3.—Morphometrlc data for Mirror Lake.
                                                                             1946*
                                          1981
                                                    Area
                                                    Maximum depth
                                                    Volume
Figure 1.—Lake Whatcom watershed and sampling stations.
                      5.46 ha           5.16 ha

                      10 m             9.2 m

                      324,441 m3        260,573

•Surveyed Feb. 26,1946 Washington State Dep. Game
                                                240

-------
                                                                                 SEDIMENT ANALYSIS
SEDIMENTATION RATES IN LAKE
WHATCOM

Sediment transport from the diversion could have in-
creased sedimentation rates in Lake Whatcom. In ad-
dition, logging, coal mining, and exposure of land sur-
faces by development could have resulted in excess
sediment transport to  the lake.  Sediment samples
were taken at approximately 30 sites on the lake. One
core each from basins 3 and 1 were dated by Cs-137
and Pb-210 methods. The average sedimentation rate
based on Pb-210 for the past century was 0.50 (basin 1)
to 0.53  (basin 3) cm/yr.  Over the past 30 years,
sedimentation rates for both basins were ~ 1.0 cm/yr.
Both rates were averages comprised of intervals with
low and higher sedimentation. Large inputs of sedi-
ment have occurred in the past as indicated by layer-
ing of the sediment and thick clay lenses.
  Sedimentation  rates appear to have accelerated
since 1962 (Fig. 3). Two  different factors may have
operated: diversion inputs in the south basin and in-
creased land surface disturbance down-lake.

EFFECTS OF MAJOR WINTER STORM

On January 9-10, 1983, heavy, continuous rains caus-
ed major slumping of unstable soils and the failure of
large debris dams on major streams. In addition, an
enormous volume of fine  particulate matter entered
the lake, as well as large trees and even a few houses.
More than 25 ha of woody debris was floating on the
lake following the storm. A salvage operation removed
most of the larger debris, but turbidity resulting from
fine particulates jumped substantially. Turbidity at all
stations and depth was less than or equal to about 1
NTU in  December, but rose to 12.8 (0 m) - 212 (80 m) in
basin 3 and  3.3 (0 m) - 5.2 (15 m)  in basin 1. Even
though most of the material sedimented  rapidly, tur-
bidity remained  higher than in past  years  through
August, when it was approximately  1 NTU at the sur-
face and 5 NTU at 80 m in basin 3. Because the major
streams enter basin 3 and the  slumping occurred
around that  basin the particulate  load was higher
there than in the other basins. Most of the material
sedimented in basin 3, because the sills separating
the  basins  blocked  transport  down-lake  while
sedimentation occurred.
  Phytoplankton populations declined greatly in 1983
(chlorophyll a 3-5 ug/l). Phytoplankton growth rates
may have been slowed by inorganic turbidity, but also
because phosphorus concentrations  have been ex-
tremely low, much less than at comparable periods in
1982 (Table 4). The fine  clay fraction could have strip-
ped  phosphorus  from  the water  column  during
sedimentation. Even though phytoplankton density
was low, oxygen depletion in  the shallower basins
resembled other years.
  The work begun in December 1982 as a Phase I
monitoring  study  designed to develop  water and
nutrient  budgets  for the lake has been confounded
greatly by a single storm event. Fortunately, we have
available a valuable long-term record of lake  water
quality, including measurements during the 3 previous
years. Even though a great deal of work remains to
analyze and assure that data base, without such infor-
mation it would be difficult to assess the effects of the
storm or to proceed to calculate loading rates.
  This has implications for monitoring any watershed
with several flashy stream inputs and varying  land
use. It is unlikely that any single year of record would
  Table 4.—Lake Whatcom water chemistry, Aug. 8,1983.
 Parameter
Station A   Station E    16/08/82
PH
Dissolved O2 (mg/l)
Turbidity (NTU)
NH3-N (pg/l)
N03-N ^g/l)
IP (Mg/l)
SRP (Mg/l)
7.75
6.51 (20)
10.14
0.19 (20)
0.60
1.9(20)
228.9
649.1 (20)
234.4
356.1 (20)
<2
<2(20)
<2
<2 (20)
7.46
6.76 (90)
9.81
9.42 (90)
5.8
5.7 (90)
26.3
72.5 (90)
301.6
503.8 (90)
<2
< 2 (90)
<2
<2(90)
—
—
0.4
—
226
—
40
                                     LAKE    WHATCOM
                                          Cs-137  DATES
               SOUTH  BASIN
                 NORTH  BASIN
        15-
        10-
  (cm)
        5-
          15-
          10-
           5-
             —i	1	1	1	1	1—«
             1952 1957  1962  1967  1972  1977
                                                                  1952 1957 1962 1967 1972 1977
Figure 3.—Cs-137 dates for Lake Whatcom cores.
                                              241

-------
LAKE AND RESERVOIR MANAGEMENT

be  adequate  for  developing  sound  loading  rate    REFERENCES
estimates. More baseline monitoring is required for
Phase I projects where a single source is not the prh-    Loranger, T., and D.F. Brakke. 1983. Temperature character-
cipal contributor of nutrients. At least 3 consecutive      lstlcs' annual heat St°ra9e a"d river diversion influence on
years of record would be justified. In the case of Lake      a monomictlc lake- "°™"- Sci. (m press).
Whatcom, at least  1 more year is necessary to deter-    Sheaffer, L. 1978. A Retrospective Examination  of Data
mine if nutrient concentrations and phytoplanktcn      from the Study of Lake Whatcom, April 1962 through April
populations return  to levels previously observed, ard      ,1.f 3h.Te?h'  ?,eP' Na 26' lnst- Watershed Stud.  Western
to quantify stream  inputs.                                Washington Umv.

ACKNOWLEDGEMENTS: The current work is supported by a
State of Washington Lake Restoration grant to Whatcom
County, with  matching funding from Whatcom County and
the city of Bellingham (subcontract fro URS Corp.). Sediment
sampling and Mirror Lake analyses were done under a con-
tract with the city of Bellingham. Several persons have been
involved in sampling  and analysis; I especially  thank Linda
Sheaffer, Susan Blake, and Chris Spens.
                                                  242

-------
SEDIMENT  DISTRIBUTION AND QUALITY  IN  A
SMALL WISCONSIN  RESERVOIR
ROBERT C. GUNKEL,  JR.
ROBERT F. GAUGUSH
ROBERT H. KENNEDY
U.S. Army Corps of Engineers
Waterways Experiment  Station
Vicksburg, Mississippi
            ABSTRACT

            Eau Galle Lake is a small Corps of Engineers impoundment on the Eau Galle River in west central
            Wisconsin. A sediment survey was conducted to document reservoir sediment characteristics in rela-
            tion to reservoir morphometry, hydrodynamics, and water quality. Sediment conditions at Eau Galle
            Lake tend to be lakelike, rather than exhibiting a more typically reservoir-like dependence on
            hydrodynamics. Sediment distribution patterns in the reservoir are primarily influenced by basin mor-
            phometry, which has resulted in sediment deposition and quality being depth related. A deep central
            basin, circular shape, and multiple inflows have contributed to the existence of two distinct sedimen-
            tary zones. The transport zone, which is characterized as a high-energy environment, exists in the
            shallow, littoral areas of the reservoir. Turbulent processes dominate the transport zone, thereby,
            discouraging the permanent deposition of fine particulates. As a result, transport zone sediments have
            a relatively large median particle size and low moisture content. The deep central basin of the reser-
            voir is characterized as an accumulation zone of low energy. Sediments in this less turbulent area
            are characterized by a smaller median particle size and high moisture content. Sediment chemical
            characteristics of nutrients, metals, and organic matter are higher in the accumulation zone. Therefore,
            the deep sediments may be more influential to water quality through exchanges at the sediment/water
            interface.
INTRODUCTION

Sediment transport and deposition is a major problem
for reservoir  management. Accumulating sediments
have  created a  significant  problem  by  reducing
valuable storage  volume. In addition to  storage
losses,  sediments may directly  or indirectly affect
reservoir water quality (Thornton  et al. 1981). In 1960,
sediments were assessed to be the major water pollu-
tant,  as well as  the major  carrier  for  pesticides,
nutrients, and pathogenic organisms (U. S. Senate
Select Comm. Nat. Water Resour. 1960).
   The transport and deposition of sediment to a reser-
voir is regulated  by a  number of  factors, including
basin  morphology,  hydrology,   and  the   influent
material settling characteristics. Sedimentary condi-
tions  in typical mainstream reservoirs are most likely
dominated  by advective transport. These reservoirs
are often long and relatively narrow, with an upper
riverine  zone of high-flow velocities  and turbulence.
Velocities and turbulence decrease as the  reservoir
widens  and  deepens, resulting  in longitudinal  gra-
dients of sediment accumulation  and  particle size.
Gunkel  et  al. (1983) observed that  expected longi-
tudinal gradients are confounded by preimpoundment
conditions and secondary  tributaries.
   Reservoirs less  influenced by flow may  be more
lake-like in  morphology and exhibit sediment deposi-
tion by focusing. Sediment focusing is defined as the
accumulation of fine particulate matter in the deepest
basins of a lake. Davis (1973) and Davis and Brubaker
(1973) found  that  initially smaller pollen grains were
preferentially deposited on littoral sediments, but dur-
ing fall circulation resuspension and deposition led to
a  higher  net  accumulation rate  in  the  deepest
sediments. Net movement of small particles from the
littoral to the deeper portions of Lawrence Lake have
also been  reported by Wetzel et  al. (1972).
  Hakanson  (1977) hypothesized  that:  (1)  Fine  par-
ticulate matter will not be deposited in  "high energy
environments"  (i.e., littoral  and turbulent areas); (2)
deposition of all particulate matter will be primarily in-
fluenced by hydrological flow patterns  and bottom
topology; and (3) the rate of deposition  will  increase
with increasing depth. His studies also demonstrated
a correspondence between sediment distribution and
sediment  moisture content. Sediments  in the trans-
port and erosion zones of Lake Vanern, Sweden, were
found to have moisture contents of 40 to 50 percent,
while those in the  accumulation zones had moisture
contents of 60 to 75 percent. In addition, Hakanson
(1977) observed that concentrations  of nutrients and
metals associated  with particulate matter varied pro-
portionally with moisture content.
  Sediment quality at  a small  northern U.  S. Army
Corps of Engineers reservoir was surveyed to gain in-
formation  concerning potential relationships between
sediment characteristics and reservoir morphometry,
hydrology, and water quality. This paper will  describe
observed depositional patterns and discuss patterns
in sediment quality.

STUDY SITE

Eau Galle Lake is a small Corps flood control  reservoir
created in  1968 by impounding Eau Galle River in west
                                                 243

-------
 LAKE AND RESERVOIR MANAGEMENT
 central Wisconsin approximately 80 km east of Minne-
 apolis-St. Paul, Minnesota. The primary land use in
 the 166 km2 watershed is agriculture; however, several
 small residential communities are located in the area.
 Total inflow for the lake is primarily accounted for by
 the Eau  Galle River  (85  percent).  Two secondary
 tributaries, Lousy and Lohn Creeks, account for  thei
 other 15 percent of flow.
   The lake is small (length of 1 km), nearly circular
 (shoreline development ratio  of 1.5),  and at  norma
 pool elevation (286.5 m msl) has a mean and maximurr
 depth of 3.2 and 9 m,  respectively. Surface area anc
 volume are 0.6 km2 and 1.9  x 106 prO, respectively
 Lake morphometry reflects  preimpoundment  con
 struction activities that included excavation of a large
 centrally  located area. Following impoundment, this
 area produced a deep central basin.
   Eau Galle Lake  is a dimictic reservoir experiencing
 hypolimnetic  anoxia,  high  nutrient concentrations
 periodically intense algal blooms,  and the develop-
 ment of macrophytes in littoral areas. Turnover occurs
 in September  to early October and ice cover persists
 from  December until late March. Water quality pat
 terns during normal flow appear to be associated with
 differences between littoral and pelagic regions. How
 ever, during high flow periods, water quality character
 istics may differ between flow-dominated regions anc
 adjacent nearshore areas.


 METHODS AND MATERIALS

 Sediment core samples were  collected at Eau Galle
 during  February  1-2,  1980.  Sample  stations  were
 located so as  to coincide with those established dun
 ing previous water quality studies or to incorporate
 site specific characteristics. Thirty-five stations were1
 selected for sampling; however, gravel prevented sam
 pie collection  at 10 (Fig. 1). Eighteen of the remaining
 25 stations provided sufficient material for both parti-
 cle size and chemical analysis, while six stations  are
 represented by only particle size data. A single station
 is represented by only chemical data.
   Sediment samples were collected using a single-
 barrel Wildco Core  Sampler (Wildco Supply Co.
 Saginaw, Mich.) fitted with polyethylene liners. The
 core  sampler  provided a means for identifying  sur-
 ficial sediments and maintaining sample integrity. The
 sampling involved collecting  two  core samples  for
 each  station:  one  for particle size analysis, and  the
   LEGEND

 O NO CORE SAMPLE
 O PARTICLE SIZE
 
-------
                                                                                      SEDIMENT ANALYSIS
Figure 2.—Sediment median particle size represented by bar
height for Eau Galle Lake stations. The 3.5 m contour line is
shown within the lake.
   Based on this observation and  the absence of any
significant correlation between median particle size
and distance from the Eau Galle  River the data were
subset for further analysis by: (1) shallow sediments
(^3.5 m), and (2) deep sediments (>3.5 m). The char-
acteristically turbulent  nature  of the  shallow  (high
energy) areas is reflected by the relatively uniform dis-
tribution of particle volume among each of the 13 par-
ticle size classes (Fig. 3). The particle size distribution
for deep sediments is skewed toward the  smaller size
classes, suggesting  preferential deposition of small
particles in the deep areas of the reservoir. This sor-
ting of  particles could be the result  of differential
transport of allochthonous inputs or material resus-
pension  by mixing.  Median  particle  size for  deep
sediments is significantly (p< .001) smaller than that
of the shallow sediments (10.41 and 21.01 j^m, respec-
tively).
       SHALLOW
        DEEP
                    PARTICLE SIZE, im\
Figure 3.—Relationship of percent total volume and particle
size for shallow and deep sediments for Eau Galle Lake.
  Sediment moisture  content  also appears  to  be
depth  related.  Hakanson  (1977)  observed  that
moisture content is inversely related to particle size.
Consistent with Hakanson's (1977) findings, Eau Galle
deep sediments  have a mean moisture content of 67
percent, which is significantly different (p< .001) from
the mean moisture content (45 percent)  for shallow
sediments. This  suggests that the deep areas of Eau
Galle are zones  of accumulation, while the shallow
sediments are subjected  to erosion  and transport.
  In addition, Hakanson (1977) reports that nutrient
and  metal concentrations  are  associated with and
vary proportionally to moisture content.  In the deep
basins of Lake Vanern he found enrichments 150 to
550 percent for  organic  matter, nitrogen, and phos-
phorus. All Eau Galle sediment concentrations exhibit
significant differences between shallow and  deep
sediments (Table 1). Concentrations of total organic
carbon, nitrogen, phosphorus, iron, and  manganese
are approximately 1.5 to 2.0 times higher in the deeper
sediments than  in shallower sediments.  These high
concentrations are possibly related  to the accumula-
tion of fine particulate  matter in the deep basin.
  A major portion  of fine particulate matter consists
of phytoplankton,  which act  as  concentrators  of
epilimnetic  carbon,  nitrogen,  and  phosphorus.  A
significant fraction of their cellular  carbon, nitrogen,
and phosphorus will be deposited with sedimenting
fine  particulate  matter, thereby enriching the  deep
sediments.  Wetzel  (1975) found   that  iron and
manganese bound  in the biomass of phytoplankton is
not lost  in the initial stages of decomposition, but
rather, moves  with sedimenting  organic detritus.
Other mechanisms for removing dissolved nutrients
and metals from the water column  would be by the
settling out of clays and  other fine inorganic par-
ticulates.
  In contrast to the other sediment variables, only in-
organic carbon has a higher mean  concentration  in
shallow  sediments;  this is probably  the result  of
precipitation  and  deposition of  CaCO3. In  littoral
areas, CaCO3 precipitation is induced when photosyn-
thesis by macrophytes and phytoplankton uses C02.
Although this precipitation  can  occur in open waters
as well as the littoral areas, little, if any, CaCO3 will be
deposited in deep sediments because of hypolimnetic
C02- Carbon dioxide content increases with depth and
in the hypolimnion  where CO2 is  often abundant,
CaCOs  is dissolved, and  bicarbonate  content  in-
creases.  This mechanism effectively prevents enrich-
ment of deeper sediments with inorganic carbon.
  As a result of accumulation in deep sediments, high
interstitial water concentrations of total and soluble
reactive  phosphorus,  iron, and  manganese  were
observed. Since  these same variables exhibit  enrich-
ment in the sediment phase it can  be expected that
dissolved/solid phase interchanges would  increase in-
terstitial   concentrations.  In addition, higher  total
organic carbon  in  the sediment fraction  should  in-
crease the rate and intensity of reduction in anaerobic
systems  (Gunnison and Brannon, 1981) and result  in
higher concentrations of soluble products. The lack of
significant enrichment  by  any  of the  interstitial
nitrogen forms may be the result of their relative ease
of mobilization from the sediments (Wetzel, 1975).
  Correlations between particle size classes for deep
sediments and various chemical and physical concen-
trations are presented  in Table 2. Percent volume  in
size  classes  2.4 through 4.7 ^m  are positively cor-
related with both sediment and interstitial nitrogen (all
forms), iron and manganese, sediment  total  phos-
                                                 245

-------
LAKE AND RESERVOIR MANAGEMENT
phorus,  and interstitital  carbon. These correlations
suggest that particles in the range 2.4 through 4.7 ^m
dominate chemical and biological activities in  deep
sediments. This results from the fact that particles  in
this size range contribute 61 percent of the deep sedi-
ments'  total surface area, while constituting only 29
percent of the total volume. Sly (1977)  stressed that
the surface area of clay-sized particles is on the order
of square meters per gram whereas the surface area
of sand grains is only on the order of  square centi-
meters  per  gram.  Fine  grain  materials have the
greatest potential  for chemical  and biological inter-
action because of the importance of surface reactions
in sediments (Jones and Bowser, 1977).
   Sediment organic carbon is correlated with percent
volumes in the 13 and  19 ^m  size  classes.  Since
organic  carbon  is  usually  correlated  with  clays
(Thomas, 1969), the  correlation  observed here with
larger-than-clay-sized particles   suggests  the  ex-
istence  of detrital  particulate organic matter rather
than an organic film on an inorganic particle.
    There is a general lack of significant correlation be-
 tween chemical  composition and percent volume in
 the 6.6, 9.4, 106,  and 150 ^m size classes. The coeffi-
 cients of variation are lowest for the 6.6 and 9.4 ^m
 size classes (CV  = 12 and 8, respectively). In addition
 to having the lowest CV's these two size classes ac-
 count for 25 percent  of the deep sediments  total
 volume,  which suggests that these two size classes
 are  relatively  evenly  distributed across  the  deep
 sediments and would therefore exhibit little relation to
 chemical composition. The high variability observed in
 the 106 and 150 ^m size classes (CV  =  149 and 316,
 respectively) and their small contribution (0.7 percent)
 to total  volume  of deep sediments is probably the
 reason for the lack of correlation.
    Shallow sediments exhibit a general lack of correla-
 tion between particle size and chemical composition.
 This lack of correlation may imply that the variability
 of littoral sediment chemical composition may be a
 function  of  localized  influences  (inflows,  macro-
 phytes, direct runoff, etc.) rather than  particle size.
                       Table 1.—Eau Galle Lake mean values for shallow and deep sediments.
               Variable
                                                    Shallow*
 Sediment chemical composition, mg/g
     Total inorganic carbon
     Total organic carbon
     Total nitrogen
     Total phosphorus
     Total iron
     Total manganese
 7.54
15.34
 2.03
 0.72
18.76
 0.76
                                                                              Deep*
 3.33
30.27
 3.14
 1.35
31.52
 1.09
                                               P"
Interstitial chemical composition, mg/l
Soluble reactive phosphorus
Total phosphorus
Total iron
Total manganese
Nitrate nitrite nitrogen
Ammonium nitrogen
Total nitrogen
Total inorganic carbon
Total organic carbon

0.19
0.19
8.42
4.97
0.01
10.92
11.57
84.40
12.29

0.30
0.33
20.18
10.13
0.02
13.43
15.02
80.56
16.63

<0.05
<0.05
<0.05
<0.05
NSt
NS
NS
NS
NS
< 0.001
< 0.005
< 0.001
< 0.001
< 0.001
<0.01
  * Number of observations on which calculations are based: for shallow, n = 10, for deep, n = 9.
 " Probability that means are euqal.
  t Nonsignificant difference (p >0.05).
       Table 2.—Significant (p^O.05) correlation coefficients for the Eau Galle deep sediments (particle size values
                                         are midpoints of size ranges).
                                                                Particle Size
              Variable
                                     150   106   75    53   38    27   19   13    9.4   6.6   4.7   3.3   2.4
 Interstitial chemical composition
     Total inorganic carbon
     Total organic carbon
     Nitrate nitrite nitrogen
     Ammonium nitrogen
     Total nitrogen
     Soluble reactive phosphorus
     Total phosphorus
     Total iron
     Total manganese
-0.44
NS
NS
NS
NS
NS
NS
NS
NS
NS*
NS
-0.44
NS
NS
NS
NS
NS
NS
-0.81
-0.82
-CI.45
-CI.79
-0.78
NS
NS
-0.78
-0.76
-0.47
-0.69
-0.62
-0.45
-0.48
NS
NS
-0.52
NS
NS
NS
-0.45
NS
NS
NS
NS
NS
NS
-0.46
-0.68
NS
-0.54
-0.56
NS
NS
-0.67
-0.50
-0.44
-0.68
•0.70
-0.55
-0.56
NS
NS
-0.54
-0.54
-0.58
-0.53
NS
-0.59
-0.57
NS
NS
-0.49
-0.64
NS
NS
0.48
NS
NS
NS
NS
0.46
NS
NS
NS
0.66
NS
NS
NS
NS
NS
NS
NS
0.69
0.64
0.48
0.51
NS
NS
0.58
NS
0.58
0.81
0.59
0.63
0.66
NS
NS
0.71
0.59
0.61
0.80
0.57
0.61
0.63
NS
NS
0.66
0.56
Sediment chemical composition
Total inorganic carbon
Total organic carbon
Total nitrogen
Total phosphorus
Total iron
Total manganese
Median particle size
Column depth

NS
NS
NS
NS
NS
NS
NS
NS

NS
NS
NS
NS
NS
NS
0.80
NS

NS
NS
-0.69
-0.85
-0.45
-0.63
0.52
NS

NS
NS
NS
-0.60
NS
NS
0.87
NS

NS
NS
NS
NS
NS
NS
0.84
NS

NS
NS
-0.68
-0.71
-0.47
NS
0.85
-0.45

NS
0.55
-0.53
-0.66
NS
NS
0.73
-0.63

0.56
0.67
NS
-0.67
NS
-0.57
NS
NS

NS
NS
NS
NS
NS
NS
-0.80
0.44

NS
NS
NS
NS
NS
NS
-0.88
0.48

NS
NS
0.53
0.61
NS
NS
-0.97
0.57

NS
NS
0.62
0.79
0.46
0.44
-0.91
0.54

NS
NS
0.56
0.77
0.45
0.45
-0.87
0.54
 * NS = correlation was not significant.
                                                   246

-------
                                                                                          SEDIMENT ANALYSIS
Turbulence in the littoral  zone has  reduced particle
size variability between stations, thereby reducing cor-
relations between particle size and chemical composi-
tion.

CONCLUSIONS

Sedimentary conditions in Eau Galle are primarily a
function of basin morphology (i.e., depth). A combina-
tion of  Eau Galle's  circular shape, multiple inflows,
and deep central basin result in sediment deposition
by focusing. In this  regard, two sedimentary environ-
ments can be distinguished within the reservoir. A
high energy environment,  which is dominated by tur-
bulent processes (e.g., flow, pool fluctuation,  wind),
comprises the littoral and inflow areas of the reser-
voir. The  turbulent  nature of this  environment dis-
courages  the  permanent deposition  of  fine  par-
ticulates. Sediments in these areas are characterized
as having a larger median particle size, low moisture
content, and lower nutrient, metal, and organic matter
concentrations.
  Conditions in the  low energy environment (i.e., deep
portions of the lake) are less turbulent and provide an
area for sediment  accumulation. Sediments in this
area are characteristically higher in moisture content
and are  comprised of relatively smaller  particles.
Higher  concentrations of  nutrients,  metals,  and
organic  matter  are  also  found  in  these  deeper
sediments. These characteristics along with expected
exchanges between the sediment and overlying water
infer that  deep sediments are likely to influence reser-
voir water quality.

ACKNOWLEDGEMENTS: This research funded by the Envi-
ronmental and Water Quality Operational studies sponsored
by the Office of the Chief, U.S. Army Engineers.

REFERENCES
Davis, M.B. 1973. Redeposition of pollen grains in lake sedi-
  ments. Limnol. Oceanogr. 18:44-52.
Davis, M.B., and L.B. Brubaker. 1973. Differential sedimenta-
  tion of pollen grains in lakes. Limnol. Oceanogr. 18:635-46.
Gunkel, R.C., Jr., et al. 1983. A comparative study of sediment
  and quality in four reservoirs. Tech. Rep. (in press). U. S. Ar-
  my Eng. Waterways Exp. Sta., Vicksburg, Miss.
Gunnison, D.,  and J.M. Brannon. 1981. Characterization  of
  anaerobic chemical processes  in  reservoirs: problem
  description and conceptual model formulation. Tech. Rep.
  E-81-6,  U. S. Army Eng. Waterways Exp. Sta., Vicksburg,
  Miss.
Hakanson, L 1977. The influence of wind, fetch, and water
  depth on the  distribution of sediment in Lake  Vanern,
  Sweden. Can. J. Earth Sci. 14.
Jenkins, T.F., et al. 1981. Chemical analysis of sediment and
  interstitial waters from selected  Corps reservoirs. ODMR
  No. WESRF 80-164. U.S. Army Eng. Cold Regions Res. Eng.
  Lab., Hanover, N.H.
Jones,  B.F., and C.J. Bowser. 1977. The  mineralogy and
  related chemistry of lake sediments. In A. Lerman, ed.
  Lake Chemistry, Geology, Physics. Springer-Verlag, New
  York.
Sly, P.G. 1977. Sedimentary processes in lakes. In A. Lerman,
  ed.  Lake Chemistry, Geology, Physics.  Springer-Verlag,
  New York.
Thomas,  R.L 1969. A note on the relationship of grain size
  clay content, quartz, and organic carbon in some Lake Erie
  and Lake Ontario sediments. J. Sediment. Petrol. 39:803-9.
Thornton,  K.W.,  et  al. 1981. Reservoir sedimentation and
  water quality-an heuristic model. In H.G.  Stefan, ed. Proc.
  Symp. Surface Water Impoundments. Am. Soc. Civil Eng.,
  New York.
U. S.  Senate Select  Committee on  National Water  Re-
  sources. 1960. Pollution Abatement, Committee Print No.
  9, 86th  Congress, 2nd session.
Wetzel, R.G., P.M. Rich, M.C. Miller,  and H.L Allen. 1972.
  Metabolism of dissolved and particulate detrital carbon in
  a temperate hard-water lake. Mem. Dell. Inst. Ital. Idrobiol.
  (Suppl.):185-243.
Wetzel, R.G. 1975. Limnology. W.B. Saunders Co., Philadel-
  phia.
                                                    247

-------
 ANALYSIS OF  SURFICIAL  SEDIMENT  FROM
 63 ILLINOIS LAKES
 M. KELLY
 R. HITE
 K. ROGERS
 Illinois Environmental  Protection Agency
 Marion, Illinois
             ABSTRACT

             Surficial sediment samples were collected from 63 Illinois lakes during summer 1979. Samples were analyzed
             for organic matter, nutrients, heavy metals and organochlorine compounds For purposes of statistical
             analysis, lakes were placed into one of several groups (glacial, artificial or miscellaneous). Most lakes
             were artificial reservoirs and these were subdivided based on geographic location. Spatial variations
             of glacial lake means were contrasted with  artificial lake studies. In an attempt to correlate fish flesh
             pesticide concentrations with sediment concentrations, no simple linear relationships were discern-
             ed. In general, the low organic content of Illinois sediments was probably attributable to the high non-
             volatile suspended solids loading characteristic of most Illinois lakes. Regardless of organic carbon
             content, the ratio of C:N remained fairly  constant at 14:1 The N:P ratio for Illinois  lake sediment of
             6:1 was somewhat less than the generally conceded 7:1 for plant material. Most Illinois lakes exhibited
             fairly low sediment metal/metalloid concentrations. Most chlorinated  hydrocarbon pesticides were
             undetected or occurred at low levels in sedirient samples. PCB's were detected only in sediments
             from seven of the lakes sampled. A classification of Illinois lake sediments based on mean consti-
             tuent concentrations and respective standard deviations was developed to facilitate interpretation of
             sediment data. The resultant four tier classification system categorized lake sediments as below nor-
             mal, normal, elevated, and highly elevated.
INTRODUCTION

Many States fortunate enough to have natural lakes of
high  quality  have  long maintained  some type of
monitoring program for protection and enhancement
of their lentic  resources.  In  other States  where
relatively   new  artificial  reservoirs  predominate,
monitoring has only recently been initiated under the
mandates of the Federal Water Pollution Control Act
of 1972 (P.L. 92-500)  and its  successor, the  Clean
Water Act of 1977 (P.L 95-217).
  Ambient monitoring of Illinois lakes was begun by
the Illinois Environmental Protection Agency (IEPA) in
1977 with  data collected on 107 lakes (Sefton, 1978).
Additional lakes were monitored in 1978 in an attempt
to examine the feasibility of classifying Illinois lake's
trophically, using  the  Earth  Resources  Technolociy
Satellite, LANDSAT (Boland et al. 1979). The most
comprehensive lake sampling program to date was in-
itiated in 1979 with the collection of a wide range of
physicochemical and biological data from 63 Illinois
lakes (Sefton et al. 1980). Included among the para-
meters was chemical analysis of surficial sediments.
Primary objectives of sediment monitoring conducted
in 1979 were to establish a data base to (1) facilitate
between-lake comparisons, (2) identify potentially tox-
ic contaminants and specific areas of contamination,
(3) aid in development of future lake monitoring  strate-
gies, (4) generate data to establish permit guidelines
for lake dredging, and (5) establish an adequate data
base to facilitate an understanding of surficial sedi-
ment chemistry in  Midwestern reservoirs.
METHODS

The 63 lakes monitored in 1979 were selected to in-
clude as much variability in  physiography, morphol-
ogy, type (e.g., glacial, artificial, backwater), hydrol-
ogy, and  watershed land use characteristics as possi-
ble. Since 94 percent of Illinois lakes are artificial im-
poundments concentrated in  the southern two-thirds
of the State (Sefton et al. 1980), most lakes sampled
were in this category.
  Samples were taken with a  petite Ponar and placed
in white porcelain pans. To insure as much uniformity
as possible only the  uppermost sediment layer (3-5
cm) was  analyzed. Sediment  was  removed by  hand,
placed into appropriate containers, and kept  frozen
until analyzed.  Replicate  samples  were routinely
taken at three sites. A few lakes were represented by
replicate  samples at  a single site.  Samples to be
analyzed  for heavy metals, Kjeldahl-nitrogen, volatile
solids,  total  phosphorus, and  chemical oxygen  de-
mand (COD) were  placed in  polyethylene bottles.
Samples  designated  for  analysis  of organochlorine
compounds were placed  in specially prepared  glass
bottles (xylene rinsed with foil or teflon lined caps). All
analyses were performed according to  accepted U.S.
Environmental Protection Agency (EPA) procedures
and IEPA quality assurance procedures  (Kelly and
Hite, 1981).  Resultant sediment chemistry data were
stored in the  EPA's STORET data base system and
statistical analyses were performed using the SAS
(1979) computer  package. For purposes of statistical
analysis,  lakes were placed into one of several groups
                                                 248

-------
                                                                                         SEDIMENT ANALYSIS
(glacial, artificial, or miscellaneous). Since most lakes
were artificial (reservoirs), this group was subdivided
based on geographic location (north,  central,  south
central, and south).

RESULTS AND DISCUSSION

In summer 1979, 273 sediment samples were collected
from 63 Illinois lakes  (X of 4.3 samples/lake). Analysis
of data generated by the  1979 study  revealed iden-
tifiable areas of contamination in some lakes, distinct
trends in concentrations of constituents within Illinois
lakes and between different types of lakes, and signifi-
cant relationships between concentrations of certain
parameters in surficial sediments. This paper synop-
sizes sediment chemistry characteristics for selected
parameters in Illinois lakes and a  briefly discusses of
the more  pertinent relationships and trends.

Organic Carbon and Related Constituents

Mean volatile solids content of Illinois lakes was 8.8
(±2.9) percent (Table 1). Analysis of variance  using
Duncan's multiple range test with  lake  type as the
classification variable (Table 2) indicated that glacial
Illinois lakes contained significantly greater levels of
volatile solids than artificial lakes did. Organic carbon
values were computed from volatile solids (Gorham et
al. 1974). Using calculated  carbon values and assum-
ing total Kjeldahl nitrogen was equal to total nitrogen
(Anderson, 1974),  the C:N ratio was  found  to remain
relatively constant regardless of sediment constituent
concentrations. The mean C:N for Illinois lakes was
14:1. Frink (1969), Gorham et al. (1974) and Brunskill et
al. (1971)  noted similar results; all reported ratios  of
12:1.
  Volatile solids, total  Kjeldahl  nitrogen,  and COD
were all highly intercorrelated. As  a  result, it is possi-
ble to predict rather accurately two of  the  three
variables after determination of one (Fig. 1).
  Illinois  lake  sediment phosphorus concentrations
were noticeably low with a mean of  703(±476) mg/kg
       o
       V)
       *.  10
       I
                     2468
                     Total  Kjeldahl  Nitrogen (g/kg)
             o
             in
                  0      60     120     180    240     300

                                COD   (g/kg)
      Figure 1.—(a) Regression of volatile solids and total Kjeldahl
      nitrogen for 259 sediment samples from  63 Illinois lakes.
      Equation of regression line is: Volatile solids (%) =  0.00189
      total Kjeldahl (mg/kg) + 2.59. (b) Regression of volatile solids
      and COD for 259 sediment samples from 63 Illinois lakes.
      Equation of regression line is: Volatile solids (%) = 5.47 x
      10-5 COD (mg/kg) + 4.09.
  Table 1.—Grand mean lake sediment concentrations of selected parameters in 63 Illinois lakes sampled summer 1979.
         Minimum and maximum values reflect highest and lowest lake means. Sample size within lakes varied.
Volatile solids (%)
Total Kjeldahl nitrogen (mg/kg)
Total phosphorus (mg/kg)
COD (mg/kg)
N:P ratio
Organic carbon* (mg/kg)
C:N ratio
n
62
63
63
63
63
62
62
Mean
8.83
3358
666
83347
5.53
44154
14.3
Standard
Deviation
2.93
1630
341
49816
2.95
14654
2.3
Minimum
0.60
245
280
5250
1.16
3000
9.5
Maximum
19.86
8180
2842
233000
16.00
99292
21.2
 "Organic carbon was computed from percent volatile solids
 Table 2.—Mean sediment percent volatile solids by lake type in 63 Illinois lakes sampled in 1979. Duncan's multiple range test
                      was used to compare lake type means, and groupings were determined.
 Lake Type (n)
Std Dev
Min
Max
 •Alpha level = 005, DF = 56, MS = 6118
Groupings*
Glacial (8)
Artificial central (24)
Artificial north (6)
Artificial south (10)
Artificial south central (8)
Miscellaneous (6)
Grand X (63)
13.15
8.52
8.49
8.01
7.92
7.23
8.83
4.21
1.76
2.04
1.69
1.46
4.14
2.93
6.32
5.78
5.52
5.94
6.03
0.60
0.60
19.86
13.64
11.02
10.75
9.80
13.38
19.86
A
B
B
B
B
B

                                                   249

-------
LAKE AND RESERVOIR MANAGEMENT
 P. The mean was 50 to 65 percent lower than concen-
 trations found in other studies (Brunskill et al. 1971;
 Williams et al. 1976; Frink, 1969). Low sediment phos-
 phorus concentrations in Illinois lakes, however, do
 not imply low loadings or that substantial amounts of
 phosphorus are  tied up in standing crop biomasses.
 Water chemistry data  (Sefton  et al. 1980) suggested
 that virtually all Illinois lakes were eutrophic and many
 hypereutrophic, based on total phosphorus concentra-
 tions (Allum et al. 1977). The relatively low percentage
 of total phosphorus in sediments attests to high non-
 volatile solids loading from watersheds predominantly
 agricultural in nature.

 Arsenic  and Heavy Metals

 Sediment analyses for heavy metals is useful for iden-
 tifying potentially toxic  metals,  establishing  back-
 ground levels, and  determining  possible pollutional
 loadings. While many monitoring studies of metals in
 lake sediments  have  been performed,  most involve
 assessment of only a  few  metals relative to known
 pollutional  loadings.  Illinois  lake sediments were
 analyzed for arsenic and eight heavy metals (Table 3);
 most  lakes exhibited low sediment metal concentra-
 tions.
  Arsenic concentrations in 273 samples exceeded 20
 mg/kg in only 12 percent of samples. Highest concen-
 trations were found in  lakes where sodium arsenate
 was historically  applied for weed control.  High sedi-
 ment  concentrations  were indicative of detectable
 concentrations in overlying water; however, water con-
 centrations were all well below the State standard of
 1.0 mg/l  for general use waters  (III. Pollut.  Control
 Board, 1977).
  Cadmium concentrations in sediments  fell within
 the range of values for lakes  not subject to  known
 pollutional sources (Kemp et  al. 1976; Mathis and
 Kevern,  1975).  Concentrations  were  below  the
 minimum  detectable level (0.5 mg/kg) in  124  of 272
 samples   analyzed.  Only  six  samples   contained
 greater than 2.0 mg/kg cadmium.
  Chromium concentrations in 271 samples analyzed
exhibited  little variability; the mean was 21.6 (±8.0)
 mg/kg with a coefficient of variation of 36.9 percent.
 Individual  sample concentrations ranged from 1 to 75
 mg/kg. Only four samples exceeded  35 mg/kg;  all
came from Skokie Lagoons, a  series of low level  ar-
tificial lakes in the  Chicago area, fed by the Skokie
 River. Correlation of iron with chromium (Fig. 2)  in-
dicated that the  ratio of iron to chromium was fairly
 constant  at 1000:1.  A significant departure from  the
expected ratio was indicative of enrichment. Skokie
 Lagoons,  for example, accounted for the  four most
elevated chromium values (Fig. 2) with sediments from
this lake containing  roughly twice the anticipated
chromium given a known iron concentration.
   Copper concentrations in individual lake sediment
samples ranged from  3 to 560 mg/kg. Highest copper
concentrations  were  found  in  reservoirs  used  for
municipal water supplies;  elevated levels were  pro-
bably attributable to use of copper sulfate for algae
control. The Illinois grand mean sediment concentra-
tion, 41 mg/kg Cu, was comparable to means for Lake
George (Schoettle and Friedman, 1974) and  Lake Erie
(Kemp et al. 1976).
   The aquatic chemistries of iron and manganese are
similar; this "is reflected geologically in their common
association in rocks of all kinds" (Bortleson and Lee,
1974).  In the absence of significant artificial inputs of
either  metal,  a close correlative  between  the  two
metals in sediments  might be anticipated and  was
found  (r = 0.5684, p = 0.0001,  n = 273). A curvilinear
trend was noted and was probably attributable to the
greater mobility of Mn  with respect to  Fe under
anaerobic conditions, since  Fe can be lost through
precipitation of  FeS while Mn remains in solution.
   Several workers (Wildung et al. 1977; Howeler, 1972;
Fillos  and Swanson,  1975;  Bortleson and Lee, 1974)
 O)
 O)
 c
 O
      50
      40
      30
      20
      10
                 17      33     49      65
                   Chromium  (mg/kg)
81
Figure 2.—Regression of iron and chromium for 273 sedi-
ment samples from 63 Illinois lakes. Equation of solid regres-
sion line is: Iron (mg/kg)  = 685 Chromium (mg/kg) + 12373.
The dashed  regression  line was computed after omitting
Skokie Lagoons sediment samples (all chromium values >40
mg/kg); the equation of the regression line is: Iron (mg/kg) =
1119 Chromium mg/kg + 3644.
 Table 3.—Grand mean lake sediment concentrations of eight heavy metals and arsenic in 63 Illinois lakes sampled summer
                                     1979. All concentrations in mg/kg.

Arsenic
Cadmium
Chromium
Copper
Iron
Lead
Manganese
Mercury
Zinc
Mean
11.17
<0.98
22.5
41.3
28631
<49.6
1313
< 0.09
111.0
Standard
Deviation
11.78

6.3
48.9
7163

955

47.8
Minimum
0.7
<0.5
3.7
5.0
5700
<5
195
< 0.04
16.5
Maximum
63.0
4.0
49.5
367.5
44667
183.3
6917
0.31
403.3
Minimum Detectable
Concentration
0.1 mg/kg
0.5 mg/kg
1.0 mg/kg
1.0 mg/kg
50 mg/kg
5 mg/kg
5 mg/kg
0.01 mg/kg
1.0 mg/kg
                                                 250

-------
                                                                                      SEDIMENT ANALYSIS
have demonstrated a positive linear relationship  be-
tween sediment total phosphorus and iron within indi-
vidual lakes. When total iron was  regressed against
total phosphorus for Illinois lakes,  the relationship
proved  significant  but the  coefficient   was  low
(r = 0.2239, p= 0.0002, n = 273).  Not enough samples
were taken from an individual lake to warrant a single
within-lake analysis; however, when correlations were
attempted within lake groups, some improvement was
found. In  the majority of lakes sampled, the ratio of
total iron to total phosphorus generally exceeded 40:1;
two  lakes  showed  extreme  ratios—12:1 (Skokie
Lagoons) and 5:1 (Gladstone Lake). The relationship of
sediment  iron to phosphorus is attributable  to  the
tendency  of phosphate "to interact with ferric  iron to
form a 'mixed' ferric hydroxo-phosphate precipitate"
(Bortleson and Lee, 1974).
  Sediment lead concentrations revealed the greatest
dichotomy between lake types.  All glacial lakes  ex-
hibited mean concentrations of 69 mg/kg  or greater,
and  with  the single exception of Skokie Lagoons
(X = 159 mg/kg), no  other nonglacial lake mean  ex-
ceeded 60 mg/kg. While values greater than 60 mg/kg
would be atypical of artificial lakes, the higher  values
found in glacial lakes as well as Skokie Lagoons may
be attributable to  proximity of these lakes  to  the
Chicago metroplex.
  Eighty-five percent of Illinois lakes exhibited mean
mercury sediment concentrations in the range  of 0.04
to 0.14 mg/kg. Very few  data are available  to indicate
at what  level mercury  can be tolerated  in surficial
sediments without  resulting in significant accumula-
tions in higher trophic levels such as fish.  It is  known
that  sediments are important in that most of the bio-
logical methylation  (by  microorganisms)  in lakes is
assumed  to take  place in the surficial  sediments
(Jernelov, 1970). Studies have determined that soluble
methylmercury compounds are taken  up and bio-
magnified in aquatic food webs. The background mer-
cury concentration  of  lake  sediments is generally
regarded  to be in  the range of 10 to 100 ug/kg  dry
weight (Sarkka et  al. 1978). The majority  of  Illinois
lakes sampled exhibited means within  this  range,
albeit in  the upper end. A few  lakes (particularly
Skokie Lagoons and glacial lakes)  exceeded what
would be  considered background levels; these  excep-
tions, however, while they do denote anthropogenic
loadings,  do not appear extreme.
  Zinc concentrations  exhibited  low between-lake
variability and were generally equal to or less than
concentrations encountered  in  essentially   uncon-
taminated lacustrine sediments (Kemp et al. 1976; Pita
and  Hyne,  1975;  Schoettle and  Friedman,  1974).
Eighty-five percent of all lake means were between 60
and 160 mg/kg. Skokie Lagoons exhibited the highest
mean concentration, 403 mg/kg Zn.
  Although zinc was highly correlated with organic
content (Fig. 3), the correlation coefficient was low
(r = 0.2676,  n = 259, p = 0.0001).  A  trend of con-
comitantly increasing zinc and volatile solids, how-
ever, is readily apparent particularly when more  de-
viant values (>200 mg/kg) are deleted. Concomitant in-
creases are expected within lakes and are generally
ascribed to binding of metals to organic matter and/or
clay.
  A  high  correlation between zinc and lead was  not
surprising (Fig. 4) considering their common geolog-
ical co-occurence.  Except  for Skokie Lagoons, only
glacial lake sediments contained more than 70 mg/kg
Pb. With the omission of glacial lake data, an approx-
imate lead to zinc  ratio of 1:2 typified Illinois lakes.
The ratio  in glacial  lakes was roughly 1:1.
Chlorinated Hydrocarbon Compounds

Sediment samples were analyzed to determine con-
centrations of nine chlorinated  hydrocarbon  pesti-
cides and polychlorinated biphenyls (PCBs).
   Aldrin and endrin  were not detected in any of 273
samples analyzed; the minimum  detectable level for
both of these constituents and dieldrin was 1  ^g/kg.
Dieldrin was detected more frequently than any other
pesticide assayed (58 percent of samples). Only 19
samples, however, contained concentrations  in ex-
cess  of  20 M9/kg; all came from artificial impound-
ments with watersheds primarily in row crop cultiva-
tion.
   The insecticide chlordane and its derivatives hep-
tachlor and heptachlor epoxide were detected in 34, 2
and 25 percent of samples analyzed. Detected concen-
trations  rarely exceeded 20 M9/kg.
   PCBs  were detected in only 14 (6 percent) samples
from seven lakes. Of the seven, four glacial and one ar-
tificial lake (Skokie Lagoons) are located within or
near the Chicago Metropolitan area. All detected con-
centrations were relatively small; several  barely ex-
ceeded the minimum detectable level (10 M9/kg).
 o
 M
 O
 >
      0          200         400        600       800

                      Zinc  (mg/kg)

Figure 3.—Regression of volatile solids and zinc for 259 sedi-
ment samples from 63 Illinois lakes. Equation of solid regres-
sion line is: Volatile solids (%) = 0.0158 Zinc (mg/kg) + 7.45.
The dashed regression line was computed after omitting all
zinc values greater than 200 mg/kg (samples from Wolf Lake
and Skokie Lagoons); the equation of the line is: Volatile
solids (%) =  0.0766 Zinc (mg/kg) + 1.28.
 O)
 E
•o
 a
                     Zinc  (mg/kg)

Figure 4.—Regression  of lead  and zinc for 273 sediment
samples from 63 Illinois lakes. Equation of solid regression
line is: Lead (mg/kg) = 0.448 Zinc (mg/kg) + 6.8. The dashed
regression line was computed after omitting Skokie Lagoons
sediment samples (all zinc values >300 mg/kg); the equation
of the regression line is: Lead (mg/kg) = 0.750 Zinc (mg/kg)
- 24.55.
                                                  251

-------
 LAKE AND RESERVOIR MANAGEMENT
   Table 4.—Classification of Illinois lake sediments: Groupings for each constituent shown are based upon 273 individual
  sediment samples collected from 63 lakes in summer 1979. Ranges of concentrations displayed and resultant groupings are
                              based on one or two standard deviations from mean.
Constituent
Volatile solids (%)
Total Kjeldahl nitrogen (mg/kg)
Total phosphorus (mg/kg)
COD (mg/kg)
N:P ratio
Organic carbon' (mg/kg)
C:N ratio'
Arsenic (mg/kg)
Cadmium (mg/kg)
Chromium (mg/kg)
Copper (mg/kg)
Iron (mg/kg)
Lead (mg/kg)
Manganese (mg/kg)
Mercury (mg/kg)
Zinc (mg/kg)
Below Normal
<5
<1650
< 225
< 32500
<2.2
-C 26500
<11


<14

< 18000
<15


<50
Normal
5-13
1650-5775
225-1175
32500-162000
2.2-9.7
26500-65550
11-17
<27
<1.8
14-30
<100
18000-36000
15-100
< 3000
<0.25
50-175
Elevated
13-17
5775-7850
1175-1650
162000-226500
9.7-13.5
65550-85100
17-20
27-41
1.8-2.6
30-38
100-150
36000-45000
100-150
3000-3900
0.25-0.40
175-250
Highly Elevated
<17
<7850
C1650
< 226500
<13.5
< 85100
<:20
<41
<2.6
<38
<150
<< 45000
<150
^3900
-C0.40
-^250
'Organic carbon values were calculated from percent volatile solids
Effects of Morphological Variables on
Sediment Chemistry

Most previous  sediment studies on individual  lakss
(Frink, 1969; Thomas and Jaquet, 1976; Pita and Hyne,
1975) have demonstrated a trend in increasing  con-
centrations of  numerous constituents (for examp e,
organic carbon, total phosphorus, chromium, coppor,
iron, manganese) in a downlake direction in reservoirs
(toward the dam) or toward the center of glacial lakes.
Whether glacial or artificial, these  increases are e.p-
parently depth dependent. It is generally believed that
increases  are  attributable to the  binding of these
substances to clay or organic particles in suspension.
The extent to which clay particles and/or organic par-
ticles and  associated constituents settle out should
be, in part, a function of lake morphology. It was anti-
cipated that correlations between certain parameters
and lake morphometric data were likely.
   Mean Site 1 (the deepest site at each lake) sediment
values for  all constituents were computed for each
lake;  these were  regressed  against mean  morpho-
metric data (such as surface area, maximum depth,
drainage area, storage capacity, retention time, mean
depth) to detect simple linear relationships. Retention
time was the  single most important morphological
variable accounting for variance in sediment consti-
tuent concentrations. Notably,  organic carbon  (total
Kjeldahl nitrogen, COD, and volatile solids), lead, and
mercury concentrations were strongly correlated with
retention time; however, these relationships (as are all
those demonstrated by regression analysis) were not
necessarily cause and effect. In fact, the lead to reten-
tion  time relationship  may  have been largely fortu-
itous. Highest sediment  lead concentrations  may be
attributable largely to atmospheric fallout (implicated
also in Kemp et al.'s 1976 study of Lake Erie) in  lakes
surrounding the Chicago area. With the exception of
Skokie Lagoons, only glacial lakes were sampled in
the Chicago area,  and since as a group glacial  lakes
exhibited typically greater retention times (i.e., mean
retention time  for glacial lakes was 3.89 years  con-
trasted to a grand mean  for all  lakes of 1.39 years), it
was not clear, for glacial lakes, whether retention time
or location exerted the greater effect.

Data Utilization

The 1979 lake sediment data base has been used to
determine which lakes are receiving unusually high
constituent loadings, and to target areas where addi-
tional  monitoring  (such  as  fish flesh analysis)  or
remedial actions are needed. Results have been used
to develop lake  and stream sediment sampling stra-
tegies  and guidelines concerning  disposal of lake
dredge material. Analysis of  surficial sediments has
additionally proved useful as a screening device for
detection  and   identification  of contaminants  not
readily detected  by routine water quality sampling pro-
cedures.
  To further facilitate interpretation of data resulting
from chemical analysis of lake surficial sediments, a
classification   of  Illinois  lake  sediments  was
developed based on mean constituent concentrations
and  respective  standard  deviations. The  resultant
four-tier  classification  system  categorized  lake
sediments  as below  normal, normal, elevated, and
highly elevated (Table 4).

ACKNOWLEDGEMENTS: The data presented represent the
coordinated effort of numerous individuals of the IEPA. The
overall monitoring effort for 1979 was coordinated by D. Sef-
ton. Regional monitoring supervisors R. Schacht, W. Tucker,
and R.  Hite were  responsible  for field collection. Analyses
were performed in IEPA Springfield and Champaign labs
under direction of J. Hurley and R. Frazier, respectively. J.
Hardin  supervised entry of  data  into STORET and was
responsible for interfacing all data with the SAS computer
package. D. Schaeffer developed and modified  numerous
programs for use on Textronixs desk top computer. Various
drafts were typed  by M. Kinsall and B. Richards.


REFERENCES

Allum, M.O., R.E.  Glessner, and J.H. Gakstatter. 1977.  An
  evaluation of National Eutrophication Survey data. Natl.
  Eutroph. Surv. Work. Pap. No. 900. Corvallis Environ. Res.
  Lab. U.S. Environ.  Prot. Agency.
                                                  252

-------
                                                                                               SEDIMENT ANALYSIS
Andersen, V.M. 1974. Nitrogen and phosphorus budgets and
  the role of sediments in six shallow Danish lakes. Arch.
  Hydrobiol. 74:528-50.
Boland, D.P., et al. 1979. Trophic classification of selected
  Illinois water bodies.  EPA-600/3-79-123. Environ. Monitor.
  Systems Lab. U.S. Environ. Prot. Agency, Las Vegas, Nev.

Bortleson, G.C., and G.F. Lee. 1974. Phosphorus, iron, and
  manganese distribution in sediment cores of six Wiscon-
  sin lakes. Limnol. Oceanogr. 19:794-801.
Brunskill, G.J., D. Povolodo, B.W. Graham, and M.P. Stainton.
  1971. Chemistry of surface sediments of sixteen lakes in
  the Experimental  Lakes Area, Northwestern  Ontario. J.
  Fish. Res. Board Can. 28(2):277-94.
Fillos, J., and W.R. Swanson. 1975. The release rate of nutri-
  ents from river  and lake sediments. J. Water Pollut. Con-
  trol Fed. 47(5):1032-41.
Frink, C.R. 1969. Chemical and mineralogical  characteristics
  of lake sediments. Soil Sci. Soc. Am. Proc. 33:369-72.
Gorham, E.J., W.G.  Lund, J.E. Sanger, and W.E. Waeter Jr.
  1974. Some relationships between  algal  standing crop,
  water chemistry,  and sediment chemistry in the English
  Lakes. Limnol. Oceanogr. 19(4):601-17.
Howeler, R.H. 1972.  The oxygen status of lake sediments. J.
  Environ. Qual. 1(4):366-71.
Illinois Pollution Control Board. 1977. Rules and Regulations,
  Chapter 3: Water  Pollution.
Jernelov, S. 1970. Release of methyl mercury from sediments
  with  layers containing inorganic  mercury at  different
  depths. Limnol. Oceanogr. 15:958-60.
Kemp, A.L.W., R.L Thomas, C.I. Dell, and J.M. Jaquet. 1976.
  Cultural impact on the geochemistry of sediments in Lake
  Erie. J. Fish. Res. Board Can. 33:440-62.
Kelly, M.H., and  R.L.  Hite.  1981. Chemical analyses  of
  surficial sediments from 63 Illinois lakes, summer 1979. III.
  Environ. Prot. Agency.
Mathis, B.S., and N.R. Kevern. 1975. Distribution of mercury,
  cadmium, lead  and thallium in a eutrophic lake.  Hydro-
  biologia 46(2):207-22.
Pita, F.W. and N.V. Hyne. 1975. The depositional environment
  of zinc,  lead, and cadmium in reservoir sediments. Water
  Res. 9:701-6.
Sarkka,  J., M.L Hattula, J.  Junatuinen, and J. Passivirta.
  1978.  Mercury  in sediments of Lake Paifanna, Finland.
  Bull. Environ. Contam. Toxicol. 20:332-9.
SAS Institute. 1979. SAS User's guide. SAS Institute  Inc.
Schoettle, M., and G.M. Friedman. 1974. Short communica-
  tion: effect  of  man's activities on distribution of trace
  elements in  sub-bottom sediments of Lake George, New
  York. Sedimentology 21:473-8.
Sefton, D.F. 1978. Assessment and classification of Illinois
  lakes. Vol. I. III. Environ. Prot. Agency. Springfield.
Sefton, D.F., M.H. Kelly, and M. Meyer. 1980. Limnology of 63
  Illinois Lakes, 1979. III. Environ. Prot. Agency. Springfield.
Thomas, R.L, and J.M.  Jaquet. 1976. Mercury in the surficial
  sediments of Lake Erie. J. Fish. Res. Board  Can. 33:404-12.
Wildung, R.E.,  R.L. Schmidt, and R.C. Routson. 1977. The
  phosphorus status of eutrophic lake sediments as related
  to  changes  in  limnological  conditions—phosphorus
  mineral components. J. Environ. Qual. 6(1):100-4.
Williams, J.D.H., J.M. Jaquet, and R.L Thomas. 1976. Forms
  of phosphorus  in the surficial sediments of Lake  Erie. J.
  Fish. Res. Board Can. 33:413-29.
                                                       253

-------
 THE  ENGINEERING  CHARACTERISTICS OF HYDRAULICALLY
 DREDGED LAKE MATERIALS
 JAMES E. WALSH
 Baystate Environmental Consultants, Inc.
 East Longmeadow, Massachusetts

 STANLEY M. BEMBEN
 Goldbert-Zoino and Associates, Inc.
 Newton Upper Falls, Massachusetts

 CARLOS CARRANZA
 Baystate Environmental Consultants, Inc.
             ABSTRACT

             A perception of the engineering characteristics of dredged lake materials is necessary for making
             decisions regarding the size and type of hydraulic dredge best suited for the proposed project, dredge
             production rates, containment area size requi'ements, various containment area operating procedures,
             the nature of potential containment area reuse, the time necessary for the filled containment area
             to become available for reuse, and potential measures that may be used to increase the rate of im-
             provement and ultimate strength characteristics of the dredged material in the containment area. The
             pertinent engineering characteristics are reasonably inferred from bulking, grain size distribution, specific
             gravity and organic content, Atterberg Limits;, consolidation and shear strength data. Typical ranges
             for these data,  including before and after dredging cases, as determined for several sites in the nor-
             theastern United States are presented. Significant differences in measured characteristics often oc-
             cur among samples taken from the lake bottom, from the dredge effluent pipeline or from the contain-
             ment area. The most important changes in dredged material composition resulting from the dredging
             and dredged material disposal process are the  loss of fines and low specific gravity material. This
             change should be accounted for when planned dredging projects employ engineering data obtained
             from samples of lake material to  be dredged.
INTRODUCTION

The  engineering  characteristics  of dredged  lake
materials must be known to make decisions regarding
(1) the size and type of hydraulic dredge best suited for
the proposed project, (2) dredge production rates, (3)
containment area size requirements, (4) various con-
tainment area operating procedures, (5) the nature of
potential  containment  area  reuse, (6)  the  time
necessary for the filled containment area to become
available for reuse, and (7) potential measures for in-
creasing the  rate  of improvement and  ultimate
strength characteristics of the dredged material in the
containment area.
  The  engineering characteristics  of dredged  lake
materials assume greater importance in the design
process for sites with  limited containment areas.
These  characteristics may  be reasonably  inferred
from  the  bulking factor,  grain  size  distribution,
specific gravity, and organic content, Atterberg limits,
consolidation and shear strength  data. This work
presents typical ranges for these data as determined
for several sites in the Northeast.
PROJECT SAMPLE SITES

The locations of the project sample sites are listed in
Table 1. These sites were selected because of ongoing,
or recently completed dredging operations or because
dredging had  recently been suggested to alleviate ex-
isting problems  within  the  particular  lake. Where
dredging was ongoing, samples were taken from the
dredge effluent pipeline and  from the  containment
area. At those sites where dredging had recently been
completed, samples were obtained from the contain-
ment areas.
             Table 1.—Project locations.
  Dredging proposed

  Watershops Pond
  Springfield, Mass.
  Lake Warner
  Hadley, Mass.
  Bass Pond
  Springfield, Mass.
  1860 Reservoir
  Wethersfield, Conn.

  Dredging Completed
  Lake Barcroft
  Fairfax, Va.
  Lake Somerset
  Far Hills, N.J.
///. Dredging in progress

   Jacklyn Pond
   Bristol, Conn.
   Forest Pond
   Southington, Conn.
   Jefferson Lake
   Jefferson Lake State Park
   Richmond, Ohio
   Tomlinson Run Lake
   Tomlinson Run State Park
   Hancock County, W.Va.
   Lake Needwood
   Lake Needwood State Park
   Derwood, Md.
   Medford Lake
   Medford, N.J.
                                                254

-------
                                                                                      SEDIMENT ANALYSIS
ENGINEERING CHARACTERISTICS

Grain Size Distribution. This has been determined for
typical representative samples obtained from  each
site. Table 2 lists the average effective diameters (D10)
and  uniformity coefficients  (U)  for lake  bottom
materials  obtained  at  11  sites  and  for  dredged
materials obtained from the seven sites with contain-
ment areas.
  In general, the grain size distributions of samples
taken within a given  containment  area showed a
higher percentage of fines for those samples obtained
near the outflow structure than for those samples ob-
tained near the inflow structure.
  The average grain size distribution curves for the
containment area dredged material and the lake bot-
tom sediment (obtained near the dredge) for both the
Tomlinson Run and Jefferson Lake sites are shown in
Figure 1. These curves indicate that some of the fines
of the lake bottom sediments were lost in the dredging
and dredged material  disposal process. This effect
was also noted at other sites. The magnitude of this
effect appears to be a function of the percentage of
the lake bottom material finer than approximately 0.01
mm, the dredged material specific gravity, and the effi-
ciency of the containment area design and operation.
The overall efficiency  of  the dredging and disposal
operations may be defined as the ratio of the weight of
material  retained  in  the  containment area to  the
weight of the material dredged.
  Samples containing fibrous organic matter usually
resulted in a grain size distribution resembling that of
a typical sand. The settlement characteristics of these
sediments will  not, however, even closely approximate
those of a sand with a  similar grain size distribution
curve. To avoid a misrepresentation of the material
properties, the grain size distribution samples con-
taining significant amounts of fibrous organic matter
should not be presented without accompanying infor-
mation regarding the sample's organic content.
  Atterberg limits.  The liquid  limits (WL) and plastic
limits (WP) were determined for 13 lake bottom sedi-
ment and containment dredged  material  samples.
 These values, as well as the plasticity index (PI) are
 listed  in Table 3. All of the samples that exhibited
 limits fall below the A-line of the Casagrande plastici-
 ty chart, and thus, are classified as silts. In addition,
 most of these limits indicate a high compressibility.
    Specific gravity and organic content. The apparent
 specific  gravities of the  solids  portion  (including
 organics) of the  17 lake bottom sediments and con-
 tainment area dredged material samples are listed in
 Table 4. These ranged from 1.35 to 2.73 with a mean
 value of 2.25  and a standard deviation of 0.39. The
 average apparent specific gravity of the lake bottom
 sediments was 2.19. The average apparent specific
 gravity of the containment area dredged materials
 was 2.41. The generally higher specific gravities of the
                        TOMLINSON RUN LAKE

                        •  LAKE BOTTOM
                        •  CONTAINMENT AREA

                           JEFFERSON LAKE

                        o  LAKE BOTTOM
                        a  CONTAINMENT AREA
   100

    90

h-  80

U  TO


{j  50
z

140
I  30

a.  20

     10
         0.1                 0.01              0.001
              GRAIN SIZE IN MILLIMETERS
  Figure 1.—Grain size distributions for lake bottom and con-
  tainment area materials.
     Table 2.—Effective diameters (D10) and uniformity coefficients (U) of lake bottom and containment area sediments.
 Sample
Effective
Diameter (D10)
(mm)
                                    Uniformity
                                    Coefficient (U)
 Lake bottom material

 Watershops-Dan Baker Cove
 Watershops-main section
 Tomlinson Run Lake
 Mountain Lake-inlet
 Mountain Lake-dam area
 Medford Lake
 Lake Warner
 Jefferson Lake
 Bass Pond
 Forest Pond
 1860 Reservoir

 Containment area material

 Tomlinson Run
 Lake Somerset
 Lake Needwood
 Lake Barcroft
 Jefferson Lake
 Medford Lake
 Jacklyn Pond
0.002
0.001
0.007
0.070
0.034
0.039
0.015
0.003
0.140
0.180
0.001
0.014
0.060
0.009
0.009
0.007
0.060
0.008
                                     3.4
                                    10.7
                                    10.0
                                     3.1
                                     7.0
                                     4.1
                                    10.0
                                    17.3
                                     2.4
                                     2.6
                                     5.4
                                    17.1
                                     1.6
                                     2.1
                                     1.4
                                    10.7
                                     4.7
                                     7.8
                                                  255

-------
LAKE AND RESERVOIR MANAGEMENT
containment  area materials reflect the loss  during
dredging of some of the organic fraction of the lake
bottom materials.
   Figure 2 shows the apparent specific gravity versus
the organic content for the containment area dredged
material and the lake bottom materials. For reference,
three  curves  of  organic  content  versus apparent
specific  gravity  are  shown for  mineral  spec fie
gravities of 2.85 and organic specific gravities of 0.50,
1.00, and 1.50, respectively.
   In general, low values of organic specific gravity
coincide with samples that had a growing organic
fraction  and probably  reflect  entrained  gases.
Samples with higher organic  specific gravity  values
generally contained dead organic matter.
   The  organic contents of the 11 lake bottom sedi-
ment samples and seven containment area dredged

 Table 3.—Atterberg limits of lake bottom and containment
                  area sediments.

Lake bottom material           WL       WP       PI

Watershops-Dan Baker
Cove                        115       99       16
Watershops-main section        56       50        6
Tomlinson Run                42       31       11
Medford Lake                 130       99       31
Jefferson Lake                44       32       12
Forest  Pond                  101       97        4
Mountain Lake                 000

Containment area material

Lake Somerset                27       25        2
Lake Barcroft                  69       48       21
Jefferson Lake                72       53       19
Jacklyn Pond                  94       89        5
Tomlinson Run                 000
Lake Need wood                25       0        0
                                                     material samples were determined by loss of weight
                                                     upon ignition; these are shown in Table 4. The organic
                                                     content is calculated as the loss of weight upon igni-
                                                     tion divided by the original total dried weight.
                                                       Consolidation properties. Consolidation  tests were
                                                     performed on relatively undisturbed  samples of the
                                                     various lake bottom and  containment area  dredged
                                                     materials. Each load increment was maintained until
                                                     the void ratio versus logarithm of time (e-log t) plot
                                                     indicated the completion of primary consolidation.
                                                     The final load of 200 kN per square meter was allowed
                                                     to remain for a 24-hour period to determine the secon-
                                                     dary consolidation characteristics of the sample. The
                                                     void ratio versus logarithm of effective stress (e - log
                                                     p) curves for four samples is shown  in Figure 3.  For
                                                     each sample, the compression index Cc was deter-
                                                     mined by taking the slope of a "best fit" straight line
                                                     through the last five loading increments on  the e - log
                                                     p curve. The Cc values for each material are given in
                                                     Table 4. The compression index values range from a
                                                     high of 2.7  for  the highly compressible Forest Pond
                                                     peat to a low of 0.07 for the Tomlinson Run  contain-
                                                     ment area fine, silty sand.
                                                       Figure 4 shows a plot of the compression index ver-
                                                     sus the liquid limit for those samples that had limits.
                                                     The trend of the data is clearly similar to that reported
                                                     by  Salem and  Krizek (1973)  for river sediments for
                                                                 o  FOREST  POND
                                                                 a  MEDFORD LAKE
                                                                 •  JEFFERSON  LAKE
                                                                 •  MT LAKE-INLET
   60
   SO
           6^2.85  Gm=2.8S
           GO= 1.00  GO = 150
                          Gm = SPECIFIC GRAVITY OFl
                               MINERAL  PORTION

                          GO = SPECIFIC GRAVITY CF
                               ORGAN C PORTION  |l
   30
u
   20
    10
     1.0      1.5       2.0      2.5       3.0
           APPARENT SPECIFIC GRAVITY (Gg)

Figure 2.—Apparent specific gravity of dredged lake mater al
versus organic content.
                                                          0.05   .10    .20.30  .50   11.0

                                                                  PRESSURE
                                                                                              1 2.0
                                                   Figure 3.—Void ratio versus logrithm of pressure curves for
                                                   lake bottom and containment area materials.
                                                256

-------
                                                                                     SEDIMENT ANALYSIS
Toledo, Ohio and Monroe, Mich. This trend is describ-
ed by the empirical equation:
              Cc = 0.02WL - 0.44
(1)
in which WL is the liquid limit expressed as a percen-
tage. Although questionable for samples with high li-
quid limits, the data developed in this study indicate
that the empirical relationship may be extended over a
broader range of samples, especially those with liquid
limits in the range of 25 to 60 percent.
  The coefficient of consolidation (cv) was calculated
for  each of the  consolidation tests. The values are
given in Table 4.  The coefficient of consolidation may
be used to approximate the rates of settlement for
containment areas where the drainage path distances
can be estimated.

  Bulking considerations. When lake sediments are
dredged and resedimented  in  a containment area,
they often assume a more open-work structure, and
hence, a greater volume than that exhibited In situ.
This increase in volume upon resedimentation  is
known as bulking. Bulking often  occurs in fine grained
and/or organic soils.
  The bulking factor, B, of a lake bottom material may
be given as the  ratio  of a given volume  of dredged
material upon resedimentation VCA, to the volume oc-
              SALEM  ft  KRIZEK  (1973)
                          Cc«(0.02wt-a44
                                      40             50            60

                                            LIQUID  LIMIT   (OIL)

Figure 4.—Compression index (Cc) versus Liquid Limit (w^ for dredged materials.
Table 4.—Specific gravity, organic contents, and consolidation properties of lake bottom and containment area sediments.
Sample
Watershops-Dan Baker Cove
Watershops-main section
Tomlinson Run Lake
Tomlinson Run State Park
Watershops-main section
Mountain Lake-inlet
Mountain Lake-dam area
Jefferson Lake
Bass Pond
Forest Pond
1860 Reservoir
Mean
Standard Deviation
Specific
Gravity
(dimensionless)
1.35
2.63
2.61
2.69
2.09
2.10
2.18
2.55
2.17
1.79
1.93
2.19
0.41
Organic
Content
(%)
26
11
6
2
9
35
10
11
4
34
16


Compression
Index
(dimensionless)
1.13

0.52
0.04
0.09
1.20
0.24
0.36
0.11
2.73
2.40


Coefficient of
Consolidation
(10-3m2/day)
83.6

5.6
74.3
9.3
4.7
2.8
5.6
6.5
0.9
1.9


Containment Area Material

Tomlinson Run                   2.66
Lake Somerset                   2.12
Lake Needwood                  2.73
Lake Barcroft                    2.20
Jefferson Lake                   2.65
Medford Lake                    2.61
Jacklyn Pond                    1.88
                     Mean      2.41
         Standard Deviation      0.33
      2
      15
      4
      8
      6
      2
      52
              0.07
              0.28
              0.10
              0.43
              0.60
              1.20
              2.60
18.6
 0.9
 4.7
 8.4
 8.4
 4.7
 0.9
                                                 257

-------
 LAKE AND RESERVOIR MANAGEMENT
 cupied by the same weight of in situ  material, VLB.
 Thus, in general,
                   B =
                        'CA
                        'LB
                 (2)
 and also, for saturated conditions,
               B =
 where Gs is the apparent specific gravity of the solids,
 WCA  is the water content of the dredged  material
 resedimented in the containment area, and WLB is the
 in  situ water  content of  the  lake bottom material.
 Water content is defined as weight of water to weight
 of solids, expressed as a percentage.

  Table 5.—Bulking factors for various dredged materials.
Johnson Huston
Material (1976) (1970)
Clay 1.2-3.1 1.45
Sandy clay 1.25
Silt 1.1-1.4 2.0
Sandy silt
Sand 1.0-1.2 1.0
Organic silt 2.0
'Lake Warner
"Mountain Lake
'1860 Resrvoir
Walsh anc
Bemben*



1.41
1.0s
1.83



 'Determined after 24 hours of settling Height of solids portion approximalely
 1-1 5 feet
                                               Several  sedimentation column  tests were run on
                                             samples of  bottom  sediments using 10-foot  high,
                                             5.5-inch  I.D.  cast  acrylic columns. These sediments
                                             were mixed with water from their  respective lakes in
                                             concentrations  representative of  dredged  material
                                             slurries. Following a 24-hour quiescent settling period,
                                             the water content of the solids phase was measured
                                             and compared to the in situ water content for each
                                             sample. The average bulking factor, B, ranged from 1.0
                                             for the Mountain Lake sand to 1.8  for the  1860 Reser-
                                             voir silt.

                                               Bulking  factors for various materials are listed in
                                             Table 5. Because of the relative ease of performing
                                             sedimentation  column  analyses, the containment
                                             area design  bulking factor should be determined ex-
                                             perimentally rather than from literature ranges.

                                               It should be  recognized that bulking is largely a
                                             reflection  of the material's  response  to effective
                                             stress. Therefore, the bulking factor  as  determined
                                             from the column tests on a given sediment will vary
                                             with the height of the solids layer, the time allowed for
                                             primary consolidation to take place, and the seepage
                                             conditions. Thus, test column heights, drainage, and
                                             test period times should represent  expected field con-
                                             ditions.

                                               Walsh and Bemben (1977) give experimental  ex-
                                             amples of the  effects of downward  seepage  in in-
                                             creasing the effective stresses throughout the depth
                                             of material in the containment area. This produces ex-
                                             tra settlement and, hence, greater storage capacity.
                                             Other factors affecting bulking may include floccula-
                                             tion caused  by  changing ionic concentrations in the
                                             water,  secondary compression,   desiccation,  and
                                             decomposition  of the organic component. For field
                 UPPER  BASIN
                   MIDDLE  BASIN
                                                                       LOWER BASIN
     0.0
     Q5
  I  I.O
  t
     1.3
             /

               2   0  O.fO 0~.20 4OO 600    I     20   0.10 0.20 400  600   I     2  0   0.10  0.20  400  600
    St        Sv

(
-------
design, the  bulking factor must  be modified by the
dredging and disposal operations efficiency factor.
  Vane shear  strength. Hand vane shear strength
measurements were  taken at seven dredged  lake
material containment areas. The time since deposi-
tion of the  dredged material ranged from a  few
minutes (Lake  Needwood and Tomlinson Run) to 14
months (Lake Barcroft,  containment area No. 1). The
depth of the dredged material in these containment
areas ranged from .5 to  2 meters,  and averaged about
1 meter.
  Vane shear  strength measurement values varied
from about  0.02 to .25 (kN/m2) throughout the seven
areas. Usually, the values at a given elevation within a
basin  increased  with increasing distance  from the
overflow structure. This  is to be  expected  as the
coarser  materials  with  more  rapid  drainage
characteristics will settle out nearer to the discharge
pipeline and consolidate more rapidly.
  Vane shear strength versus depth data is shown for
the Jacklyn Pond containment area in Figure 5. The
area consisted of three small basins, one flowing into
the other. Neither shear strength nor water content
differed significantly for  the materials  in the three
basins. The effects of  surface desiccation were  ap-
parent in the upper and lower basins.
CONCLUSIONS

The engineering behavior of  hydraulically  dredged
lake materials may be reasonably inferred from grain
size, Atterberg limit, specific gravity, organic content,
                                 SEDIMENT ANALYSIS

and sedimentation  column data for the  lake bottom
materials. Typical  ranges for  these parameters  as
determined  for  several  lakes in the  northeastern
United States have been  presented.  This information
may be used for the preliminary feasibility evaluation
of lake dredging projects in order to assess contain-
ment  area design,  size requirements, operating pro-
cedures, and potential reuse. Site specific knowledge
of these  engineering  characteristics is, of course,
necessary for final design.

ACKNOWLEDGEMENTS: Much of the work presented here is
a portion of a more extensive study funded,  in part, by an
award to Dr.  Bemben from the Office of Water Resources
Research, Department of the Interior, as authorized under
the Water Resources  Research Act of 1964. Additional fun-
ding was provided by the Mud Cat Division of National Car
Rental, Inc., by a grant to Dr. Bemben. The results of that
study are available as Walsh and Bemben (1977).


REFERENCES

Huston, J.  1980. Hydraulic Dredging. Cornell Maritime Press,
  Inc., Cambridge, Md.
Johnson,  L.D. 1976. Mathematical  model for  predicting  the
  consolidation of dredged material in confined disposal areas.
  U.S. Army Eng. Waterways Exp. Sta., Vicksburg, Miss.
Salem,  A.M., and R.J. Krizek. 1973. Consolidation character-
  istics of dredging slurries. J. Waterways Harbors Coastal
  Eng.  Div. Am. Soc. Civil Eng. 99 (WWA).
Walsh, J.E., and S.M. Bemben. 1977. Disposal and Utilization of
  Hydraulically Dredged  Lake Sediments  in Limited Contain-
  ment Areas. Publ.  92. Water  Resour.  Res. Center, Univ.
  Massachusetts, Amherst.
                                                   259

-------
                                   Comparative  Analysis
                                                          of   Reservoirs
REGIONAL COMPARISONS OF LAKES AND RESERVOIRS:
GEOLOGY, CLIMATOLOGY, AND MORPHOLOGY
KENT W. THORNTON
Ford, Thornton, Norton & Associates,  Ltd.
Little Rock, Arkansas
           ABSTRACT

           Lakes and reservoirs typically have been considered synonomous; in fact, Hutchinson classifies reser-
           voirs as lake type 73. Processes such as internal mixing, redox reactions, nutrient cycling, and primary
           production obviously occur in both lakes and reservoirs. The forcing functions or driving variables
           for lakes and reservoirs, however, may not be identical so the response of these two systems may
           be different. Regional considerations of the climatic, geologic and geographic differences between
           lakes and reservoirs indicate why our understanding and available predictive techniques for lakes
           should be tempered for the proper management or reservoirs. Lakes and reservoirs generally are
           distributed in different parts of the U.S. Most lakes occur in the glaciated portion of the U.S. while
           most reservoirs are located in the southeastern, central, southwestern, and western U.S. Geologic
           differences in these areas have implications for differences in the loading of dissolved and paniculate
           constituents to lakes and  reservoirs. Climatic differences such as precipitation-evaporation interac-
           tions also result in distinct areas. In the eastern U.S., precipitation exceeds evaporaton, water is plentiful
           and lakes are prevalent. In the western U.S., evaporation exceeds precipitation, water is scarce and
           reservoirs are prevalent. Geographic differences influence the management of lakes and reservoirs.
           Operation of reservoirs for hydropower, irrigation, industrial and public water supply can influence
           the response of the system to external and internal inputs. The distribution of lakes and reservoirs
           in conjunction with geologic, climatologic, and geographic patterns would imply potential differences
           in the limnological response of lakes and reservoirs. Proper management of our water resources re-
           quires that we consider these potential differences in the decision process.
INTRODUCTION

Lakes

Limnologists have historically studied lakes and small
streams.  Lakes  Zurich,  Mendota,  Windemere,
Untersee, Obersee, and Lindsey Pond are well known
throughout the limnological world and provide for our
present understanding of the structure and function of
lakes. As a discipline apart from natural history, how-
ever,  limnology is a young  science. The classic
studies  of Birge and Juday,  Gessner, Hutchinson,
Hynes, Mortimer, Muller, Nauman, Ohle, Ruttner, and
Thienemann were conducted primarily during this cen-
tury. The study of large rivers and  reservoirs, until
recently, has not contributed to our understanding of
lentic and lotic systems.
  The  distribution  of  about 2,300 natural  lakes
throughout the world exhibits a trimodal distribution
(Schuiling, 1976) (Fig. 1). The majority of lakes are
distributed between latitudes 35-55° in both the nor-
thern and southern hemisphere with a maxima in both
                                             261

-------
LAKE AND RESERVOIR MANAGEMENT
         60      40
 Figure 1.—Latitudinal distribution of over 2300 natural lakes
 throughout the world (after Schuiling, 1976).
 around 45° N latitude represents  reservoirs  on the
 Missouri  and Columbia River systems and some
 smaller New England  projects. While  this reservoir
 distribution represents only CE reservoirs, TVA and
 many Bureau of Reclamation reservoirs also occur at
 these latitudes.
   Geology  and climatology differ along  latitudinal
 and  longitudinal  gradients. These two factors and
 their interaction with basin  morphology influence the
 limnological response of freshwater systems.  If these
 factors differ between  lakes and reservoirs, perhaps
 the response of lakes and reservoirs also differs. It is
 the system response, not the limnological processes,
 that are being considered. It is assumed the limnolo-
 gical  processes  of primary production, respiration,
 decomposition, and  so on  are similar in  lakes and
 reservoirs. However, these processes also are similar
 in lakes and streams but the responses of these two
 systems, lentic versus lotic, are certainly distinct.
hemispheres around 48° latitude. The third mode ex-
tends from  15° N  to 20° S  latitude  with a  maxima
around the equator. The minima in both the northern
and southern hemispheres occurs from approximately
20-40° latitude.
  Our limnological knowledge and experience is bas-
ed primarily  on  the study  of lakes  located around
latitude 45°N. These lakes are natural features of the
landscape created through geologic time and forces.

Reservoirs

Reservoirs  are constructed to store  water for flood
control, municipal,  industrial, and agricultural  water
supply,  hydroelectric power generation,  or other pur-
poses  for which they are  designed. Reservoirs, in
many cases, are constructed where available  water
supply is not adequate. Although reservoirs have been
dated to 400-300 B.C. (Biswas, 1975), U.S. reservoirs
are relatively new  features on the landscape.  Many
reservoirs in the United States are less than 50 years
old. Morris Dam  in Tennessee, one of the first reser-
voirs in the TVA system, was finished in 1936. Since
reservoirs are engineered systems, engineering con-
siderations of firm yield, spillway design for maximum
probable floods,  sediment trap efficiency, and similar
factors  have received considerable study. Reservoir
limnology, however, has not had a similar focus and
analysis.
  It generally is assumed that lakes and reservoirs are
synonomous; in  fact, reservoirs have been classified
as lake type 73 (Hutchinson, 1957). Knowledge of lake
limnology has been considered sufficient to under-
stand reservoir limnological structure and function.
As an  initial evaluation of  this assumption, we can
review the distribution of reservoirs and lakes  in the
United States.
  A distributional  comparison of U.S.  natural lakes
and Corps of Engineers (CE)  reservoirs sampled dui-
ing EPA's National Eutrophication Survey (1972-1975)
indicated a  bimodal distribution  of lakes with  a
unimodal distribution of reservoirs (Fig. 2). The reser-
voir  maxima generally corresponds with the lake
minima. The majority of lakes occur  in the glaciated
portion of the United States with the secondary mode
representing  Florida solution lakes. The distribution
of U.S.  lakes corresponds with the distribution of
lakes in the northern hemisphere (Fig. 1). The  majority
of reservoirs are located throughout the southeastern,
central, southwestern, and western United States. The
small secondary mode in the reservoir distribution at
 Figure 2.—Latitudinal distribution of 309 natural lakes and
 107 CE reservoirs sampled during the NES.
  There are two  purposes for this paper.  First, the
hypothesis:
    H0: Lake Response ~ Reservoir Response
will be reviewed.  If this null hypothesis is rejected,
then  alternative  approaches may  be required to
predict reservoir water quality and manage reservoir
ecosystems. The second purpose is to provide a back-
ground for subsequent papers in this  chapter con-
trasting lakes and reservoirs.
GEOLOGY

Watershed  geology  influences  lake and reservoir
water quality since the major mechanisms controlling
water chemistry are precipitation, dominant rock for-
mations,  and  the  evaporation-crystallization  pro-
cesses (Gibbs, 1970). Two water quality variables that
indicate the dissolution and transport  of dissolved
and particulate constituents from the watershed are
total dissolved solids (TDS) and suspended sediment
(SS).
  Longitudinal and latitudinal geological differences
can be inferred by examining the annual  average TDS
and SS concentrations  in  freshwater streams. TDS
concentrations generally were higher in the western
United States than the eastern United States (Fig. 3).
The  northeastern  and  southeastern areas  of the
United States  have low TDS concentrations,  which
also corresponds  with  low alkalinities throughout
these regions (Omernick and Powers, 1982). TDS con-
                                                 262

-------
                                                                     COMPARATIVE ANALYSIS OF RESERVOIRS
centrations  are considerably higher  in  the  Great
Plains and southwestern United States.
  Streams located in geographic areas where natural
lakes predominate generally have lower SS concentra-
tions than streams in major  reservoir zones (Fig. 4).
The  upper Midwest, Appalachian region,  and  north-
eastern U.S.  have relatively low  stream SS con-
centrations.  The Piedmont area of the southeast, the
Great Plains,  and southwestern United States have
higher stream SS concentrations and are located in
unglaciated  areas.
  These two variables, TDS and SS, have implications
for  light penetration and water clarity, nutrient and
contaminant transport, and productivity in lakes and
reservoirs.
CLIMATOLOGY

Precipitation patterns are directly related to geolog-
ical weathering  and hydrologic  transport.  Precipita-
tion-evaporation interactions result  in  two distinct
areas  in the  United  States In  the eastern United
States, precipitation generally exceeds evaporation,
water supply is  plentiful,  and lakes  are prevalent  in
glaciated areas (Fig. 5).  In the western United States,
evaporation exceeds precipitation, water generally is
scarce, and reservoirs  are prevalent (Fig.  5). These
precipitation-evaporation  relationships   are  ex-
emplified by the riparian doctrine in the east and prior
appropriation doctrine in the west for water usage.
   Precipitation intensity and duration also can be im-
portant factors in geological weathering and hydro-
logic transport of various water quality constituents. A
comparison of the annual average number of days
with thunderstorms indicates the southern and south-
western States and Florida have a high thunderstorm
frequency (Fig. 6). This area is influenced by the warm
humid air from the Gulf of Mexico. Excess water (pre-
cipitation-evaporation) in  the South and  Southeast,
therefore,  is  relatively  abundant.  Orographic  in-
fluences in the Smokey and Rocky Mountains also are
apparent. While thunderstorm  frequency is  similar
over a major portion of the United States, it must be
remembered water is abundant in the eastern  United
States and scarce in the western  United States.
MORPHOLOGY

Morphometric characteristics of both lentic and lotic
systems are recognized as playing a significant role in
the limnological response of aquatic systems. Mean
depth, for example, has  been implicated in relation-
ships  ranging  from  nutrient  loading  models   to
morphoedaphic indices.
   A comparison of morphometric characteristics for
NES natural lakes and CE reservoirs indicated reser-
voirs had greater drainage and surface areas, drain-
age area/surface area (DA/SA) ratios, mean and maxi-
mum depths, shoreline development ratios, and areal
water loads than natural lakes (Table 1). The greater
DA/SA ratio indicates a  potential  for greater hydro-
logic and constituent transport and loading to reser-
voirs. This also is  reflected in the greater areal water
load and shorter  hydraulic  residence time of reser-
Figure  3.—TDS concentrations  (mg/l) in  streams located
throughout the United States.
Figure 5.—Precipitation-evaporation  balance in the  con-
tinental United States.
                                                                          . V
Figure 4.—Suspended  solids (SS) concentrations (mg/l) in    Figure 6.—Annual average number of days with thunder-
streams throughout the United States.                     storms.
                                                  263

-------
 LAKE AND RESERVOIR MANAGEMENT
  Table 1.—Comparison of morphometric characteristics of natural lakes and Corps reservoirs (after Thornton et al. 1981).
Variable
Drainage area (Km2)
Surface area (Km2)
Drainage/surface area (DA/SA)
Mean depth (m)
Maximum depth (m)
Shoreline development ratio
Areal water load (m/yr)
Hydraulic residence time (yr)
Natural
Lakss
(N = !I09)
222
5.Ci
33
4.5.
10.7'
2.9 (N == 34)1
6.J.
0.74
CE
Reservoirs
(N = 107)
3228
34.5
93
6.9
19.8
9.0 (N = 179)2
19
0.37
Probability
Means Are
Equal
< 0.0001
< 0.0001
< 0.0001
< 0.0001
< 0.0001
< 0.001
< 0.0001
< 0.0001
 'Hutchinson, 1957
 ' Letdy and Jenkins, 1977
 voirs. Greater shoreline development ratios in reser-
 voirs indicates there are probably more coves, ern-
 bayments, and  backwater areas  in reservoirs than
 lakes.
 DISCUSSION

 The limnological response of an aquatic ecosystem is
 not controlled or regulated by any single factor but re-
 presents the interaction and integration of all the ex-
 ternal and internal processes and forces. Since inter-
 nal processes such as primary production, nutrient
 cycling, consumption, and decomposition  are assum-
 ed to be similar in lakes and  reservoirs, we need 1o
 consider the  similarity of  external forces in deter-
 mining the response of the system. Similar processes
 also occur  in lakes and rivers but the limnological
 response of these two systems are markedly influenc-
 ed by different external forces. Water control,  for ex-
 ample,  represents one obvious difference between
 lakes and reservoirs.
  The interaction of these three  broad categories—
 edaphic, climatic, and morphometric—on the produc-
 tivity of lakes has been known for some time (Rawson,
 1955) and served as the original  basis for the study
 and  analysis  of eutrophication (Vollenweider,  1969,
 1976; Vollenweider and Kerekes, 1980). TDS has been
 used as a  surrogate variable for  lake productivity
 (Ryder, 1964) and to represent  dissolved nutrient con-
 tributions  from the watershed  (Hutchinson,  1957;
 Vollenweider,  1969; Stumm  and Morgan,  1971). The
 same abiotic  factors that influence TDS  concentra-
 tions are assumed to  affect phosphorus  concentna-
 tions (Kemp, 1971). SS concentrations not  only trams-
 port nutrients adsorbed to  the particles but also in-
 fluence the light regime of the system and potential
 utilization of nutrients by plankton. Reservoirs typical-
 ly occur in areas with high SS  concentrations.
  Precipitation  contributes  both   to  geological
weathering  and   nutrient   loading.  Storm  events
generally contribute  at least an order of  magnitude
more phosphorus than occurs during a comparable
baseflow period. One major storm  event  to DeGray
 Lake, Ark. contributed a phosphorus load comparable
to the entire  annual  phosphorus load during  base
flow. Since reservoirs  are designed to store  water,
storm flow and its  constituent  loads are generally re-
tained within the reservoir, particularly  where evapora-
tion  exceeds  precipitation.  Where precipitation is
abundant, many reservoirs store water for flood con-
trol.  Lakes generally have an inflow = outflow rela-
tionship.
   Morphometric relationships also support the poten-
 tial for greater SS and nutrient loads. Drainage areas,
 DA/SA ratios, and  areal water loads are greater for
 reservoirs than lakes. The greater shoreline develop-
 ment ratios indicate a potential for high biological pro-
 ductivity  (Wetzel,   1983).  Shoreline   development
 generally relates to  more extensive  littoral  areas,
 which are more productive than the pelagic areas of
 lentic  systems.
   Different geologic locations, larger drainage basins
 in areas with extensive geologic weathering of  both
 dissolved and particulate constituents,  storage  of
 constituent  loads,  and extensive coves and  em-
 bayments suggest  the range  of reservoir responses
 may  differ  from lake  responses.  Nutrient  loading
 models and  other  predictive  techniques developed
 from a cross-sectional data base of lake responses,
 therefore, may not be applicable for reservoirs or may
 have to be modified for reservoir application. Manage-
 ment strategies developed for lakes may need to be
 modified for effective use in reservoirs.
   The purpose of this paper was to provide circum-
 stantial evidence that reservoir responses may be dif-
 ferent  than  lake  responses because external forces
 may differ between these two types of systems. The
 intent  was  to caution  managers,  scientists,  and
 engineers to closely  examine the characteristics of
 the systems used to develop an approach or  tech-
 nique. If these systems are similar to the lake or reser-
 voir under study, then, the approach or technique may
 be applicable regardless whether it is a lake or reser-
 voir. Lake limnology has provided us with valuable in-
 sight  into functional relations  and processes occurr-
 ing within  lentic  systems. This  knowledge  must be
 placed in perspective with the unique characteristics
 of your lake or reservoir  for  the  development  and
 implementation of cost-effective and environmentally
 sound  management approaches.
REFERENCES

Biswas, A.K. 1975. A short history of hydrology. Pages 57-79
  in A.K. Biswas, ed. Selected Works in Water Resources.
  Int. Water Resour. Ass. Champaign, II.
Gibbs, R.J. 1970. Mechanisms controlling world water chem-
  istry. Science 170: 1088-90.
Hutchinson, G.E.  1957. A Treatise on Limnology. Vol. 1. John
  Wiley and  Sons, New York.
Kemp, P.H. 1971. Chemistry of natural waters. VI. Classifica-
  tion of waters. Water Res. 5: 945-56.
                                                 264

-------
                                                                             COMPARATIVE ANALYSIS OF RESERVOIRS
Leidy, G.R.,  and R.M. Jenkins.  1977. The  development of
  fishery compartments and population rate coefficients for
  use in reservoir ecosystem modeling. Contract rep. CR
  4-77-1. U.S. Army Corps Eng. Waterways  Exp. Sta., Vicks-
  burg,  Miss.
Omernick, J.M., and C.F. Powers. 1982. Total alkalinity of sur-
  face waters—a national map. EPA-600/D-82-333. U.S. En-
  viron. Prot. Agency, Corvallis, Ore.
Rawson, D.S. 1955. Morphometry as a dominant factor in the
  productivity of large lakes. Verh. Inst. Ver. Limnol. 12:
  164-75.
Ryder, R.A. 1964. Chemical characteristics  of Ontario lakes
  as  related to  glacial history. Trans. Am. Fish. Soc. 93:
  260-8.
Schuiling, R.D. 1976. Sources and composition of lake sedi-
  ments. Pages 12-18  In H.L. Golterman,  ed. Interactions
  Between  Sediments  and  Freshwater.  Dr.  Junk  B.V.
  Publishers. The Hague.
Stumm, W., and J.J. Morgan. 1971. Aquatic Chemistry. John
  Wiley and Sons, New York.
Thornton,  K.W.,  et al.  1981. Reservoir sedimentation  and
water quality—an heuristic model. Pages 654-61 in H.G.
Stefan, ed. ASCE Proc.  Symp. on Surface Water Impound. I:
654-61.
Visseman, W. Jr., J.W. Knapp, G.L. Lewis, and T.E. Harbaugh.
  1977.  Introduction to Hydrology. Harper and  Row, New
  York.
Vollenweider,  R.A. 1969.  Scientific fundamentals of the
  eutrophication of lakes and flowing waters, with particular
  reference to nitrogen and phosphorus as factors in eutro-
  phication. Organ. Econ. Coop. Develop.  Paris.
        . 1976. Advances in defining critical  loading levels
  for phosphorus in lake eutrophication. Mem. 1st. Ital. Idro-
  biol. 33:53-83.
Vollenweider, R.A., and J.  Kerekes. 1980.  The loading con-
  cept as a basis for controlling eutrophication—philosophy
  and  preliminary results of  the OECD programme  on
  eutrophication. Prog. Water Technol. 12: 5-38.

Wetzel,  R.G.  1983.  Limnology. Saunders Publishing Co.
  Philadelphia.
                                                       265

-------
LAKE-RIVER INTERACTIONS: IMPLICATIONS FOR NUTRIENT
DYNAMICS IN  RESERVOIRS
ROBERT H. KENNEDY
U.S. Army Corps of  Engineers
Waterways  Experiment Station
Vicksburg, Mississippi
             ABSRACT

             Unlike small drainage lakes which receive nutrient inputs from relatively diffuse sources, reservoirs
             and other river-fed lakes receive a majority of their nutrient income from a single large source located
             distant from the lake's outlet. Interactions between lake and river, which are governed by hydrology,
             the thermal structure of the lake, and lake mo'phology, will, therefore, play an important role in deter-
             mining the impact and ultimate fate of influent nutrients. These interactions may also affect the man-
             ner in which nutrients are recycled and/or distributed within the  reservoir or river-fed lake and lead
             to the establishment of longitudinal gradients, in concentration. Studies conducted at  DeGray Lake,
             Ark., and West Point Lake, Ga., both large Corps of Engineers  hydropower reservoirs, provide in-
             structive examples of some of the effects of these interactions. Both lakes exhibit longitudinal gra-
             dients in nutrient and chlorophyll concentrations which are influenced by the hydrologic characteristics
             of the major tributary. Exchanges of water and material between hypolimnion and epilimnion are also
             affected by flow regime. These exchanges play an important role in the seasonal dynamics of nutrients
             and metals.
INTRODUCTION

Comparisons of natural and manmade lakes suggest
a number of substantive differences, particularly with
respect to physical characteristics (Ryder, 1978; Bax-
ter, 1977; Thornton  et al. 1981; Walker,  1981).  As ,a
group, reservoirs are larger, deeper, and morpholog-
ically more  complex.  Watershed area to lake area
ratios are also greater for reservoirs than for natural
lakes. These characteristics, and the fact that reser-
voirs are commonly constructed on  relatively  large
rivers and  streams, account for observed differences
in water residence time and nutrient loading rates. In
marked  contrast to most natural  lakes, particular!/
drainage lakes, inflows to reservoirs commonly enter
via  a single  large  tributary  located  a considerable
distance from the point of discharge. The implications
of this are serious since nutrient loads, as measured
at the point of inflow, may be  significantly modified
with respect to both quantity and quality by processes
occurring in reservoir headwater areas. For example,
losses  by  sedimentation  would  progressive!/
decrease nutrient quantity  or  availability along the
length of the reservoir. This would  in turn foster the
establishment of similar longitudinal  gradients h
sediment quality and phytoplankton productivity.
  Differences in water density between river inflows
and lake surface waters may also affect the distribu-
tion of influent materials. In stratified reservoirs, rivef
waters enter and progress through shallow upstream
areas of the reservoir as a well-mixed flow. As the
reservoir basin widens and deepens,  and inertia
decreases, cooler river waters plunge to depths of
similar density resulting in the occurrence of  inte'-
flows or underflows. Such hydrodynamic occurrences
potentially   affect   nutrient  distributions  and the
establishment of longitudinal gradients, since down-
stream areas of the reservoir's epilimnion  would be
partially isolated from nutrient inputs during summer
stratified months.
   Such relations between nutrient and  chlorophyll
distributions,  sedimentation,  and  hydrodynamics
prompted Thornton et  al.  (1981) to propose an
heuristic  model  describing  the  establishment  of
longitudinal water  quality  gradients  in  reservoirs
which are strongly influenced by advective forces. The
model  identifies three zones — riverine,  transition,
and lacustrine—extending from headwater to dam.
The riverine zone is characterized as having a river-like
flow regime, high nutrient and suspended solid con-
centrations, and reduced  phytoplankton  standing
crop. The lacustrine, or lake-like, zone is less influenc-
ed by river inputs, has lower nutrient and suspended
sediment   concentrations,  and  a  moderate
phytoplankton standing  crop. The transition zone,
which is located near the point at which river waters
plunge  below lake  surface waters, exhibits in-
termediate characteristics and is potentially the most
productive portion of the lake. While recognizing that
the boundaries between zones are  difficult to define,
the  model provides a  framework within  which  to
discuss changing water quality characteristics along
the length of a reservoir.
   Presented here are results of studies conducted at
three reservoirs  having  differing  nutrient loading,
water quality,  morphometric, and hydrologic charac-
teristics. Similarities  and differences in  observed
water quality characteristics are discussed in relation
to factors affecting the establishment of water quality
gradients.

STUDY SITE DESCRIPTIONS
Comprehensive water quality studies were conducted
during the period 1976-1980 at DeGray, Red Rock, and
                                                266

-------
                                                                    COMPARATIVE ANALYSIS OF RESERVOIRS
West Point Lakes, three U.S. Army Corps of Engineer
reservoirs having different physical characteristics
(Table 1). DeGray Lake, a large, deep, tributary reser-
voir operated primarily for hydroelectric power produc-
tion, was created in 1970 by impoundment of the Cad-
do River in south-central Arkansas. Located in a pre-
dominately forested watershed with minimal urban or
residential development, the lake exhibits few water
quality problems. River flows are markedly seasonal
with highest flows occurring in spring (Montgomery
and Kennedy,  1984). During this period, river inflows,
which occur as interflows, are detectable in the upper
reaches  of the reservoir as plumes of turbid water
(Ford et al. 1980).
  Red  Rock Lake and West Point Lake,  both main-
stem impoundments,  are  located on relatively large
rivers and were constructed primarily for flood control
and hydropower, respectively. Red Rock Lake was im-
pounded in  1969 on the  Des Moines River in south-
eastern Iowa approximately 96 km downstream from
the city of Des Moines. Sediment and nutrient export
rates for this large,  predominately agricultural water-
shed are excessive and lake waters are extremely tur-
bid. In addition, sediment deposition has resulted in
the formation  of an expansive submerged delta in the
upstream third of the  reservoir (Gunkel et al. 1983).
  West  Point  Lake, the largest of the three study
sites, receives a majority of its water and material in-
puts from the Chattahoochee River.  Sediment and
nutrient loads transported by this river, which drains a
large urban and agricultural  watershed  in  central
Georgia, are also high. The reservoir has  a complex
morphology which features numerous  coves, shallow
near-shore  areas,  and   two  major   embayments.
Riverine  influences  are  apparent in  the  narrow,
shallow  upstream third of the  reservoir and the loca-
tion of the plunge line is often easily discernible as a
region of rapidly changing  turbidity (Kennedy et al.
1982).
 METHODS

 Two types of water quality studies have been con-
 ducted at each of the sites: long-term monitoring and
 short-term,  intensive studies. Long-term monitoring
 programs involved routine data collection at selected
 stations located along the  length of each  reservoir
 and were designed to delineate longitudinal differ-
 ences in water quality or to detect temporal changes.
 Short-term studies were designed to provide detailed
 information concerning specific processes or to better
 define spatial and temporal  patterns in water quality.
 While each of the  studies was  conducted on a dif-
 ferent time-frame,  and in  some cases by  different
 research groups, methods of sample collection and
 analysis were similar  and  generally conformed to
standard  analytical  methods  (such  as  Standard
Methods, 1980).
  For  present  purposes,  material  input  (and  dis-
charge) estimates have been based on mean monthly
flow and  concentration data  observed  at stations
located immediately upstream (and downstream) from
the reservoir. Gradients in surface water quality are
based on normalized station conditions by expressing
average concentrations at each station for each reser-
voir relative to that reservoir's maximum average con-
centration. Using this convention, normalized station
concentrations  range from a minimum of zero to a
maximum of one. Since flow rates are expected to in-
fluence spatial gradients,  two hydrologic seasons
were defined. Those months during which flow and
lake volume values would yield a theoretical hydraulic
residence time less  than  that computed on annual
averages were  considered to be  high flow months.
Months during  which hydraulic residence times ex-
ceeded the average annual value were considered to
be low flow  months. The use of hydraulic residence
times thus incorporates the effect of changes in both
flow and pool elevation (that is, storage).
  Sediment quality studies and more detailed studies
of water quality  patterns were  conducted at each
study site. Sediment studies involved collecting and
analyzing the top 10 cm portion of cores obtained at
several stations located throughout each reservoir
(Gunkel et al., 1983). Intensive water quality surveys
were conducted over 1-2 day periods during summer
and involved sample collection at several  depths at
30-60  stations  (depending on  reservoir size) located
along each reservoir's major axis.

RESULTS

Annual areal nutrient loading rates, nutrient retention
coefficients and  hydraulic residence times  differed
markedly between reservoirs (Table 2).  As expected,
nutrient loads for the mainstem reservoirs  located on
larger  rivers draining  urban and agricultural water-
sheds were higher than for DeGray Lake, the tributary
reservoir having a  predominately forested watershed.
While  nitrogen retention coefficients  were low and
similar for DeGray and Red Rock Lake (there were in-
sufficient data for this calculation for West Point
Lake),  phosphorus retention in West Point and DeGray
Lakes was higher than the small rapidly-flushed main-
stem reservoir.
   Differences among hydrologic seasons and season-
al differences between reservoirs are also apparent
(Table 2). Hydrologic seasons were similar for DeGray
Lake (low flow from July through October) and West
Point Lake (low flow from June through October). In
contrast, the low flow season for Red Rock Lake, the
only site experiencing snow-fall and freezing condi-
tions, extended from November to February.
Table 1.— Physical
Characteristic
Impoundment type
Major tributary
Volume (ID6™3)
Surface area (Km2)
Length (Km)
Mean depth (m)
Maximum depth (m)
Average hydraulic residence time (yr)
characteristics of DeGray,
DeGray Lake
tributary
Caddo River
808
54
32
14.9
60
2.0
Red Rock, and West Point
Red Rock Lake
mainstem
Des Moines River
78
26
12
3.1
11
0.05
lakes.
West Point Lake
mainstem
Chattahoochee River
746
105
53
7.1
31
0.13
                                                  267

-------
LAKE AND RESERVOIR MANAGEMENT
   Nutrient loading rates, which  are  expected to be
related to inflow rate, were higher during the high flow
season for each reservoir. Seasonal differences were
most noticeable in the case of phosphorus loading to
DeGray Lake. While the thirty-five-fold increase during
the high flow period may reflect more intensive, storm-
oriented sampling (Kennedy et al. 1983b; Montgomery
and Kennedy,  1984),  flows were  more  seasonally
variable in the Caddo River than in the river's influent
to the other lakes. Spring storm events frequently in-
creased the Caddo River's  flow by several hundred-
fold over relatively short periods (days), and attendant
increases in nutrient concentrations  (and thus load)
were often extreme (Montgomery and  Kennedy, 1984).
Storm-related changes in flow in the larger rivers wer«
less extreme and of longer duration. This characteris-
tic, and the fact that these rivers  receive  numerous
point-source loading from urban areas, may account
for the less pronounced difference between seasonal
loading rates.
   Gradients or patterns in four water quality variables
(total phosphorus, total  nitrogen, suspended solids,
and chlorophyll), were evaluated on a seasonal basis;
in surface waters of each reservoir. Only total nitrogen
exhibited  no trend of longitudinal change in all three
reservoirs. Gradients in DeGray Lake (Fig. 1), are con-
sistent with those observed during previous studies
(Thornton  et  al.  1982).   Total   phosphorus   and
chlorophyll  concentrations  were higher and more
variable at upstream stations (headwater  and mid-
lake) than at the near-dam station. The greater degree
of variability observed during the  high flow season
was presumably related to variations in flow and the
extent to which river inputs  affected downstream
areas. Gradients in suspended solids were not ap-
parent.
   Patterns in Red Rock Lake were variable and only
total  phosphorus  and  suspended solids  displayed
consistent decreases from headwater to dam (Fig. 2).
The effect of high sediment inputs and of changes in
influent suspended solid concentrations, particularly
following  storm  events (Kennedy  et al.  1981),  are
reflected in high and variable suspended solid concen-
trations at the headwater station.  Steeper  gradients
during  highflow  months  reflect  elevated  inflow
suspended solid concentrations. Patterns in  chloro-
          phyll  distribution  were apparently more related to
          riverine transport than to processes occurring within
          the reservoir. Soballe (1981) determined that, despite
          high nutrient  levels,  rapid flushing and  diminished
          light availability limit phytoplankton standing  crops
          and  that  observed  chlorophyll concentrations  are
          maintained by continuous input from upstream areas.
            Similarly to the other reservoirs, total phosphorus
          concentrations decreased from headwater to dam in
          West  Point Lake (Fig. 3). The observed variability in
          suspended solids concentration at mid-lake stations
          was potentially related to changes in phytoplankton
          standing   crop since  maximum concentrations  of
          chlorophyll occurred  at  mid-lake  stations  in  West
          Point Lake. This observation is related to the fact that
          riverine flows and inorganic turbidity limit phytoplank-
          ton standing crops at  upstream locations (Kennedy et
          al. 1982). As river flow rates decrease within the reser-
          voir and influent suspended loads are reduced by sedi-
          mentation, conditions for phytoplankton  growth  are
          greatly improved.
            § 2r
            cr
            O
            Q.

            O 1
            o o
            o
            o
            en
            K
            z


            o
DEGRAY LAKE
(3.23) (2.0
" 1 f


T
1 ,
9)
u> 2
O
T O
(A
8 1
" ii §
(i i .
HMD "' "
2
- '
} H P I'
0
j 	 1 	 \ — n
r



I
IT r
, ; |
HMD
! I- »
                  HMD        "HMD
                                 STATION
         Figure  1.—Total phosphorus, suspended  solids,  total
         nitrogen and chlorophyll concentrations for surface (V 5 m)
         waters at  headwater (H), mid-lake (M) and near-dam (D) sta-
         tions in DeGray  Lake for the period 1978-80. Symbols in-
         dicate normalized mean (± 1 SD.) values for high (•) and
         low (A) flow seasons.
         Table 2.—Comparative leading and retention characteristics for DeGray, West Point, and Red Rock lakes.
Variable

Phosphorus loading (gm P/m2 yr)
Nitrogen loading (gm N/m2 yr)
Hydraulic residence time (yr)
Phosphorus retention coefficient
Nitrogen retention coefficient

Phosphorus loading (gm P/m2- yr)
Nitrogen loading (gm N/m2- yr)
Hydraulic residence time (yr)
Phosphorus retention coefficient
Annual
— DeGray Lake —
0.28
3.51
2.06
0.51
0.03
— West Point Lake —
8.31
—
0.13
0.82
Hydrologic
High Flow

0.35
4.55
1.11
0.53
0.05

9.35
—
0.09
0.79
Season
Low Flow

0.01
1.43
3.39
0.48
0.01

6.86

0.18
0.86
 Nitrogen retention coefficient
 Phosphorus loading (gm P/m2- yr)
 Nitrogen loading (gm N/m2- yr)
 Hydraulic residence time (yr)
 Phosphorus retention coefficient
 Nitrogen retention coefficient
Red Flock Lake
     53.54
    705.77
      0.05
      0.25
      0.08
 63.20
941.14
  0.02
  0.27
  0.12
 34.21
235.02
  0.12
  0.25
-0.05
                                                  268

-------
                                                                     COMPARATIVE ANALYSIS OF RESERVOIRS
  While normalized concentrations in surface waters
provide a rough index for examining water quality gra-
dients or patterns in each of these reservoirs, informa-
tion  gathered during intensive  samplings  provided
greater insight to relations between water quality pat-
terns and morphologic  and hydrodynamic  features.
Presented in Figure 4 are vertical and longitudinal pat-
terns in the distribution of total phosphorus in each
reservoir on a single day during the summer stratified
period. The  effect of tributary inflows and  resultant
mixing patterns is apparent. Concentrations in Red
Rock, while decreasing  substantially along the  longi-
tudinal axis of the reservoir, were relatively uniform
from  the top to bottom suggesting  plug-flow condi-
tions and  the occurrence  of vertical  mixing. This
agrees well with data obtained by Kennedy et al. (1981)
during and following the passage of a storm hydro-
graph when inflowing river waters were traced for a
4-day period.
  Distributional patterns for total phosphorus in West
Point Lake indicated vertical mixing and riverine con-
ditions in upper reaches of the  reservoir and a well-
defined interflow at mid-lake. Observations of surface
conditions (for example, changing turbidity and debris

                   RED ROCK LAKE
         III
     2r
o
o
cc
I-
z
«1
<
K
o
1-
              ft
(2.32)
8 2
O
in
Q
u 1
Q
Z
UJ
a.
in
3
' .











i
f






li
° H M D
(3.7) (2.09)
2
_i
I
Q.
O H
o: 1
O
Z
o
n
' I












T
1
t
;







accumulation) and subsequent dye studies (Kennedy
et al. 1983a) indicate that this is  a  common  hydro-
dynamic condition. The  extent and  location  of the
anoxic zone were also influenced by the river inflow.
High flow velocities in the upstream portion  of the
reservoir prevented the establishment of anoxic condi-
tions in all but the deeper, downstream third  of the
reservoir.
  Conditions in DeGray Lake also indicated a river in-
fluence. While inflow rates were substantially lower in
this lake during summer months, morphologic effects
were more  apparent.  Unlike West Point Lake, the
plunge  point was located near the point of  inflow
under normal summer flow conditions and high phos-
phorus  concentrations  were limited to  upstream
areas. Interflows, because of the steeper thermal gra-
dient in this  reservoir, were confined  to a narrow
stratum immediately above the thermocline (Ford  et
al. 1980). The deposition of both allochthonous and
autochthonous organic material to the  shallow up-
stream  portion of the hypolimnion resulted in  anoxic
conditions here (James and Kennedy, 1984) and con-
sequent releases of phosphorus  from  bottom sedi-
ments (Kennedy et al. 1983b). As a  result, phosphorus
concentrations were considerably higher than those
of the inflowing river water.
   Patterns of longitudinal gradients in nutrient distri-
butions in the water column and the effects of riverine
inflows were  reflected in  sediment quality (Fig.  5).
Sediment phosphorus concentrations  were signifi-
cantly  (P^.05) lower downstream from  the point  at
which  riverine influences diminish. Similar to water
column conditions, no longitudinal differences were
found for sediment nitrogen. Particularly noteworthy
in light of the foregoing discussion of chlorophyll gra-
dients was the observed peak in sediment organic car-
bon  concentration at a mid-reservoir location in West
Point Lake. Patterns  in the distribution  of sediment
organic carbon in DeGray Lake, however, appear to  be
more related  to preimpoundment  conditions (timber
and  terrestrial detritus inundated  during lake filling)
than to lacustrine processes (Gunkel et  al. 1983).
                         STATION

 Figure 2.—Total phosphorus, suspended  solids,  total
 nitrogen and chlorophyll concentrations for surface waters
 at headwater (H), mid-lake (M) and near-dam (D) stations in
 Red Rock Lake (based on data reported by Baumann et al.
 1981). Symbols indicate normalized mean (±  1 SD.) values
 for high (•) and low (A) flow seasons.
                   WEST POINT LAKE
       II
            f   I'
             M
           (235)
 Figure 3.—Total phosphorus, suspended solids and chloro-
 phyll concentrations for surface waters at headwater (H),
 near-dam (D) and two mid-lake (M) stations in West Point
 Lake (based on data compiled by Walker, 1981). Symbols in-
 dicate normalized mean (± 1 SD.) values for high (•) and
 low (A) flow seasons.
  TOTAL PHOSPHORUS CONCENTRATION,
           (Shaded areas are zones of anoxia)

                             DEGRAY
                              7/26/78
                                                        RED ROCK
                                                          8/9/79
                                                                         WEST POINT
                                                                           6/6/79
 Figure 4.—Vertical  and longitudinal  distribution of total
 phosphorus (JJQ  P/l) in  Red  Rock, DeGray and West Point
 Lakes. Shaded areas indicate zones of anoxia.
                                                  269

-------
  LAKE AND RESERVOIR MANAGEMENT
  DISCUSSION

  Water quality gradients, common features of reser-
  voirs and other large river-fed lakes (Gloss et al. 1980;
  Johnson and Merritt, 1979; Peters, 1979; Thornton et
  al. 1982; Walker, 1983; Kennedy et al. 1982), result fron
  the combined effects of morphology, hydrology  and
  hydrodynamics, and  sedimentation. While limited to
  three lakes, observations of gradients presented hero
  conform to those suggested by the heuristic model of
  Thornton et al. (1981). Material inputs to these reser-
  voirs are transported variable distances within  the
  reservoir. The extent to  which this occurs is directly
  dependent upon the development of volume (longi-
  tudinal changes in basin width and depth) and rate of
  flow, and  indirectly  upon  sedimentation.  In   this
  regard,  DeGray and  Red Rock  Lake  represent ex-
  tremes.  Red  Rock  Lake, since it  is  smaller  and
  receives higher flows is strongly influenced by its
  tributary.  Despite   high  rates   of  sedimentation,
  nutrient  concentrations  and turbidity  remain  high
  along the entire length  of the  reservoir and auto-
  chthonous production  is  lower  than would  be  ex-
  pected.  The reservoir appears to lack  a lacustrine
  zone. DeGray  Lake, on the  other  hand,  exhibits
  riverine, transition, and lacustrine zones. However,
  lower inflow rates and greater volume  development
  result in relatively small  riverine and transition zones
  which are frequently  located  proximate to the tribu-
  tary inflow. Greatest productivity occurs in upstream
  areas of the reservoirs and the potential for internal
  phosphorus loading from shallow hypolimnetic areas;
  is high (Kennedy et al. 1983b).
    West  Point  Lake,  which also  exhibits  three lake
  zones, represents an  intermediate situation. River in-
  flows penetrate to mid-reservoir resulting in the estab
  lishment of large riverine and transition  zones. Auto
  chthonous production reaches a maximum at mid
  reservoir, coincident with the location of the transition
  zone.
         WEST POINT LAKE        DEGRAY LAKE
            10    20   30   40   0    10    20   30   40
      20


      1 5


      10


      05


       0
       0


      25

      20 .


      1 5 -

      1 0 -

      05 -

       „_-.
10    20    30   40  0    10    20    30   40
           10   20    30    40   0    10   20    30    40

                    DISTANCE FROM DAM, km
Figure 5.—Longitudinal changes  in sediment phosphorus,
nitrogen and organic carbon concentrations for DeGray and
West Point Lakes (based on data reported by Gunkel et al
1983)
                                              CONCLUSIONS

                                              The  observance  of  gradients and  an evaluation of
                                              various processes leading  to their establishment in
                                              reservoirs and  other river-fed lakes provides a basis
                                              for an improved  understanding of  water quality dy-
                                              namics in these aquatic systems. The assumption of
                                              complete mixing, commonly employed for evaluating
                                              trophic  state responses to  loading, may be inappro-
                                              priate for  many reservoirs.  The efficacy of applying
                                              trophic  state indices based on "average" conditions
                                              should also be questioned for lakes exhibiting marked
                                              longitudinal gradients. Also, stratified sample designs
                                              for monitoring programs may often be required to ade-
                                              quately  define water quality conditions.
                                             ACKNOWLEDGEMENTS: This research funded by the Envi-
                                             ronmental and Water Quality Operational studies sponsored
                                             by the Office of the Chief, U.S. Army Engineer.
 REFERENCES

 Bauman, R.E., C.A. Beckert, D.L Schulze, and D.M. Soballe
   1981. Water Quality Studies—EWQOS Sampling: Red Rock
   and Saylorville Reservoirs, Des Moines River, Iowa. Annu
   Rep. Eng. Res. Inst., Iowa State Univ., Ames.

 Baxter, R.M. 1977. Environmental effects of dams and impound-
   ments. Ann. Rev. Ecol. System. 8:255-83.

 Ford, D.E., M.C. Johnson, and S.G. Monismith. 1980. Density in-
   flows to DeGray Lake, Ark. Second Int. Symp. on Stratified
   Flows. Trondhiem, Norway.

 Gloss, S.P., L.M. Mayer, and D.E. Kidd. 1980. Advective control
   of nutrient dynamics in the epilimnion of a large reservoir
   Limnol. Oceanogr. 25:219-28.

 Gunkel,  R.C.,  et al. 1983. A comparative study  of  sediment
   quality in four reservoirs. Tech. rep. (in press). U.S. Army Eng
   Waterways Exp. Sta., Vicksburg, Miss.

 James, W.F., and R.H. Kennedy. 1984. Patterns of sedimenta-
   tion in DeGray Lake, Ark. Tech. rep. (in prep.) U.S. Army Eng
   Waterways Exp. Sta., Vicksburg, Miss.

 Johnson, N.M., and D.H. Merritt. 1979. Convective and advective
   circulation  of Lake  Powell, Utah-Arizona,  during 1972-75
   Water Resour. Res. 15:873-84.

 Kennedy, R.H., K.W. Thornton, and J.H. Carroll. 1981. Sus-
   pended-sediment gradients in Lake Red Rock. Proc. Symp.
   Surface-Water Impoundments. Am. Soc. Civil Eng. Minne-
   apolis, Minn. 11:1318-28.

 Kennedy, R.H., K.W. Thornton, and R.C. Gunkel. 1982. The
   establishment of water quality  gradients in  reservoirs
   Can. Water Res. J. 7:71-87.

 Kennedy, R.H., R.C. Gunkel, and J.V. Carlile. 1983a. Riverine
   influences on the  water quality  characteristics  of West
   Point  Lake. Tech. rep. (in press). U.S. Army Eng. Water-
   ways Exp. Sta., Vicksburg, Miss.

 Kennedy, R.H., R.H. Montgomery, W.F. James, and J. Nix.
   1983b.  Phosphorus, dynamics in  an Arkansas Reservoir:
  the importance of seasonal loading and internal recycling.
   Misc.  Pap. E-81-1. U.S. Army Eng. Waterways Exp. Sta.,
  Vicksburg,  Miss.

Montgomery, R.H., and R.H. Kennedy. 1984. Material loading
  to DeGray  Lake, Ark. by the Caddo River. Tech. rep. (in
  prep).  U.S. Army  Eng. Waterways Exp. Sta., Vicksburg,
  Miss.

Peters, R.H.  1979. Concentrations  and  kinetics of phos-
  phorus fractions along the trophic gradient of Lake Mem-
  phremagog. J. Fish. Res. Board Can. 36:970-9.

Ryder, R.A. 1978. Ecological heterogeneity between north-
  temperate reservoirs and glacial  lake systems due to dif-
  fering  succession  rates  and cultural  uses.  Verh  Int
  Verein. Limnol. 20:1568-74.
                                                    270

-------
                                                                           COMPARATIVE ANALYSIS OF RESERVOIRS
Soballe, D.M. 1981. The fate of river phytoplankton in Red
  Rock Reservoir. Ph.D.  Dissertation.  Iowa State  Univ.,
  Ames.
Standard Methods for the Examination of Water and Waste-
  water.  1980.  15th  ed.  Am.  Pub. Health  Assoc., Ass.
  Washington, D.C.
Thornton, K.W., R.H. Kennedy, A.D. Magoun, and G.E. Saul.
  1982. Reservior water quality sampling design. Water Res.
  Bull. 18:471-80.
Thornton, K.W. et al. 1981. Reservoir sedimentation and
  water quality—a heuristic model. Proc.  Symp.  Surface-
  Water  Impoundments. Am. Soc. Civil Eng., Minneapolis,
  Minn. 1:654-61.
Walker, W.W. 1981. Empirical methods for predicting eutro-
  phication in impoundments. Phase I: Data Base Develop-
  ment. Tech. rep. E-81-9. U.S. Army Eng. Waterways Exp.
  Sta., Vicksburg, Miss.
	1983. Empirical methods for predicting eutrophica-
  tion in  iimpoundments. Phase II: Model Testing. Tech. rep.
  E-81-9. U.S. Army Eng. Waterways Exp. Sta., Vicksburg,
  Miss.
                                                       271

-------
  INTERMOUNTAIN WEST RESERVOIR LIMNOLOGY
  AND MANAGEMENT OPTIONS
 JERRY MILLER
 U.S. Bureau of Reclamation
 Salt Lake City, Utah
 INTRODUCTION

 The Upper Colorado Region of the U.S. Bureau of Re-
 clamation includes the Colorado River from its head-
 waters to Lees Ferry below Glen Canyon Dam, and the
 Bonneville Basin which drains into the enclosed Great
 Salt Lake in Utah (Fig. 1).
   The entire flow of the Colorado  River has  been
 stored and regulated for most of the  last 50 years. In
 the runoff of 1983, the entire Colorado River Reservoir
 complex was filled beyond capacity,  and for the first
 time since 1935 major flooding occurred below Hoover
 Dam. During this extended period  of initial  reservoir
 filling, a number of physical, chemical, and biological
 changes occurred.  Only  since about 1975  have wo
 been able to  observe limnological  conditions ap-
 proaching dynamic equilibrium  in  the river/reservoir
 complex as a whole (U.S. Dep. Inter. 1983).
   Reclamation has  focused  on understanding the
 long-term  impacts  of  this developing  reservoir
 dynamic equilibrium on several key management con-
 cerns: (1) impacts to basinwide salinity; (2) eutrophica-
 tion,  particularly in  the  reservoir inflow areas; (3)
 downstream temperature; (4) temperature and nutrient
 interactions of upstream  reservoir releases on down-
 stream reservoirs; (5) in-reservoir management alterna-
 tives  such as  selective withdrawal,  chemical treat-
ment, and aeration; and finally, (6) more conventional
watershed and nutrient control management options.
  Reclamation has a number of ongoing studies to
define the  future dynamic reservoir  limnological
equilibrium in the Colorado River Basin (Miller et al.
1983). I will describe in broad, general terms two reser-
voirs—Flaming Gorge in the Colorado River Basin and
Deer Creek Reservoir  in the Bonneville Basin. Finally, I
will use the limnological history of these two reser-
voirs to discuss some of  the critical  water quality
problems and management alternatives for reservoir/
river/reservoir interactions. This paper will rely on
some of the  reservoir  limnology hydrodynamics
discussed in  other papers in this chapter.
CLIMATE/METEOROLOGY/LIMNOLOGY

The Intermountain West is an area of broad contrast.
Mountain ranges from 3,000 to 4,250  meters  ac-
cumulate heavy winter snowpacks that provide much
of the flow of the major rivers traversing through semi-
arid valleys. The climate,  hydrology,  morphometry,
and limnology of the reservoirs are summarized in
Table 1.
   °K£GON'      '   °  A  H  o

                                                            O  M I  N  G
      4
           NEVADA
                                                                                       \	
Figure 1.—Upper Colorado Region.
                                             272

-------
                                                                   COMPARATIVE ANALYSIS OF RESERVOIRS
                 Table 1.—Flaming Gorge and Deer Creek Reservoir morphometry and limnology.
Reservoir
Flaming Gorge
Volume (m?)
4672.8 X 106
MX.
Depth (m)
135
Mean
Depth (m)
25.5
Hydraulic
Retention Time
2.4 yrs.
Length
km
150
Outlet
Works
pre 1978
hypolimnion
post 1978
selective
Elevation
(m)
1840
Deer Creek
                187.5 x 106
                     42
                                           17
                                                .57 yrs.
                                                                       9.5    hypolimnion
                                                                                             1650
  Spring runoff in May and June produces the bulk of
reservoir inflow and nutrients. The warm temperatures
that induce snowmelt runoff may also cause thermal
stratification. The setup of thermal stratification and
hypolimnion  flushing   rate for bottom withdrawal
reservoirs during the runoff period are very important
to limnological conditions the rest of the year.


FLAMING  GORGE  RESERVOIR

Flaming Gorge is a large reservoir (150 km long, 120 m
maximum  depth,  and  a 2.4 year hydraulic retention
time)  and  exhibits downreservoir  trophic segmenta-
tion (Fig. 2).
   During much of the year Flaming Gorge is tempera-
ture limited. The productivity fluctuates widely, but by
fall blue-green algae become dominant in the inflow
area (Fig. 2).
   Thermal stratification and summer stagnation, par-
ticularly in the inflow area, can vary from early May to
                                         late September in a hot/dry year, or from mid-July to
                                         late August in a cool/wet/windy year. Summer stagna-
                                         tion and dissolved oxygen (DO) sag below the thermo-
                                         cline can range from 45 days to over 110 days depen-
                                         ding primarily on the meteorological conditions of the
                                         year.
                                            During most years the  inflow area develops anoxia
                                         and accumulates high  available phosphorus concen-
                                         trations below the thermocline. The development of
                                         blue-green algal blooms seems to follow fall turnover
                                         which begins in September. By late October, turnover
                                         induced reaeration has usually reestablished satura-
                                         tion DO levels, and the quickly declining temperatures
                                         terminate noxious blue-green blooms in November.
                                            Storage began in 1963; by 1967 the U.S. Geological
                                         Survey studies (Bolke and Waddel, 1975) showed that
                                         a chemocline had developed in the hypolimnion just
                                         below the penstock  withdrawal elevation. This moni-
                                         molimnion remained anoxic and  did not completely
                                         mix from 1967 until the fall  and  spring  turnover in
       0 Km
         I
                              40 Km
                                                     80 Km
                        -LACUSTRINE ZONE-
   M-elev
   M-elev
                                             I
 lt-6000 -
 M- 1825
-t9«l
meters

Spillway elev. June '75-'83
 	e.ev.6020        OL.GOTROPH.C

  7-75/EC-590
 [9-68/EC-600]"

 (5-83/EC-650)
 (7-83/EC-600)
 M- 1800
 tt-5900 -
 M- 1 784
 ft-5800 -
 M- 1 750
 tt-5700 -
 M-1725 -
 ft-5600 -
 M-1700
                                              	1	(Dec.-ApriO-
                                                          (Aug.-Oct.)
                                                  I	TRANSITION ZONE	1
                                                                I	RIVERINE ZONE	
                                                                       (May-July)
                                               MESOTROPHIC  BLUE-GREENS-Sept.-Oct.
                                              	~~              EUTROPHIC
                           THERMOCLINE
                                                                 low D.O,
                                          high  O.O.
[9-68/EC-675]
    I/EC_810]	 -^gftEMpCLINE  ^*f?S&
                                                                      Limnological  History:
 9-73/EC-880
pre  1975-anoxic/
 9-75/EC-830    f'
[9-68/EC-10001
1981-aerobicx^/
15 a 3 /
   EC-810),
I7-83/
  E C - 7 0 0),
 9-7S/
  EC-89C
1983-  ...
aerobic*
                        A.  Penstock Outlet  elevation  1965  no  1978.  B. River Outlet—flood
                        control  releases were made  in  1975 and 1983.  C. Selective withdrawal
                        structure  (1978) for  downstream temperature control.  D. Chemocline—
                        sharp  increase  in  TDS prevented  turnover  until  1981-82.  Spills from
                        the  river outlet  in  1983  resulted  in  release of most of  the deep
                        high  TDS  water;  chemical  stratification  is  no longer significant.
                        E.  The  text refers  to  the  inflow area which  includes the riverine
                        and  transition  zones where  fall  blue-green algae dominance and low
                        dissolved  oxygen conditions  occur.


                        75/EC-590   refers   to  the  July  1975  electrical   conductivity  at
                        approximately the depth shown  (EC  X  .7 = TDS).
 Figure 2.—Flaming Gorge Reservoir profile.
                                                 273

-------
 LAKE AND RESERVOIR MANAGEMENT
 1981-82. Since the early 1970's the total  dissolved
 solids (IDs) began gradually declining in  the moni-
 molimnion (Miller et al. 1983). By 1983 the reservoir
 had completely mixed  and only a trace of chemical
 stratification remains (Fig. 2). Dissolved oxygen  was
 observed  near saturation  levels through the entire
 depth by May 1983 (U.S. Bur. Reel. 1978-83).
   During initial filling, completed in 1973, salts were
 leached from newly inundated geologic formation!;.
 This initial leaching (documented by Bolke and Wad-
 dell, 1975) was probably largely  responsible for the
 development of the chemocline. The fact that reser-
 voir sediments eventually  covered  most of the salt-
 bearing geology, plus the lack of mechanical action
 below the inactive storage zone greatly decreased salt
 leaching, particularly in the monimolimnion. There are
 no readily apparent  mechanisms to reestablish the
 chemocline.
   Prior to 1978 hypolimnion releases  from Flaming
 Gorge  Reservoir  caused cold downstream tempera-
 tures to vary from 4° C to 10° C annually. These temp-
 eratures severely limited potential fishing and recrea-
 tion. In 1978 the Bureau of Reclamation added a selec-
 tive withdrawal structure to the penstock to facilitate
 warmer eplimnion releases (Sartoris, 1976).  Tempera-
 tures  from May  through November now  generally
 range from 12 to  16° C to obtain maximum  trout prc-
 ductivity. An outstanding coldwater fishery  is now
 available in about 30 miles of the Green River belov/
 the dam (Gosse, 1982).
   Following operation  of  the selective withdrawal
 structure, several interesting developments occurred
 in the reservoir. The chemocline decay seemed to ac-
 celerate and the inflow area eutrophication appeared
 worse. In fact, in 1978-9  such strong dissolved oxygen
 sags occurred in  the inflow area during the fall, that
 the trout completely disappeared from that portion of
 the reservoir (Dufek and Wengert, 1979-80).
   The State of Wyoming became alarmed and asked
 Reclamation to determine if the selective withdrawal
 120  to 150  km downreservoir was related  to  the
 depressed dissolved oxygen levels and loss of trout
 habitat in the inflow area. Reclamation did investigate
 this possibility.
   Limnological data was limited in 1978-9, but it ap-
 pears that the reservoir stratified fairly early, and lonci
 summer stagnation occurred; dense blue-green alga!
 blooms were also observed  although not quantified
 The  lengthy summer stagnation  periods were  ap
 parently induced  by meteorological  and hydrologic
 conditions. Similar summer stagnation  periods were
 observed in several other reservoirs in the region (U S
 Bur.  Reel. 1978-83).
  Since 1980 the summer stagnation has tended to be
 much shorter, again  primarily a function of  meteoro-
 logical conditions. Based on the data presently avail-
 able, no correlation appears between the operation of
 the selective withdrawal  structure and the extreme de-
 pression of DO and subsequent temporary loss of trout
 habitat  in the inflow area in 1978-9. The trout have
 returned to this area (even through the fall) since 1980.
 Again, dynamic chemical and  biological equilibrium
 has not yet occurred; major shifts,  such as the de-
 struction of the chemocline, are still  being observed.
  A number of factors contribute to the  development
of  noxious  biue-green  algae  (Aphanizomenon and
Anabaena) in the tributary  arms of  Flaming Gorge.
 Natural  phytoplankton pulses, the mixing of phos-
phorus released  from the sediments in the anoxic
hypolimnion  during  fall  turnover, erosion and
 agricultural  nonpoint  sources,  municipal  point
 sources, high phosphorus releases from an upstream
 reservoir on the Green River during summer stagna-
 tion, and the hydrodynamics of the inflow area may all
 contribute.
    Most of Flaming Gorge  is either mesotrophic or
 oligotrophic, particularly as water depths begin to ex-
 ceed 20 m. Nutrient control into the inflow area may
 be difficult or impractical. The problem is the impact
 of the depressed  DO on the trout fishery  and the
 nuisance  of blue-green  algae to recreation. In  this
 case, reducing or eliminating  the dominance of blue-
 green algae is  much  more desirable than  reducing
 total productivity or trophic status of the entire reser-
 voir. Even better would be management alternatives
 designed to provide recreation facilities 15 to 20 km
 farther downreservoir, use other  fish species in the in-
 flow area, or eliminate the blue-green algae. Unfortun-
 ately, an effective method to eliminate the blue-green
 algae has not been demonstrated, and providing new
 roads and recreational facilities is a very expensive
 proposition in this rugged region. The length of the
 eutrophic reservoir in the inflow area should be con-
 sidered before locating recreation facilities.
 DEER CREEK RESERVOIR

 Deer Creek is a major municipal and irrigation water
 supply reservoir for the metropolitan Salt Lake City/
 Provo, Utah area. The morphometry and limnology are
 shown in Table 1 and Figure 3. It has a deep hypo-
 limnion outlet and mean hydraulic  retention time of
 210 days. Because of the internal hydrodynamics, an
 excellent downstream trout  fishery temperature is
 maintained without a selective withdrawal structure.
   Deer Creek has shown a strong tendency to stratify
 and turn anaerobic during the summer  and  occa-
 sionally under the ice during a hard  winter. Aphanizo-
 menon (blue-green algae) blooms occur from August
 into October, and combined with  the  anoxic  hypo-
 limnion cause taste and odor problems for the potable
 water supply. For about 20 years the Salt Lake Metro-
 politan Water District has treated the reservoir with
 copper sulfate  to  reduce  the  blue-greens. Weekly
 temperature profiles and algae counts were taken dur-
 ing the summer  to determine  when to treat (Hershey,
 pers. comm.). Most of these data have not  been
 published.
   In 1982  a new nondischarging  land  application
 wastewater treatment plant began operation in Heber
 City. This reduced phosphorus by about  15 percent,
 perhaps even more significantly to the inflow area dur-
 ing the fall bloom period.
   Daily upcanyon winds in  the summer tend to keep
 the Aphanizomenon stacked into   the inflow  area;
 periodic downreservoir migration occurs when a fron-
 tal passage changes  the wind direction.  This algae
 movement  pattern masked  what were thought to be
 algal reductions following copper sulfate treatment.
 Copper sulfate  treatment  was terminated in  1981
 because of the reduced phosphorus  inflow and ques-
 tionable effectiveness. Additional point and nonpoint
 source  nutrient  reductions have also  been  imple-
 mented or are planned.
   Unfortunately, the past 2 water years have been very
 wet, cool, and windy. Deer Creek has not stratified un-
til mid-July, and summer stagnation has not exceeded
30 to 45 days. Most years, summer stagnation would
exceed 90 days.  The flushing  rate and bottom outlet
                                                274

-------
                                                                    COMPARATIVE ANALYSIS OF RESERVOIRS
are important  factors producing an  early turnover,
usually beginning in late August.
  Laboratory anaerobic and NaOH-P sediment extrac-
tions to determine available  phosphorus were con-
ducted in  Deer Creek Reservoir (Messer and  Ihnatt,
1983). We  hope to determine the importance of inter-
nal versus  external  phosphorus  loading  to the
Aphanizomenon blooms. Several more years of more
typical conditions will be required to determine the im-
pact of the external phosphorus reductions.
  Simulations of  the proposed  upstream Jordanelle
Reservoir with Water Quality for River Reservoir Sys-
tems (WQRRS) (Smith, 1978; Wegner,  1983) are being
used to help determine management alternatives to
decrease nutrient availability to Deer Creek. Timing
and quantity of phosphorus and temperature releases
from the proposed upstream reservoir are primary con-
siderations because of downstream blue-green algae
problems. A multiagency water quality program is now
being developed to protect Deer Creek and the propos-
ed Jordanelle Reservoir. The Deer Creek/Jordanelle
management policy will be based on prioritizing water
user/water quality requirements.


SUMMARY

Flaming Gorge and Deer Creek  present a number of
management problems.  A reduction of primary pro-
ductivity or trophic status is important because of the
municipal  water use  from Deer  Creek Reservoir, but
eliminating the anoxic  hypolimnia and  blue-green
algae is the key to success. On the other hand, most
of Flaming Gorge is already oNgotrophic or meso-
trophic; only the inflow area has a seasonal problem.
Reducing  nutrient inflow and  possibly total reservoir
productivity  may  not be  desirable,  particularly  for
                      fishery management. The problem is not so much the
                      eutrophic classification of the inflow area, but rather
                      the seasonal recreation nuisance of blue-green algae
                      and the impact on the coldwater fishery. What are the
                      realistic management options?
                        A  eutrophic classification is  not  necessarily  bad,
                      depending on the type of water use, and  it does not
                      necessarily  mean   blue-green  algae  dominance.
                      Reducing or eliminating blue-green algae should be a
                      higher priority than reducing the overall trophic status
                      of Flaming Gorge Reservoir's inflow area.
                        Selective withdrawal from Flaming Gorge has great-
                      ly improved the coldwater fishery in the river below the
                      dam and may have helped destroy the chemocline in
                      the reservoir. We have not been able to determine that
                      selective withdrawal  had any negative impacts on the
                      reservoir to date. Selective withdrawal is certainly not
                      necessary in every reservoir, and before deciding to
                      use  it potential negative impacts  to  the  reservoir
                      should be determined.  Selective withdrawal could
                      have negative impacts on a reservoir, depending on its
                      unique hydrodynamic characteristics.
                        Selective withdrawal, upstream-downstream reser-
                      voir  interactions,  reservoir  limnology, aeration,
                      chemical treatment, alternatives to control blue-green
                      algae, and downstream temperature and water quality
                      must all  be given careful  consideration. Appropriate
                      modeling techniques need to be used with judgment,
                      and  water-user requirements for water quality must be
                      prioritized and  balanced to develop the  best multi-
                      purpose water resource development alternatives.
                        Each reservoir has unique limnological and hydro-
                      dynamic  characteristics  and  should  be  carefully
                      analyzed  to determine  the  effectiveness  of  each
                      management alternative.  Managers should  be ex-
                      tremely careful not to make costly decisions based on
                                          Summer Wind
                      Masotrophic
                                                                                b|U.-ar.,n rnlga*
CO
CE
2
o
LU
_l
LU
                         (Auqu«t-Thermoclln«)
      1610
                                               The reservoir  is  filled
                                            during  May-June   runoff;
                                         thermal   stratification sets
                                      up  between  May  and  July  depend-
                                    ing    on    local    meteorological
                                 conditions. By August  most of  the cool
                              hypolimnion  has  been discharged;  surface
                           elevation  and  the  thermocline have  been
                         drawn down. During  years of   early  stratifi-
                       cation   the   cool  snowmelt   inflow  may  form  a
                    density current  which passes  thru the  reservoir
                 resulting in  a turbid outflow. In the  fall, blue-green
              algae  become dominant  in  the   inflow  area.   Dissolved
          oxygen  declines  rapidly  below the euphotic   zone  indicating
       the algae  are  circulated deeper than the light extinction  depth.
    Daily  upcanyon  summer  winds  confine the  blue-green  algae  to the
inflow  area except during temporary  frontal disturbances.
Figure 3.—Deer Creek Reservoir profile.
                                                 275

-------
 LAKE AND RESERVOIR MANAGEMENT
 1 or 2 years of limited reservoir data with models, par-
 ticularly lake models, that are inappropriate for  the
 system. Both empirical  and computer hydrodynamic
 modeling methods can be extremely useful when they
 are used with judgment and experience. Comparisons
 with   similar  reservoirs  should  be  used  whenever
 possible.
REFERENCES

Bolke,  E.L, and J.M. Waddell. 1975. Chemical quality and
  temperature of water in Flaming Gorge Reservoir, Wyom-
  ing and Utah, and the effect of the reservoir on the Green
  River. U.S. Geol. Surv. Water Supply Pap. 2039-A.
Dufek, D.J., and W. Wengert. Pers. comm. 1979-80. Wyoming
  Game and Fish, Green  River.

Gosse, J.C. 1982.  Microhabitat of Rainbow  and Cutthroat
  Trout in the Green River Below Flaming Gorge Dam. Vol. I,
  II. Utah Div. Wildl. Resour. Contract No. 0-07-40-S1357  U s'
  Bur. Reel. (WPRS), Salt  Lake City, Utah.

Hershey,  S. 1980.  Pers. comm.  Unpubl. Temperature  pro-
  files  and algae counts, 1963-80. Salt Lake Metro. Water
  Distr.

Messer, J.J., and J.M. Ihnat. 1983. Reconnaissance of sedi-
  ment-phosphorus relationships in some  Utah Reservoirs
  UWRL-Q-83-03. Utah State Univ., Logan.
 Messer, J.J, J.M. Ihnat, and D.L. Wegner, 1984. Phosphorus
   release from the sediments of Flaming Gorge Reservoir
   Wyoming, USA. Ver. Int. Verein. Limnol. 22. In press.
 Miller, J.B., D.L Wegner. and D.R. Bruemmer. 1983 Salinity
   and phosphorus routing through the Colorado River/Reser-
   voir System. Pages 19-41  in  Proc. 1981  Symp. Aquatic
   Resources Management of the Colorado River Ecosystem
   Ann Arbor Sci. Publ., Ann Arbor, Mich.

 Sartoris, J.J. 1976. A mathematical model for predicting river
   temperatures—application to the Green River below Flam-
   ing Gorge Dam. Appl. Sci. Br. Div. Gen.  Res. Enq  Res
   Center, U.S.  Bur. Reel., Denver, Colo.

 Smith, D.J. 1978. Water quality for river/reservoir systems
   U.S. Army Corps Eng. Hydraul. Eng. Center, Davis, Calif.
 U.S.  Bureau of Reclamation. 1978-83. Unpubl. limnological
   survey water quality records, Flaming Gorge, Strawberry
   and Deer Creek Reservoirs and Lake Powell. Upper Colo'
   Reg. Off. Salt Lake City, Utah.

 U.S. Department of Interior. 1983. Quality of water, Colorado
   River Basin. Progr. Rep. No. 11. U.S. Bur. Reel. Salt Lake Ci-
  ty, Utah.

Wegner, D.L. 1983. Unpubl. WQRRS computer simulations of
  Deer Creek and proposed Jordanelle Reservoir, Upper Col-
  orado Region. U.S. Bur. Reel. Salt Lake City Utah
                                                    276

-------
FACTORS CONTROLLING PRIMARY PRODUCTION  IN  LAKES
AND RESERVOIRS: A PERSPECTIVE
BRUCE L KIMMEL
Environmental Sciences Division
Oak Ridge National Laboratory
Oak Ridge, Tennessee

ALAN W. GROEGER
Department of Zoology
University of Oklahoma
Norman, Oklahoma


            ABSTRACT

            Phytoplankton productivity and biomass fluctuations are controlled by the same energy and
            nutrient inputs and the same balance of gain and loss factors in natural and manmade lakes.
            However, some significant physical and hydrodynamic differences between lakes and reservoirs
            do exist which influence (1) the relative contributions of various primary producers to their food-
            webs, (2) the relative importance of certain limiting factors to primary production (e.g., turbidity,
            nutrient availability, flushing rate), (3) the spatial variability of primary production within reser-
            voirs, and (4) the applicability of lake-based classifications and empirical relationships to reser-
            voirs. An important distinction between most natural and manmade lakes is that reservoirs are
            semi-fluvial environments that fall between rivers and lakes on a continuum of aquatic eco-
            systems. A wider recognition of the riverine influences on reservoir ecosystems will enhance our
            understanding of the spatial and  temporal heterogeneity associated with manmade impound-
            ments and, thereby, permit more effective management of reservoir resources.
 INTRODUCTION

 Attempts to  manage  reservoir  water quality  and
 biological productivity are often based on preconcep-
 tions derived  largely from our knowledge of natural
 lakes.  Considerable  overlap  occurs  and  many
 similarities  exist  between  natural  and  manmade
 lakes; however, some obvious differences (e.g., basin
 morphology, hydrodynamics and hydraulic retention
 times, extent of water level fluctuations, surface ver-
 sus subsurface outflow)  separate the "typical"  lake
 and reservoir. In fact, reservoirs  combine numerous
 features of  both riverine  and lacustine environments
 and can be viewed as occupying an intermediate posi-
 tion on a river-lake continuum (Fig. 1).
   This paper compares the primary  production and
 the environmental  factors controlling  primary produc-
 tion in natural and manmade lakes. Because autoch-
 thonous primary production usually provides a major
 portion of the organic matter base for lacustrine food-
 webs, it is necessary to  understand the factors that
 control it to effectively manage the biological produc-
 tivity of lakes and reservoirs. It is also important to
 distinguish  any differences in factors controlling lake
 and reservoir productivity and the management impli-
 cations of those differences.
 PRIMARY PRODUCTION IN LAKES AND
 RESERVOIRS

 Primary producers in natural and manmade lakes fall
 into  the same  major categories: planktonic  algae
 (phytoplankton),  periphytic  algae (periphyton), and
 macrophytes  (larger  rooted  aquatic  plants).  Other
 primary producers  (e.g.,  photosynthetic and chemo-
 synthetic  bacteria, algae in symbioses with  other
organisms) may be present, but are usually minor con-
tributors to the system's total primary production. The
relative importance of the phytoplankton, periphyton,
and macrophyte contributions to the total primary pro-
duction depends  on basin morphology, water clarity,
substrate suitability, and the extent of water level fluc-
tuations.  In many reservoirs, abiogenic turbidity (due
to suspended silts and clays) and large water level
fluctuations (resulting  from flood control and hydro-
power operations)  often  restrict the establishment
and development of attached algal and rooted macro-
phyte communities, and thereby, enhance the impor-
tance of  phytoplankton production  (e.g.,  Ellis, 1936;
Ryder, 1978).
  As a group, reservoirs appear to be somewhat more
productive  than  natural  lakes  in  terms  of  phyto-
plankton  productivity. We compared the annual phyto-
plankton  productivities of a number of natural lakes
and  reservoirs  for which appropriate  data were
available. Reservoir productivity values fell within the
same range as those for temperate and tropical lakes;
however, oligotrophic reservoirs were much rarer than
oligotrophic lakes  and  the  relative  frequency  of
eutrophic impoundments  was  considerably  higher
than that of eutrophic lakes (Fig. 2). This likely results
from differences in watershed size and fertility bet-
ween lakes and reservoirs.  Most  reservoirs have
significantly higher drainage-area-to-lake-surface-area
ratios and receive higher sediment and nutrient loads
(per unit surface area per annum) than do most natural
lakes (Thornton et al. 1981; Benson, 1982).
  Probably more reservoirs than  natural lakes  are
located in fertile drainage basins and, therefore, the
natural "trophic equilibrium" level (Hutchinson, 1973;
Kimmel and Groeger, in press) is, in general, higher for
                                                 277

-------
 LAKE AND RESERVOIR MANAGEMENT
 most reservoirs than for most natural lakes. Fishery
 biologists may have the opposite impression: i.e., that
 in terms of fishery production or yield per unit areia,
 natural lakes tend to be more productive than reser-
 voirs. If this is the case, it can probably be explained
 by the primary production and favorable fish habitat
 associated with littoral zones in natural lakes. In con-
 trast, many reservoirs,  especially flood control and
 hydropower impoundments, essentially lack  littoral
 zone vegetation as a result of the effects of pronounc-
 ed seasonal water level  fluctuations.


 ENVIRONMENTAL FACTORS
 CONTROLLING LAKE AND RESERVOIR
 PHYTOPLANKTON PRODUCTION

 The  basic factors controlling phytoplankton produc-
 tivity (temperature, light availability, macro- and micro-
 nutrient availability)  have been  reviewed extensively
 elsewhere (e.g., Tailing, 1961; Lund, 1965; Goldman
 1968; Fogg, 1975; Harris, 1978; Westlake et al. 1980). In
 general, phytoplankton production is controlled by the
 same energy and nutrient inputs (Brylinsky and Mann,
 1973; Schindler, 1978; Brylinsky, 1980) and the same
 gain and loss processes (Jassby and Goldman, 1974;
 Kalff and Knoechel, 1978; Crumpton and Wetzel, 1982)
 in reservoirs  as  in  natural  lakes.  Fluctuations  in
 phytoplankton standing  crop reflect changes in the
 net  balance  of  biomass gains (by advection  and
 growth) and losses (by advection, respiration, sinking,
 grazing, and  other sources of  mortality and loss).
 Although the relative importance  of some  of thes.e
 controlling factors and processes  may vary between
 lakes and reservoirs, the extent of  variation  probably
 does not greatly exceed  that which occurs commonly
 within individual impoundments (Fig. 3).
             Whereas lotic ecosystems are characterized by
           longitudinal gradients  in channel morphology,  flow
           velocity, water temperature, substrate type, and biota
           (e.g., Hynes, 1970; Cummins, 1974; Vannote et al. 1980;
           Minshall  et  al.  1983),  vertical  gradients of  light,
           temperature, dissolved substances, and production
               60
               50
            > 40
            o
            o
            "J  30
            oc
               20
            UJ
            CC
               10
                      |    1 NATURAL LAKES

                      Y///A RESERVOIRS
                 OLIGOTROPHIC  MESOTROPHIC  EUTROPHIC
           Figure 2.—Frequency distribution of the trophic status of
           natural lakes and reservoirs, as reflected by the average daily
           phytoplankton production on an annual basis. Trophic state
           categories are those of Likens (1975) and Wetzel (1975); oligo-
           trophic: 300  mg C m-2 day-'. The data set, which included
           102 natural  lakes  and 64 reservoirs, was compiled from
           Wetzel (1975), Brylinsky (1980), and Kimmel et al. (in press)
                                       RESERVOIRS
         MAINSTREAM
   "RUN-OF-THE-RIVER"
         RESERVOIRS
MAINSTREAM
  STORAGE
 RESERVOIRS
 TRIBUTARY
  STORAGE
RESERVOIRS
                          INCREASING  RETENTION  TIME

                 t°CfCUPy an in(t,ermediate P°sition between rivers and natural lakes on  a continuum of aquatic
nnnnHr    i         of.nverine lnf|uence and the hydraulic retention time determine the relative positions of various im-
poundment types (e.g., mainstream run-of-the-river, mainstieam storage, tributary storage) along the river-lake continuum
                                             278

-------
                                                                    COMPARATIVE ANALYSIS OF RESERVOIRS
and  decomposition  processes  are  predominant
features  of lentic  environments (e.g.,  Hutchinson,
1957; Wetzel, 1975). Reservoirs possess both horizon-
tal and vertical gradients of the environmental factors
that control primary production. Longitudinal changes
in basin  morphology and  flow velocity result  in dif-
ferences  in  suspended   particle  concentrations,
nutrient levels,  mixing depth, and thereby,  in varia-
tions in light  and nutrient availability for primary pro-
duction in various parts of the reservoir.
  Because a transition from a riverine environment to
a lacustrine environment occurs within the reservoir
basin,  reservoirs characteristically exhibit a striking
degree of spatial heterogeneity in phytoplankton pro-
ductivity and biomass (Kimmel et al. in press). Similar
longitudinal productivity gradients occur in riverine
estuaries (Stress and Stottlemyer, 1965) and in natural
lakes receiving  substantial inflows or point source
nutrient inputs (Gascon and Leggett, 1977).
  What are the management implications of the mark-
ed spatial variation of primary production within reser-
voirs?
   1. Plans for sampling and monitoring of reservoirs
must take into account  this characteristic spatial
heterogeneity. Obviously, the standard practice of ob-
taining a  representative  sample or profile  in the
deepest part of the lake basin is inadequate for char-
acterizing  such  spatially  heterogeneous systems.
Thornton et al. (1982) provide a detailed discussion of
reservoir sampling strategies.
   2. Similarly,  because of  longitudinal gradients  in
water  quality  and productivity  within  reservoirs,
classic trophic classifications and indices are less ap-
propriate for impoundments than for  natural  lakes.
Within a reservoir basin, the fertility of the mixed layer
(i.e., in terms of the maximum  photosynthesis rate
(Pmax)  or °f  tne phytoplankton  productivity m-3 of
euphotic zone) generally decreases downlake  as the
advective nutrient supply decreases (Fig. 3). Trophic
state (as reflected by Secchi depth, phosphorus levels,
chlorophyll concentrations, phytoplankton productivi-
ty, dissolved oxygen  depletion,  or indices based on
these parameters) shifts from more eutrophic to more
oligotrophic conditions  along  the riverine-transi-
tional-lacustrine zone gradient (e.g., Thornton et al.
 1981; Hannan et al. 1981; Kennedy et al 1982; Kimmel
et al. in press).
  RIVERINE ZONE
                  TRANSITIONAL ZONE     LACUSTRINE ZONE
                     JCED SUSP SOLIDS 1
 Figure 3.—Longitudinal zonation in environmental factors
 controlling  primary productivity, phytoplankton  biomass,
 and trophic state within reservoir basins.
  3.  Finally, the longitudinal gradient in productivity
within reservoirs suggests that a corresponding zona-
tion may occur in the relative suitability of portions of
reservoirs for various uses. For example, water supply
intakes located in the less productive, lacustrine zone
will likely have fewer filter clogging and taste and odor
problems than intakes located farther uplake. Boaters
and swimmers may prefer the clearer, lacustrine por-
tions of impoundments, whereas fishermen will usual-
ly choose the more productive transitional and riverine
regions.  Siler and Foris (in press) documented longi-
tudinal trends in nutrient levels,  phytoplankton bio-
mass, forage fish standing stock, sport fish harvest,
and fishing pressure in  Lake Norman reservoir,  N.C.;
they suggest that managing a reservoir as a uniform
biological entity may be ineffective because of the ex-
isting spatial heterogeneity and propose that applying
different fishery management  strategies  in various
portions  of  reservoir  basins  could  significantly
enhance reservoir fishery production.
APPLICABILITY OF NUTRIENT LOADING
MODELS TO LAKES AND RESERVOIRS

The cultural eutrophication of freshwater resources is
of  international  concern  and,  therefore,  much
research has been conducted to assess and predict
the trophic status of lacustrine systems (e.g., Vollen-
weider, 1968, 1976;  Likens,  1972;  Lee et al.  1972;
Reckhow, 1979; Vollenweiderand Kerekes, 1980). Con-
sequently,   numerous  nutrient   loading-trophic
response models have been developed and are now in
common use (reviewed by Reckhow, 1979). Many of
these empirical models, which predict the response of
the average phytoplankton biomass level to the  an-
nual phosphorus loading rate as influenced by basin
morphometry and hydraulic retention time, are based
on data derived primarily from natural lakes. Although
the primary producers and environmental factors con-
trolling primary production are  very similar in  lakes
and reservoirs, these correlative relationships should
be applied with caution to reservoirs for a number of
reasons:
   1. Many  reservoirs  are located  in geographic
regions which are poorly represented in  most nutrient
loading model data sets.
   2. In reservoirs receiving  high  concentrations of
abiogenic suspended solids,  a significant fraction of
the  total phosphorus  loading  may be  either  bio-
logically unavailable (Sonzogni et al. 1982) or rapidly
lost to the sediments (Chapra, 1980; Gloss et al. 1981).
However, this problem may be correctable by applying
more suitable phosphorus sedimentation coefficients
for reservoirs (Jones and Bachmann, 1978; Canfield
and Bachmann, 1981; Higgins et al. 1981).
   3. Phosphorus availability is not necessarily  the
primary factor limiting algal growth in reservoirs as is
assumed in most nutrient loading models, but is only
one of many environmental factors that can control
algal abundance. In  particular,  low light availability
often moderates the effects of nutrient  loading in  tur-
bid reservoirs (Kimmel and Lind, 1972; O'Brien, 1975;
 Kimmel, 1981; Marzolf, in  press) and in  deeply mixed,
riverine impoundments (Placke,  1983).
   4. Nutrient retention in shallow, rapidly flushed  sys-
tems (e.g., mainstream reservoirs) is low compared to
 lakes with long retention times. Different mechanisms
 may  govern  nutrient  loading-trophic  response
 relationships in  rapidly flushed lakes and reservoirs
(Chapra, 1975;  Higgins et  al.  1981; Turner et al. 1983).
                                                  279

-------
  LAKE AND RESERVOIR MANAGEMENT
   5. Because density flows and hypolimnetic outlets
 commonly occur in reservoirs, the  annual nutrient
 loading to a reservoir may significantly overestimate
 the actual nutrient supply to the phytoplankton during
 the growing season. Additionally, the average hydrau-
 lic retention time (total annual inflow/reservoir volume)
 of a reservoir may have little relationship to the actual
 flushing rate of the  trophogenic zone.
   As  concluded by  Reckhow (1979), to appropriately
 use  empirical  models  and   indices  for  reservoir
 management, careful consideration must be given to
 their underlying assumptions,  the limitations  of the
 data sets upon  which they are  based, and the degree
 of uncertainty associated with  their predictions.

 CONCLUSIONS

 The primary producers and the basic environmental
 factors and processes controlling primary production
 in natural and manmade lakes are identical. However,
 some  significant  physical  and hydrodynamic  differ-
 ences between  lakes and reservoirs do exist that in-
 fluence (1) the relative contributions of various primary
 producers to their foodwebs,  (2) the relative impor-
 tance of certain  limiting factors to primary production
 (e.g., turbidity, nutrient  availability,  flushing  rate), (3)
 the  spatial variability of  primary  production within
 reservoirs,  and  (4)  the   applicability  of  lake-based
 classifications  and  empirical relationships to reser-
 voirs.
   Perhaps the  primary  distinction  between natural
 and manmade lakes  is that reservoirs are  semi-fluvial
 environments that fall between  rivers and lakes on a
 continuum of aquatic ecosystems (Fig. 1). We suggest
 that such a view, which recognizes riverine influences
 on reservoir  ecosystems,  can improve  the under-
 standing of the spatial  and temporal heterogeneity
 associated with  manmade impoundments, and, there-
 by, the management of reservoir resources.

 ACKNOWLEDGEMENTS: We thank S.  M.  Adams, D. M.
 Soballe, W. Van Winkle, and C. W. Gehrs for their comments
 on the manuscript. Research  sponsored  by the Office of
 Health  and  Environmental  Research, U.S.  Department of
 Energy. A. W. Groeger  was supported by a DOE Laboratory
 Research Participation  predoctoral  fellowship administered
 by Oak Ridge Associated Universities. Oak Ridge National
 Laboratory is operated by the Union Carbide Corp. for the
 U.S.  Department  of Energy  under  contract  W-7405-eng-20.
 Publ. No. 2270, Environmental Sciences Division, ORNL


 REFERENCES

 Benson, N.G. 1982. Some observations on the  ecology and
  fish management of  reservoirs in the United  States Can
  Water Res. J. 7:2-25.
 Brylinsky, M. 1980. Estimating the productivity of lakes and
  reservoirs. Pages 411-453 in E.D.  LeCren and R.H. Lowe-
  McConnell, eds. The Functioning of Freshwater Ecosys,-
  tems. Cambridge Univ. Press, London.
 Brylinsky, M., and  K.H. Mann. 1973. An analysis of factors
  governing  productivity in  lakes  and reservoirs. Limno.
  Oceanogr  18:1-14.
Canfield, D.E., and R.W. Bachmann.  1981. Prediction of total
  phosphorus concentrations, chlorophyll a and Secchi dep-
  ths in natural and artificial lakes.  Can. J. Fish. Aquat. Sc
  38:414-23.
Chapra, S.C. 1975. Comment on " An empirical method of
  estimating the retention of phosphorus in lakes," by W.E.
  Kirchner and P.J. Dillon. Water Resour. Res. 2:1033-4.
 	1980.  Application  of  phosphorus  loading
   models to river-run lakes and other incompletely mixed
   systems. Pages 329-334 in Restoration of Lakes and Inland
   Waters. EPA 440/5-81-010. U.S. Environ. Prot.  Agency
   Washington, D.C.

 Crumpton, W.G., and R.G. Wetzel. 1982. Effects of differ-
   ential growth and mortality in the seasonal succession of
   phytoplankton  populations  in Lawrence  Lake,  Mich
   Ecology 63:1729 39.

 Cummins, K.W. 1974. Structure and function of stream eco-
   systems. Bioscience 24:631-41.

 Ellis, M.M. 1936. Erosion  silt as a factor in aquatic environ-
   ments. Ecology 17:29-42.

 Fogg, G.E. 1975. Algal Cultures and Phytoplankton Ecology.
   2nd ed. Univ. Wisconsin Press, Madison.

 Gascon, D., and W.C. Leggett. 1977. Distribution, abundance,
   and resource utilization of littoral zone fishes in response
   to a nutrient/production gradient in Lake Memphremagog
   J. Fish. Res.  Board Can. 34:1105-17.

 Gloss, S.P., R.C. Reynolds, L.M. Mayer, and  D.E.  Kidd. 1981.
   Reservoir influences on salinity and nutrient fluxes in the
   arid Colorado River basin. Pages 1618-1629 in H.G. Stefan,
   ed. Proc. Symp. Surface Water Impoundments. Am  Soc
   Civil Eng., New York.

 Goldman, C.R. 1968. Aquatic primary production  Am  Zool
   8:31-42.

 Hannan, H.H., D.  Barrows, and D.C. Whitenberg. 1981. The
   trophic status of a deep-storage reservoir in Central Texas
   Pages 425-434  in H.G.  Stefan, ed.  Proc.  Symp. Surface
   Water Impoundments. Am Soc. Civil Eng., New York.
 Harris, G.P. 1978. Photosynthesis, productivity and growth:
   The physiological ecology of phytoplankton. Arch. Hydro-
   biol. Beih. Ergeb. Limnol.  10:1-171.

 Higgins,  J.M.,  W.L Poppe,  and  M.L  Iwanski. 1981. Eutro-
   phication analysis of TVA reservoirs. Pages  404-412 in
   H.G. Stefan,  ed.  Proc.  Symp.  Surface Water Impound-
   ments. Am. Soc. Civil Eng., New York.

Hutchinson, G.E. 1957. A Treatise on Limnology. Vol. 1. John
  Wiley and Sons, New York.

	1973. Eutrophication. Am. Sci. 6:269-79.
Hynes, H.B.N. 1970 The Ecology of Running Waters Univ.
  Toronto Press, Toronto.
Jassby, A.D., and C.R. Goldman. 1974. Loss rates from a lake
  phytoplankton community. Limnol. Oceanogr. 19:618-27.
Jones, J.R., and  R.W. Bachmann. 1978. Phosphorus removal
  by sedimentation in  some Iowa reservoirs.  Verh.  Int.
  Verein  Limnol. 20:1576-80.
Kalff, J.,  and  R.  Knoechel.  1978.  Phytoplankton and their
  dynamics in oligotrophic and eutrophic lakes.  Ann.  Rev
  Ecol. Syst. 9:475-95

Kennedy, R.H., K.W. Thornton, and R.C. Gunkel.  1982. The
  establishment of water quality gradients in reservoirs.
  Can. Water Res. J. 7:71-87.

Kimmel, B.L 1981. Land-water interactions: Effects of intro-
  duced nutrients and soil particles on reservoir productivi-
  ty. Tech. Comple. Rep. Proj. No. A-088-OKLA. Office Water
  Res. Technol.  U.S. Dep. Inter., Washington, D.C.
Kimmel, B.L, and A.W. Groeger. In press. Limnological and
  ecological changes associated with reservoir aging. In
  G.E. Hall, ed. Reservoir Fisheries Management: Strategies
  for the 80's. Alien Press, Lawrence, Kans.

Kimmel, B.L, and  O.T. Lind. 1972. Factors affecting phyto-
  plankton production in a eutrophic reservoir. Arch. Hydro-
  biol. 71:124-41

Kimmel, B.L, O.T. Lind, and L J. Paulson. In press. Reservoir
  primary production. In K.W. Thornton, ed. Perspectives on
  Reservoir Ecosystems. John Wiley and Sons, New York.
                                                    280

-------
Lee, G.F., M. Rast, and R.A. Jones. 1972. Eutrophication of
  water bodies: Insights for an age-old problem. Environ. Sci.
  Tech. 12:900-8.
Likens,  G.E.   1972.  Nutrients and   Eutrophication:  The
  Limiting  Nutrient Controversy. Spec. Symp. 1. Am. Soc.
  Limnol. Oceanogr.
Lund, J.W.G. 1965. The ecology of the  freshwater phyto-
  plankton. Biol. Rev. 40:231-93.
Marzolf, G.R. In press. Reservoirs in the Great Plains of North
  America. In F.B. Taub, ed. Lake and Reservoir Ecosystems.
  Elsevier Sci.  Publ., Amsterdam.
Minshall, G.W., et al. 1983. Interbiome comparison of stream
  ecosystem dynamics. Ecol. Monogr. 53:1-25.
O'Brien, W.J. 1975. Factors limiting primary productivity in
  turbid Kansas  reservoirs.  Tech.  Comple.  Rep.  Proj.
  A-0520-KAN. Office Water Res. Technol., U.S. Dep. Inter.,
  Washington, D.C.
Placke, J.F. 1983.  Trophic status evaluation  of TVA reser-
  voirs. Tech. rep. TVA/ONR/WR-83/7. Tenn. Valley Author.
Reckhow, K.H. 1979. Empirical lake models for phosphorus:
  Development, applications,  limitations and uncertainty.
  Pages  193-221  in D.  Scavia  and  A. Robertson,  eds.
  Perspectives on Lake  Ecosystem Modeling. Ann Arbor
  Sci., Ann Arbor, Mich.
Ryder, R.A. 1978.  Ecological  heterogeneity between north-
  temperate  reservoirs and glacial lake systems due to dif-
  fering succession rates and  cultural uses.  Verh.  Int.
  Verein. Limnol. 20:1568-74.
Schindler,  D.W.  1978.  Factors regulating phytoplankton
   production  and  standing crop in the world's freshwaters.
   Limnol. Oceanogr. 23'478-96.
Siler, J.R., and W.J. Foris. In press.  Heterogeneity in  fish
   parameters  within a reservoir. In G.E. Hall, ed. Reservoir
   Fisheries  Management: Strategies for the 80's. Allen
   Press, Lawrence,  Kans.
Sonzogni, W.C., S.C. Chapra, D.E. Armstrong, and T.J. Logan.
   1982. Bioavailability of  phosphorus inputs to lakes. J. En-
   viron. Qual.  11:555-63.
                  COMPARATIVE ANALYSIS OF RESERVOIRS

Stress, R.G., and J. Stottlemeyer. 1965. Primary production in
  the Patuxent River. Chesapeake Sci. 6:125-40.
Tailing, J.F. 1961. Photosynthesis under natural conditions.
  Ann. Rev. Plant Physiol. 12:133-54.

Thornton, K.W., R.H. Kennedy, A.D. Magoun, and G.E. Saul.
  1982. Reservoir water quality sampling design. Water Res.
  Bull. 18:471-80.
Thornton, K.W.  et  al.  1981.  Reservoir sedimentation and
  water quality — an heuristic model. Pages 654-661 in H.G.
  Stefan, ed.  Proc.  Symp. Surface Water Impoundments.
  Am. Soc. Civil Eng., New York.
Turner, R.R., E.A. Laws, and R.C. Harris. 1983. Nutrient reten-
  tion and transformation  in relation to hydraulic  flushing
  rate in a small impoundment. Freshw. Biol. 13:113-27.
Vannote, R.L, et al. 1980. The river continuum concept. Can.
  J. Fish. Aquat. Sci. 37:130-37.
Vollenweider,  R.A.  1968. Scientific  fundamentals of the
  eutrophication of lakes and flowing waters, with particular
  reference to  nitrogen  and  phosphorus  as  factors in
  eutrophication.  Tech.  rep.  DAS/CS 1/68.27:1-159. Organ.
  Econ. Coop. Dev., Paris.
	.  1976. Advances in defining critical loading levels
  for phosphorus  in lake eutrophication. Mem.  1st.  Ital.
  Idrobiol. 33:53-83.
Vollenweider,  R.A., and J.J. Kerekes. 1980. Background and
  summary results  of the OECD  cooperative program on
  eutrophication. Pages  25-36 in  Restoration of Lakes and
  Inland Waters. EPA 440/5-81-010. U.S. Environ. Prot. Agen-
  cy, Washington, D.C.
Westlake, D.F., et  al. 1980. Primary production. Pages 141-
  246 in  E.D. LeCren and R.H. Lowe-McConnell, eds. The
  Functioning of Freshwater Ecosystems. Cambridge Univ.
  Press. London.
Wetzel, R.G. 1975.  Limnology. W.B. Saunders, Philadelphia.
                                                        281

-------
 ORGANIC MATTER SUPPLY AND PROCESSING IN
 LAKES AND RESERVOIRS
 ALAN W.  GROEGER
 Department of Zoology
 University of Oklahoma
 Norman, Oklahoma

 BRUCE  L KIMMEL
 Environmental Sciences Division
 Oak Ridge National  Laboratory
 Oak Ridge, Tennessee



             ABSTRACT

             The estimation of annual organic matter budgets for lakes and reservoirs can be an important in-
             itial step in investigating aquatic ecosystem structure and function. Comparison of organic mat-
             ter budgets for a number of aquatic systems indicates that hydraulic retention time exerts a
             critical influence on the efficiency with which those systems process and trap their organic mat-
             ter inputs. Generally, lakes or reservoirs with short  retention times have high drainage basin
             area: water surface area ratios, and receive large quantities of allochthonous organic matter.
             These rapidly flushed ecosystems are relalively inefficient  in retaining and processing their in-
             puts and transport large quantities of organic matter downstream. Lakes or reservoirs with long
             retention times have low drainage basin area: water surface area ratios and are dominated by
             autochthonous inputs. These ecosystems are considerably more efficient in retaining and pro-
             cessing their  inputs, and lose little organic matter downstream. Lakes and reservoirs with
             similar retention times and drainage basin area: water surface area ratios tend  to behave
             similarly, both in the relative importance of allochthonous inputs and in the efficiency with
             which organic matter inputs are processed  and trapped within the ecosystem.
INTRODUCTION

Lake and reservoir ecosystems depend on inputs of
both autochthonous and allochthonous organic mal-
ter for their energy base. Photosynthetic generation of
organic matter by algae and aquatic macrophytes
comprise the  autochthonous organic  matter  inpul.
Allochthonous organic matter is formed outside the
ecosystem, and is imported via precipitation, dry falI-
out,  ground water,  lateral transport, and inflowing
rivers and streams.  The quantity and quality of this
organic matter supply and the relative proportion of
allochthonous to autochthonous inputs influence
foodweb structure and determine the potential secon-
dary productivity for lakes and reservoirs.
   Because dams are barriers to riverine drainage
systems, reservoirs often receive large allochthonous
organic matter loads relative to most  natural  lakes
where autochthonous production is more commonly
the predominant source of organic  matter (Wetzei,
1975).  Here  we  compare  annual organic  matter
budgets for a number of natural lakes and reservoirs,
and address two questions: (1) Do lakes and reservoirs
differ significantly in the organic matter supply they
receive? and (2) are lakes and reservoirs fundamental-
ly  different in their efficiencies of organic matter pro-
cessing and retention?
METHODS

Data on annual organic matter budgets were compiled
from both published and unpublished sources (Table
1). We compared organic matter budgetary dynamics
in lakes  and  reservoirs  by examining  values  for
organic matter retention efficiency:
                organic matter
            retention efficiency (%) =
x 100
where I  equals the total annual organic matter input
and L the annual downstream loss of organic matter.
  Organic matter retention efficiency is not an ideal
index for comparing ecosystems because the differ-
ence between inputs and outputs does not distinguish
between community respiration, or the actual pro-
cessing of organic matter, and organic matter loss by
permanent  sedimentation. However, organic matter
retention does provide a basis for initial comparisons
of organic matter dynamics within lake and reservoir
ecosystems, and can have important implications for
downstream ecosystems.
RESULTS AND DISCUSSION

Lakes and reservoirs can generally be separated by
certain hydrologic and morphometric characteristics
(Thornton et al. 1981). A particularly important distinc-
tion is that reservoirs tend to  have larger drainage
basin  area: water surface area ratios (Ad:A0) than do
natural lakes. As the Ad:A0 increases, the hydraulic
retention time (when expressed in days  =  (lake
volume/annual  inflow  volume)  x  365)  generally
decreases. Assuming comparable areal rates of ter-
restrial production, larger watersheds contribute more
allochthonous organic matter to aquatic systems than
do small watersheds, and accordingly, an inverse rela-
tionship exists between areal allochthonous organic
matter loading and retention time (Fig. 1). In this data
                                                282

-------
                                                                    COMPARATIVE ANALYSIS OF RESERVOIRS
set, as  retention time approaches  1  yr, the alloch-
thonous  loading rates  are  approximately <. 100  g
organic matter m -2 yr-1. The two systems with reten-
tion times  <15  days have loading  values > 1,500 g
organic  matter m-2  yr-1. Lakes and  reservoirs with
intermediate retention times tend to also have inter-
mediate allochthonous loading values. The high areal
loading value for Slapy Reservoir can  be at least  par-
tially explained by its mean depth >  22  m, which  is
about three times greater than the  average mean
depth for the data set.
  Organic  matter retention efficiency is directly
related  to  the  hydraulic  retention  time (Fig. 2).  A
number of interrelated factors help explain this situa-
tion. As already shown (Fig. 1), ecosystems with short
retention times tend to have high areal organic matter
loading.  In systems  such  as  these,  which  are
dominated  by fluvial  processes, their allochthonous
inputs are composed largely of dissolved organic mat-
ter  (DOM)  (Odum and  Prentki, 1978).  While DOM
passes through  a lake or reservoir  at the same  rate
that water  is exchanged, particulate  organic matter
(POM) passes through  at a much  slower  rate as a
result of both biological incorporation and physical
settling. The allochthonous DOM pool  tends to be of a
largely refractory nature, and its decomposition  rate
may be very slow relative to the retention time of these
rapidly flushed systems. In general, the physical reten-
tion and biological processing of organic matter is a
function of the residence time of the organic matter
within that aquatic ecosystem. In Findley Lake, which
has a moderate retention time of about 50 days, 88
percent of  the allochthonous POM  is retained while
virtually all of the DOM  is lost downstream (Wissmar
et al. 1977; Richey et al. 1978;  Richey and Wissmar,
1979). It should be noted, however, that because of the
large absolute dominance of allochthonous DOM  in
some of these systems,  if even a small fraction of this
pool is utilized it may be an important energy contribu-
tion to the  foodweb.
  Lakes and reservoirs with high Ad:A0 ratios often
undergo periods of extremely rapid flushing (retention
times of a  few days  or  less) during heavy watershed
runoff.  These hydrologic events, which occur  over
short periods, account for a major fraction of both the
annual water and organic matter inputs  and outputs
of such ecosystems (Lind, 1971; Richey and Wissmar,
1979). Groeger and King (subm.) found that 40 percent
of the annual allochthonous organic matter input to a
mainstream Michigan reservoir (Lake Isabella) entered
and flowed through the system during a 4-week period
of high spring flow.  This dramatic  pulsing of inputs
and outputs, where large quantities of allochthonous
organic matter enter and flow through systems  very
rapidly, may also be a  key factor in explaining their
low organic matter retention efficiencies.
  Autochthonous organic matter is generally believed
to be higher in  energy  and  nutrient content per unit
weight  (Hallegraeff 1978; Wissmar et al. 1977), and
more biologically labile (Wetzel, 1975) than  alloch-
thonous material. These characteristics, plus the fact
that autochthonous  inputs are largely in particulate
form, suggest that autochthonous  organic matter is
differentially  processed  relative  to allochthonous
organic  matter. Slapy  Reservoir,  Findley  Lake  and
Lake Paajarvi (retention times of 38, 51,  and 1,200
days, respectively) have  low  autochthonous inputs
relative to  their  allochthonous loads (autochthonous:
allochthonous  input  ratios of 0.03,  0.06, and  0.59,
respectively) and lie far  below predicted values  of
organic matter retention efficiency (Fig. 2). In contrast,
Lake Wingra, the only system that lies far above the
regression  line in  Figure 2, has an  autochthonous:
allochthonous input ratio of 11.5.


CONCLUSION

The  hydraulic retention time of a lake or reservoir
plays a critical role in determining the relative efficien-
  10,000
                             ORNL-DWG 83-12592A
£
~5?
Q
O
    1000
I
O
o

<

 g

O
     100
      10
                              • LAKE
                              o RESERVOIR
            Ma
                         Fi
                              La«
                                           Ba
                                        I
             10           100          1000
               LOG,0  RETENTION  TIME (d)
Figure 1.—Log10 allochthonous organic matter loading (g
m-2 yr-i) versus Log10 retention time (days) for a number of
natural and  man-made lakes, (r  =  -0.80, P < 0.01). The
points are labeled  to correspond to lakes and reservoirs
listed in Table 1.
    100
                            ORNL-DWG 83H2593A
              10            100           1000
               LOG,0  RETENTION TIME  (d)


Figure 2.—Organic matter retention efficiency (%) versus
Log10 retention time (days) for a number of natural and man-
made lakes (r = 0.81, P <0.01). The points are labeled to cor-
respond to lakes and reservoirs listed in Table 1.
                                                 283

-------
Table 1.—Annual organic matter budget data for lakes and reservoirs used in this analysis. Conversion factors

      used to convert reported values into g organic matter: (1) Kcal x 0.214, (2) g organic carbon x 2,
;*
m

System
Lakes Lake
Balaton (Ba)
Mirror
Lake (Mi)
Lawrence
Lake (La)
Red
Lake (Re)
Wingra
Lake (Wi)
Lake
Kinneret (Ki)
Lake
Paajarvi (Pa.)
Findley
Lake (Fi)
Marion
Lake (Ma)
Reservoirs DeGray
Reservoir (De)

Rybinsk
Reservoir (Ry)
Lake
Waco (Wa)
Kiev
Reservoir (Kv)
Kremenchug
Reservoir (Kr)
Ivankovo
Reservoir (Iv)
Lake
Isabella (Is)
Slapy
Reservoir (SL)

Allochthonous
Inputs
(gm-2yr-i)
10-13
36
50
84
104
121
126

164
1580
99

232
364
408
410
917
2500
6361
I
Autochthonous
Inputs
(gm-2yr-i)
250
•"V)
342
269
1200
1312
74

10
160
145

163
626
589
—
1376
600-
1000
183
\<3f KJ x u.uo i , ana \
Autochthonous:
Allochthonous
Inputs
19-25
2.8
6.8
3.2
11.5
10.8
0.6

0.06
0.1
1.5

0.7
1.7
1.4
—
1.5
0.2-
0.4
0.03
i) g i,vu x
Retention
Time
(days)
1350
365
274
420
175
900-
1736
1205

51
5.5
374

251
72
30-41
162
41
13
38
u./o.
A^
9.7
5.7
7.1
2.0
11.4
15.1
18.2

11.3
977
22.1

—
145
24.2
—
—
83
939

Mean
Depth
(m)
3.2
5.6
5.9
6.6
2.4
24
14.4

7.8
2.4
9.0

5.6
4.0
4.0
6.0
3.5
2.6
22.3

Organic Matter
Retention
Efficiency (%)
—
85
80
86
95
99
64

15
14
74

77
65
58
—
57
27-36
28

Conversion
Factors
1
2
2
1
2
2
3

2
2
2

2
1,2
1
2
2

4

Primary
Source
Olah, 1978
Jordan & Likens,
1975
Wetzel et al. 1972
Andronikova et al.
1972
Loucks & Odum, 1978;
Richey et al. 1978
Serruya et al. 1980
Sarvala et al 1981

Loucks & Odum, 1978;
Richey et al. 1978
Loucks & Odum, 1978;
Richey et al. 1978
R.H. Kennedy,
pers. comm.
Romanenko, 1978
Una, 1971;
Kimmel & Lind, 1972
Gak et al. 1972
Denisova&
Palamarchuk, 1977
Kadukin et al. 1980
Groeger & King,
submitted
Hrbacek et al. 1966
AND RESERVOIR
-z.
a
m
m
Z
H

















-------
                                                                           COMPARATIVE ANALYSIS OF RESERVOIRS
cy of organic matter retention and processing. Reten-
tion time also affects thermal structure and stability
(Johnson et al.  1978), sedimentation characteristics
(Rausch  and  Schreiber,  1981),  nitrogen  and  phos-
phorus loading (Vollenweider, 1976), nutrient recycling
(Devol  and  Wissmar, 1978), nutrient retention  and
primary productivity  (Turner et  al. 1983), and  chlor-
ophyll dynamics (Soballe, in prep.) within lentic  sys-
tems. Canfield and Bachman (1981) and Kimmel et al.
(in press) have suggested that lakes and reservoirs are
not distinctly different aquatic ecosystems but repre-
sent a range of limnological  conditions that can be ar-
ranged along  a  continuum.  We suggest further  that
retention time is a characteristic upon which such a
continuum or gradient can be constructed. In answer
to  the  questions  posed   initially,  because  the
"average" lake and reservoir can be separated by their
Ad:A0 relationships and retention times, they differ in
such basic properties as  areal organic matter loading
rates and organic matter retention efficiencies. Con-
versely, in dealing with  individual ecosystems  that
have similar retention times reservoirs and lakes ap-
pear to function in a  quite similar fashion.

ACKNOWLEDGEMENTS: We  thank  S.M.  Adams,  D.M.
Soballe, C.W. Gehrs, and W. Van Winkle for their comments
on  the manuscript. We would  like to thank R.H. Kennedy,
U.S. Army Engineer Waterways Experiment Station, for sup-
plying  unpublished data on DeGray Lake.  Research spon-
sored by the Office of Health and Environmental Research,
U.S. Department of Energy. A.W. Groeger funded under ap-
pointment to the Laboratory Graduate Participation Program
administered by Oak Ridge Associated Universities for the
U.S. Department of Energy. Oak Ridge National Laboratory is
operated by the Union Carbide Corp., for the U.S. Department
of Energy under contract W-7405-eng-26. Publ. No. 2269, En-
vironmental Sciences Division,  ORNL.


REFERENCES

Andronikova, I.N., et al. 1972. Biological productivity of the
   main communities of the Red Lake. Pages 57-71 in Z. Ka-
   jak and A. Hillbricht-llkowska, eds. Productivity Problems
   of Freshwaters. Polish Sci. Publ., Warsaw and Krakow.
Canfield, D.E., and R.W. Bachman. 1981.  Prediction of total
   phosphorus  concentrations,  chlorophyll  a and  Secchi
   depths in natural and artificial lakes. Can. J. Fish. Aquat.
   Sci. 38:414-23.
Denisova, A.I., and I.K. Palamarchuk. 1977. Balance of bio-
   genie and organic materials in the Kremenchug Reservoir.
   Water Resour. 4:48-59.
Devol, A.M., and R.C. Wissmar. 1978. Analysis of five North
   American  lake ecosystems. V.  Primary  production and
   community structure. Verh. Int. Verein. Limnol. 20:581-6.
Gak, D.Z., et al. 1972. Productivity of aquatic organism com-
   munities of  different trophic levels in  Kiev Reservoir.
   Pages 447-55 in Z. Kajak and A.  Hillbricht-llkowska, eds.
   Productivity  Problems of Freshwaters. Polish Sci. Publ.,
   Warsaw and Krakow.
Groeger, A.W., and R.H. King.  Organic matter budget for a
   Michigan reservoir. Freshw. Biol. (Subm.)
 Hallegraeff, G.M.  1978.  Caloric content and elementary com-
   position of seston of three Dutch freshwater lakes. Arch.
   Hydrobiol. 83:80-98.
 Hrbacek, J., L. Prochazkova, V.  Straskrabova-Prokesova, and
   C.O.  Junge. 1966. The relationship between the chemical
   characteristics of the Vltava River and  Slapy  Reservoir
   with  an Appendix: Chemical  budget for Slapy Reservoir.
   Pages 41-84 in  J. Hrbacek, ed. Hydrobiological Studies 1.
   Academia Publ. House  of the Czechoslovak Acad. Sci.,
   Prague.
Johnson, N.M., J.S. Eaton, and J.E. Richey. 1978. Analysis of
  five North American lake ecosystems. II. Thermal energy
  and  mechanical  stability.  Verh. Int.  Verein.  Limnol.
  20:562-7.
Jordan, M., and G.E. Likens. 1975. An organic carbon budget
  for an  oligotrophic lake in New Hampshire,  U.S.A. Verh.
  Int. Verein. Limnol. 16:994-1003.
Kadukin, A.I., et al. 1980. Balance  of organic  matter, bio-
  genie elements, and trace elements in the Ivankovo Reser-
  voir. Water Resour. 7:346-55.
Kennedy, R.H.  Pers. comm. U.S. Army Corps of Engineers,
  Vicksburg, Miss.
Kimmel, B.L., and O.T. Lind. 1972. Factors affecting phyto-
  plankton  production in a eutrophic reservoir. Arch. Hydro-
  biol. 71:124-41.
Lind, O.T. 1971. The organic matter budget of a central Texas
  reservoir.  Pages  193-202  in  G.E.  Hall,  ed.  Reservoir
  Fisheries  and Limnology.  Am. Fish.  Soc., Washington,
  D.C.

Loucks, O.L, and W.E. Odum. 1978. Analysis of five North
  American lake ecosystems. I.  A strategy for comparison.
  Verh. Int. Verein. Limnol. 20:556-61.
Odum, W.E., and R.T. Prentki. 1978. Analysis of five North
  American lake ecosystems. IV. Allochthonous carbon in-
  puts. Verh. Int. Verein. Limnol. 20:574-80.
Olah, J.  1978.  The annual energy budget of Lake Balaton.
  Arch. Hydrobiol. 81:327-38.
Rausch, D.L, and J.D. Schreiber.  1981. Sediment and nutrient
  trap efficiency of a small flood-detention reservoir. J. En-
  viron. Qual.  10:288-93.
Richey, J.E., and R.C. Wissmar. 1979. Sources and influences
  of allochthonous inputs on the productivity of a subalpine
  lake.  Ecology 60:318-28.
Richey J.E., et al. 1978. Carbon flow  in four lake ecosystems:
  A structural approach.  Science 202:1183-6.
Romanenko, V.I. 1978. Balance of organic matter in the eco-
  system of the Rybinsk Reservoir. Pages 121-131 in Proc.
  First  and Second USA-USSR  Symp. Effects of Pollutants
  upon Aquatic Ecosystems. Vol. 1 EPA 600/3-78-076. U.S.
  Environ.  Prot. Agency, Washington, D.C.
Sarvala, J., V.  llmavirta, L. Paasivirta, and K. Salonen. 1981.
  The ecosystem of the oligotrophic Lake Paajarvi: Secon-
  dary  production and an ecological energy budget of the
   lake. Verh. Int. Verein. Limnol. 21:454-9.
Serruya, C., M. Gophen, and U. Pollinger.  1980. Lake Kinneret:
  Carbon flow patterns and ecosystem  management. Arch.
   Hydrobiol. 88:265-302.
Soballe, D.M.  Predicting  trophic status in rapidly  flushed
   lakes: The  role of riverine phytoplankton transport  and
   biological response time, (in prep.).
Thornton,  K.W., et al. 1981. Reservoir sedimentation  and
   water quality—an heuristic model. Pages 654-61  in H.G.
   Stefan, ed. Proc. Symp. on Surface Water Impoundments.
   Am. Soc. Civil Eng., New York.
Turner, R.R., E.A. Laws, and R.C. Harris.  1983. Nutrient reten-
   tion and  transformation in relation to hydraulic flushing
   rate in a small impoundment. Freshw. Biol. 13:113-27.
 Vollenweider,  R.A. 1976. Advances in defining critical loading
   levels for phosphorus in lake eutrophication. Mem. 1st. Ital.
   Idrobiol. 33:53-83.
 Wetzel, R.G. 1975. Limnology. W.B.  Saunders, Philadelphia.
 Wetzel,  R.G.,  P.H. Rich,  M.C. Miller,  and H.L Allen. 1972.
   Metabolism of dissolved and paniculate detrital carbon  in
   a temperate hardwater lake. Mem. 1st. Ital.  Idrobiol.
   29:185-244.
 Wissmar, R.C., J.E.  Richey, and D.E. Spyridakis. 1977. The
   importance of allochthonous paniculate carbon pathways
   in a subalpine lake. J.  Fish. Res.  Board Can. 34:1410-18.
                                                      285

-------
 MIXING  EVENTS IN EAU GALLE LAKE
 ROBERT F. GAUGUSH
 U.S. Army Corps of  Engineers
 Waterways  Experiment Station
 Vicksburg, Mississippi
             ABSTRACT

             Eau Galle Lake (Spring Valley, Wis.), a U S. Army Engineers reservoir, is more susceptible to
             weather-related mixing events than most north temperate lakes. Bottom withdrawal at Eau Galle
             leads to considerable heat storage in the hypolimnion and by late summer there is only 6-8° C
             difference between surface and bottom temperatures. Steep metalimnetic gradients  never
             develop and as a result the reservoir has a relatively low thermal stability, which implies a
             susceptibility to weather-induced (i.e., wind or cold front passage) mixing events. Two types of
             mixing events were observed in Eau Galle _ake in the summers of 1981 and 1982: (1) Small scale
             mixes similar to those observed in lakes. These events lead to net movement of nutrients into
             the epilimnion and a resultant increase ir chlorophyll a concentrations. (2) Large scale mixes
             that function essentially as short-lived turnovers. These large scale mixes are preceded by a
             cooling trend in air temperature which results in heat loss from the surface of the reservoir. Sur-
             face  cooling reduces  the already small temperature differences between surface and bottom
             and sufficient wind can produce considerable mixing. These events lead to  the introduction of
             oxygen into the previously anoxic hypolimnion and rather than increasing epilimnetic concentra-
             tions of nutrients, there is a general loss throughout the water column. Low thermal stability and
             susceptibility to mixing may be a common feature of bottom withdrawal reservoirs and may ex-
             ert a considerable influence on nutrient and phytoplankton dynamics.
INTRODUCTION

Weather related mixing events have been shown 1o
act as an important mechanism for epilimnetic inter-
nal nutrient loading in lakes during the summer wh€'n
external  loadings  can be  expected to  be minimail.
Stauffer and Lee (1973) demonstrated that cold front
passage and  wind stress resulted in  thermocline
migrations in  Lake Mendota. These migrations  in-
creased epilimnetic nutrient concentrations and were
followed by  increased  chlorophyll  concentrations.
Stefan and Hanson (1981) observed significant phos-
phorus transport from the anoxic hypolimnion to the
epilimnion  associated with mixing  in  five  shallow
lakes in south-central Minnesota. Phosphorus trans-
port was followed by intense  algal blooms in these
lakes. Kortmann et al. (1982) reported the occurrence
of  algal blooms  in  response to  the  thermocline
descending  below the anaerobic  interface  in Lake
Waramaug.
  The influence of weather-induced mixing events  n
reservoirs has not  been considered. A comparison of
309  natural  lakes and  107  U.S.  Army Corps  of
Engineers reservoirs included in the 1972-75 U.S. En-
vironmental  Protection Agency National Eutrophica-
tion Survey indicated that reservoirs are generally
larger, deeper, morphologically more complex, and
had shorter hydraulic residence times  than  natural
lakes (Thornton et al. 1982). These differences coupled
with the importance of advective and unidirectional
transport in reservoirs (Baxter, 1977) and  the presence
of either selective or bottom withdrawal may alter a
reservoir's thermal regime in such a way  as to make it
more susceptible to mixing events.
  The  primary objective  of this  paper is to  demon-
strate the occurrence and extent of mixing events ard
their effect on epilimnetic nutrient loading and phyto-
plankton abundance in Eau Galle Lake, a Corps reser-
voir. A secondary  objective is to determine if  the
design and operation of  this reservoir influence its
susceptibility to mixing events. Eau Galle Lake is well
suited for this determination because it is an atypical
Corps reservoir in that it is morphologically similar to
natural lakes but does have a low  level gated release
which may alter its thermal regime.
SITE DESCRIPTION

Eau Galle Lake, created in September  1968 by im-
poundment of the Eau Galle River, is located in west-
central Wisconsin approximately 80 km east of Minn-
eapolis-St. Paul,  Minn. The reservoir's primary pur-
pose is to provide flood control. Normal releases are
through an uncontrolled morning glory and a low-level
gated outlet conduit located at normal pool elevation
and 5  m  below normal  pool elevation, respectively.
Eau Galle is a small, shallow reservoir (Table 1). Pool
elevation  rarely fluctuates more than 0.3 m above its
normal 286.5 m. The regularity of the 4 km shoreline is

  Table 1.—Physical characteristics for Eau Galle Lake.
Elevation
Surface area
Volume
Maximum depth
Mean depth
Length
Shoreline  length
Shoreline  development ratio
Drainage area
Residence time
286.5 m msl
  0.6 km2
  1.9 x 106 m3
  9 m
  3.2m
  1 km
  4km
  1.5
166 km2
  0.07 yr
                                                 286

-------
                                                                     COMPARATIVE ANALYSIS OF RESERVOIRS
indicated by the shoreline development ratio of 1.5.
Land uses are primarily agricultural (crops and dairy
pasture). Eau Galle is dimictic and characterized by
high nutrient  concentrations, hypolimnetic anoxia,
periodically intense algal blooms, and the develop-
ment of macrophytes in the littoral areas. Fall turnover
occurs  in late September or early October and ice
cover usually begins in December and persists until
late March. Mean  annual values for selected water
quality variables are presented in Table 2.


METHODS

Routine sampling  for nutrients,  metals, and  chloro-
phyll a was conducted on a biweekly basis throughout
the study period (1981 and 1982).  Water samples were
collected from the  surface to the bottom at  1 m inter-
vals using a pump and hose sampler. Dissolved ox-
ygen and temperature were  measured in situ weekly
with a  Hydrolab surveyor (Hydrolab Corp.,  Austin,
Tex.). Details concerning sampling techniques and the
specific methods used  for determining phosphorus,
nitrogen, chlorophyll a,  iron, and manganese can be
found  in Johnson  and Lauer (in  prep.). Six stations
were sampled at each visit but only the data derived
from the deepest station will be  presented here.
   Meteorologic data was obtained from the nearest
first-order weather station located at Minneapolis-St.
Paul (Natl. Weather Serv.). Discharge records for Eau
Galle were obtained from the U.S. Geological Survey.
   Thermal stability per unit lake surface area is the
amount of work required to  mix  the entire volume of
the lake to a uniform temperature (Birge, 1915). Stabili-
ty (S, g.cm cm -2) was calculated  from the integral
given by Hutchinson (1957):
                  Zm
       S = Ao-1  / -«z -  zg)Az(1 - Pz))dz
                0

where z = depth,  m
     zm = maximum depth
      A0 = lake surface area, m2
      Az = area enclosed by contour of depth z
      pz = density of water at depth*z

and the center of gravity (zg) of the lake is
                 Zm
zg =
                    zAzdz
 where V = lake volume, m3

 Volume-weighted mean lake temperature is defined as
                 Zm
       TL = V-1 /  TzVzdz
                0
                                             cent strata relative to the density difference between
                                             water at 5 and 4°C:

                                                    RTRMZ = (pz - pz + i)/(p5»c  - P4°c)
                                              RESULTS AND DISCUSSION

                                              The development of Eau Galle's heat content general-
                                              ly follows the typical pattern for north temperate lakes
                                              (Fig. 1, upper). Heat content is at a minimum when the
                                              lake is ice-covered and begins to increase with ice-out,
                                              reaching a maximum sometime in July. With the onset
                                              of cooler weather the lake  begins  to lose heat and
                                              reaches minimum heat content  in early  December.
                                              Development of stability (Fig. 1, lower) is  highly cor-
                                              related with the changes in heat content (r  = 0.86, p V
                                              .001). Eau Galle begins to stabilize in mid-April, attains
                                              maximum stability at or a  few weeks prior to max-
                                              imum heat content,  and reaches a minimum in mid-
                                              October.
                                                The months  of June, July, and August of both 1981
                                              and 1982 were marked by  a series of decreases  in
                                              stability;  the  large  destabilizations are  associated
                                              with  considerable heat loss.Significant heat  losses
                                              and decreased stability might be  expected in late
                                              August, but the months of  June  and July should be
                                              characterized  by a steady  increase in  heat  content
                                              and stability.
                                                Eau Galle has  two features that may make it more
                                              susceptible to these summer destabilizations. First,
                                              its hypsographic, or depth area, curve (Fig. 2a)  il-
                                              lustrates  that most of the lake's area (and volume) lies
                                              above the 4 m contour. The predominance of depths
                                              less than 4 m results in a very shallow center of gravity

                                                    	                ICE COVER
                                                   J  F M A M J  J  ASONDJ  F  M A M J J  ASONO
                                                                        ICE COVER
                                                   J  F M A M J  J  A
                                                            1981
                                                                   O N D J F
M A M J J
      1982
                                                                                      A S  O  N D
                                              Figure 1.—Heat content (upper) and thermal stability (lower)
                                              of Eau Galle Lake, 1981 and 1982.
 where Tz = temperature at depth z
       Vz = stratum volume at depth z

 Other symbols are as given. Lake heat content, the
 store of heat that it could  impart to  its surroundings
 on cooling to 0°C, is defined as

       HL  = cTLV

 where c = specific heat of water, 103 kcal deg - 1 m - 3
 and other symbols are as given.  Relative thermal
 resistance to mixing (Vallentyne, 1957) was obtained
 by converting water  temperature  to  density  and
 calculating the density difference between two adja-
                                                  Table 2.—Selected water quality characteristics of
                                                                Eau Galle Lake*.
                                              Total alkalinity
                                              Total carbon
                                              Total inorganic carbon
                                              Total organic carbon
                                              Total phosphorus
                                              Total nitrogen
                                              Chorophylla(^g l~1)
             152.3
             47.4
             35.2
              7.4
              0.091
              1.65
             31.9
                                               •Mean surface values (0—2m) from six lake stations for years 1978-80
                                               Units are mg I ~ unless otherwise noted
                                                  287

-------
 LAKE AND RESERVOIR MANAGEMENT
 (zg = 2.7 m). Given otherwise equal conditions, stab li-
 ty will increase with the descent of the thermocline,
 reaching a maximum value when the thermocline lies
 at the center of gravity. As the thermocline descends
 past this depth stability decreases (Rutner, 1963). The
 shallow center of gravity may make Eau  Galle more
 susceptible to mixing.
   Second, a considerable fraction of the discharge
 from Eau Galle  is through the low-level release gate
 located  5  m  below the surface (Fig. 2b). Low-level
 releases are essentially constant at approximately 3
 x 104m3day-i but outflow from the surface changes
 with pool elevation. Snowmelt and spring runoff rai:;e
 pool elevation and reduce the relative contribution of
 the low-level release. During the stratified  period (late
 April through September) the low-level release ac-
 counts for 50 percent or more of the discharge. The
 output of relatively colder water from below the sur-
 face can result in considerable hypolimnetic heating
 as outflowing water  is replaced with  warmer water
 from above. Hypolimnetic heating is evident from the
 progression of isotherms below 6 meters (Fig. 3), and
 acts to reduce the density differences from surface lo
 bottom which, in turn, reduces stability.
  The combination of a shallow center of  gravity and
 considerable hypolimnetic heating make  Eau  Galle
 particularly susceptible to destabilization during rs
 stratified period. Figure 4 presents the  meteorologic
 conditions associated with changes in stability and
 mean lake temperature. Meteorologic variables have
 been smoothed,  using  a  3-day moving averagB.
 Periods of destabilization  are generally accompanied
 by decreases in  both maximum  and  minimum air
         APR    MAY   JUN    JUL    AUG    SEP
                           1981
         APR    MAY    JUN    JUL    AUG    SEP
                           1982

   Figure 3.—Isotherms (°C) for April through September, 1981
   (upper) and 1982 (lower).
              CUMULATIVE AREA, PERCENT
               20     40     60     80
    £
    LLJ
    Q
                                           100
      10 I-
                                                     100
   75
                                                      50
                                                      25
                                                   UJ
                                                   O
                                                   cc
                                                          J   FMAMJ   JASON   D
                                                                            1981
ai
O

-------
                                                                     COMPARATIVE ANALYSIS OF RESERVOIRS
temperature,  increasing barometric pressure,  winds
out of the northwest, and precipitation.
  West-central  Wisconsin  is  affected by  distur-
bances, originating in the northwest, which migrate
eastward and are followed by colder polar air masses
(Natl. Weather Sen/.). As  such a cold front passes
barometric  pressure increases,  air  temeperature
drops abruptly,  wind velocity increases, and  winds
shift from the southwest to the northwest (Blair and
Fite, 1957). Surface heat loss during  the clear, cool
nights that follow the passage  of  a cold  front will
result in convection currents that can act to a depth of
3 meters  (Wetzel, 1975).  Once convection currents
have  decreased  stability, winds  can  mix a con-
siderable fraction of the lake's volume.
  In Eau Galle, decreases in stability occur in either of
two ways: (1) with a  large decrease in TL (significant
heat loss), or (2) with little or no change in TL (a re-
distribution of heat). It will be shown that  these  dif-
ferent  types  of  destabilization have  profoundly  dif-
ferent effects on epilimnetic nutrient concentrations
and phytoplankton abundance.
  Although  decreases  in  stability indicate mixing,
stability, as defined, expresses the condition  of  the
lake as a whole and cannot be used to determine the
extent of a mixing event. Relative thermal resistance
to mixing (RTRM), which is calculated for each depth
stratum, can indicate the  depth to which mixing has
occurred. A temperature change of 1°C m~1  in  the
range of 10 to 30°C produces RTRM values ranging
from 10 to 37. A RTRM value of 30 was found to most
closely approximate the position of the thermocline
(Kortmann  et  al.  1982)  and  that  convention was
adopted for Eau Galle.
  Thermocline  depth  (RTRM   =  30)  is  extremely
dynamic in Eau Galle during its stratified period (Fig.
5, upper) and the extent of hypolimnetic anoxia (Fig. 5,
lower) closely follows changes in thermocline depth.
Hypolimnetic  anoxia  increases in  extent  during
periods when the thermocline lies high  in the water
column. Large-scale mixing events, characterized by a
large, rapid descent of the thermocline, introduce ox-
ygen into  the hypolimnion and reduce the extent of
anoxic conditions. Note that  when anoxia is at a max-
imum (peaks in  July and August) the epilimnion is
separated from the hypolimnion by no  more than 1
meter. Small-scale mixing events during  maximum
anoxia, exhibiting only a slight descent of the thermo-
cline and having no effect on the extent of anoxia, can
be expected  to  easily transport  nutrients from the
hypolimnion and result in an increase in phytoplank-
ton abundance.
  Phytoplankton dynamics (expressed by changes in
chlorophyll a concentrations) are clearly influenced by
changes in thermocline depth and its proximity to the
anoxic hypolimnion (Fig. 6).  Both years exhibit a spr-
ing bloom prior to the onset of stratification and hypo-
limnetic anoxia. Later blooms occur when anoxia is at
a maximum and the nutrient-rich hypolimnion lies
relatively  close  to the surface. These blooms are
separated by large-scale mixing events that reduce
surface concentrations of chlorophyll a. Decreases in
chlorophyll a concentrations in response  to a large-
scale mix  result from two factors: (1) an initial dilution
Figure 4.—Meteorologic conditions associated with changes in stability (AS,	) and mean lake temperature (ATL, —) for
April through September, 1981 (left) and 1982 (right). Shaded areas correspond to periods of decreasing stability.
                                                  289

-------
 LAKE AND RESERVOIR MANAGEMENT
 of epilimnetic concentrations caused by the expan-
 sion of the epilimnion; and (2) if mixing continues and
 the mixed layer extends  below the  photic zone, the
 phytoplankton will spend a greater amount of time
 under poor light conditions, thereby reducing growth
 rates (Stefan et al. 1976).
   The stratified period for Eau Galle  can be separated
 into a series of three  different events  based  on
 changes  in stability, heat content, and the extent of
 anoxia. A stable period can be identified by increases
 in stability, heat content, and anoxia. Mixing periods
 are characterized by losses in stability and heat con-
 tent. Large and  small scale mixes  can be differen-
 tiated  based  on the  large loss of  heat  and  tne
 decrease in the extent of anoxia associated with large
 scale mixes. Representative changes associated with
 these three different events are presented in Table 3.
   Stable  periods are typified by surface losses of both
 total phosphorus and nitrogen  while greater deptis
 exhibit considerable gains. Epilimnetic losses result
 primarily  from particulate matter settling out during
 these  relatively   calm periods.  Concentrations  of
 dissolved metals  generally show gains as would ae
 expected given the  increased extent of hypolimnelic
 anoxia. Surface concentrations of chlorophyll a either
 show no  change or decrease during  stable periods
   Significant  epilimnetic loading of nitrogen  and
 phosphorus occur during small-scale mixes. Internal
 loading rates during these mixes are one to two orders
 of magnitude  higher than  average  external loading
 rates. External loading during the late spring and sum-
 mer averages about 0.02 mg P m - 2 day -1 and 0.20 mg
 N m-2 day-1  (Montgomery, in prep.). Net positive n-
 ternal loading of nutrients to the epilimnion results in
       APR    MAY    JUN    JUL     AUG    SEP
                        1981
      APR   MAY    JUN    JUL    AUG    SEP
                        1982
Figure 6.—Isopleths of chlorophyll  a concentrations (Mg
m-3), April through September 1981 (upper) and 1982 (lower).
 t
        APR    MAY    JUN     JUL    AUG    SIEP
                                                      6 -
                                                         APR    MAY    JUN    JUL    AUG    SEP
        APR    MAY    JUN    JUL    AUG    SIEP
                          1981
   APR    MAY    JUN     JUL    AUG    SEP
                       1982
Figure 5.—Isopleths of RTRM (upper) and dissolved oxygen (Mg I - \ lower) for April through September, 1981 (left) and 1982
(right).
                                                  290

-------
                                                                         COMPARATIVE ANALYSIS OF RESERVOIRS
              Table 3.—Changes associated with stable periods and large- and small-scale mixing events1.
              Year
                                                                             Mixing Events
                                           Stable
                                                                 Small-Scale
                                        Large-Scale
                                      1981
1982
                                                               1981
                        1982
                                                                                       1981
 'Units are mg m ~ 2 day ~ 1 unless otherwise noted
 Calculations based on entire lake (0-8 m)
 'Calculations based on (0-3 m)
                                                 1982
Time period (month/day)
A Stability (g • cm cm -2)
A Heat content (x1Q9 kcal)
A Anoxic volume ( x 104 m3)
Total phosphorus
0 - 3 m
4-8m
Total nitrogen
0- 3 m
4-8m
'Dissolved iron
dissolved maganese
'Chlorophyll a (mg m-3 day-1)
5/19-6/2
14.5
4.3
1.1

-0.2
3.5

-6.0
4.7
0.0
17.9
0.0
6/15-6/29
23.0
3.3
1.4

-1.0
33.5

-40.8
19.8
17.5
2.3
-1.3
6/30-7/14
-18.0
-0.1
0.0

10.3
11.3

37.7
24.1
50.3
5.6
7.6
7/13-7/27
-5.7
-0.7
0.0

8.9
3.5

267.8
74.8
24.2
10.6
5.6
7/14-7/28
-22.1
-4.5
-10.4

-0.2
-6.3

-13.2
-8.7
-33.0
-22.3
-4.0
7/27-8/10
-27.4
-4.4
-5.3

-2.6
0.2

-177.3
-35.6
9.3
-10.1
-0.8
increased chlorophyll a concentrations during small-
scale mixes.  These  mixes have little effect on  the
hypolimnion given that concentrations of nitrogen,
phosphorus, and dissolved metals exhibit gains dur-
ing these events.
   Large-scale mixes act as short-lived turnovers and
result in epilimnetic losses of phosphorus, nitrogen,
and chlorophyll. The effects of the introduction of ox-
ygen associated with these mixes are evident in the
general loss of phosphorus,  nitrogen, and dissolved
metals  from the hypolimnion. Although  phosphorus
and dissolved iron showed positive changes during a
large-scale  mix in 1982, the rate of change is  con-
siderably lower than those  observed  during small-
scale mixes.
CONCLUSIONS

Two different types of mixing events can be identified
in Eau Galle Lake. First, small-scale mixes occur when
the epilimnion and the anoxic, nutrient-rich hypolim-
nion lie relatively close together. The effect of these
mixes is  similar to those reported for natural lakes
with an internal loading of nutrients to the epilimnion
increasing phytoplankton abundance. Second, large-
scale mixes occur that function essentially as short-
lived turnovers. With the introduction of oxygen into
the  previously  anoxic  hypolimnion,  nutrient and
dissolved metal  concentrations decrease  in the lake
as a whole. These  large-scale mixes have not been
reported for natural lakes and may be a result of lower
stability in Eau Galle Lake caused by the effect of low-
level releases on its thermal regime.

ACKNOWLEDGEMENTS: The author wishes  to thank the
members of the Eau Galle Reservoir Laboratory (J. Carroll, S.
Ashby, G.  Lauer, D. Johnson, and W. James) for sample col-
lection and chemical analyses. This research was sponsored
by the  Office,  Chief of Engineers,  U.S. Army Corps of
Engineers  as part of the Environmental Water Quality and
Operations Studies (EWQOS), Work Unit VIIA.
      REFERENCES

      Baxter, R.M. 1977.  Environmental effects of dams and im-
        poundments. Ann. Rev. Ecol. Syst. 8:255-83.
      Birge,  E.A. 1915. The heat  budgets of American and Euro-
        pean lakes. Trans. Wis. Acad. Sci. Arts Lett. 18:166-213.
      Blair, T., and R. File. 1957. Weather Elements. Prentice-Hall,
        Inc.  Englewood Cliffs, N.J.
      Hutchinson, G.E. 1957. A Treatise on Limnology. Vol. I. Geo-
        graphy, physics and chemistry. John Wiley, Inc. New York.
      Johnson,  D., and G. Lauer. (In prep.) General methods. In.
        R.H.  Kennedy, ed.  Limnological  Investigations of Eau
        Galle Reservoir,  Wis. Vol.  I. Tech. Rep. U.S. Army Eng.
        Waterways Exp. Sta. Vicksburg, Miss
      Kortmann, R.W., D.D.  Henry, A.  Kuether and S. Kaufman.
        1982. Epilimnetic nutrient loading by metalimnetic erosion
        and resultant  algal responses in Lake Waramaug, Conn.
        Hydrobiologia 92:501-10.

      Montgomery, R. (In prep.). Material loadings.  In R.H. Ken-
        nedy, ed. Limnological Investigations of Eau Galle Reser-
        voir, Wis. Vol. II. Tech.  Rep. U.S. Army Eng. Waterways
        Exp. Sta., Vicksburg, Miss.
      Rutner, F. 1963.  Fundamentals of Limnology. Univ. Toronto
        Press, Toronto.
      Stauffer,  R.E., and G.F. Lee. 1973. The role of  thermocline
        migration in regulating algal blooms.  Pages 73-82 in E.J.
        Middlebrooks, D.H. Falkenborg, and  T.E. Maloney, eds.
        Modeling the Eutrophication Process. Ann  Arbor  Sci.
        Publishers, Ann Arbor, Mich.
      Stefan, H., and M.J. Hanson, 1981. Phosphorus recycling in
        five shallow lakes. J. Environ. Eng. Div. Am. Soc. Civil Eng.
        107:713-30.
      Stefan, H., T. Skoglund, and R.O. Megard. 1976. Wind control
        of algae growth in eutrophic  lakes. J. Environ. Eng. Div.
        Am. Soc. Civil Eng. 102:1201-13.
      Thornton, K.W.,  R.H. Kennedy, A.D. Magoun, and G.E. Saul.
        1982. Reservoir  water quality sampling design.  Water
        Resour. Bull. 18:471-80.
      Vallentyne, J.R.  1957. Principles of  modern limnology. Am.
        Sci. 45:218-44.
      Wetzel, R.G. 1975. Limnology.  W.B.  Saunders Co., Phila-
        delphia.
                                                     291

-------
 EMPIRICAL PREDICTION  OF CHLOROPHYLL  IN  RESERVOIRS
 WILLIAM W. WALKER, JR.
 Environmental Engineer
 Concord, Massachusetts
             ABSTRACT

             The use of nutrient loading models for predicting the trophic status of lakes and reservoirs is
             based partially  upon empirical phosphorus/chlorophyll relationships which  were originally
             developed using data from northern natural lakes. Recently, increased attention has been given
             to the effects of other potentially limiting factors, including nitrogen, light, and flushing rate, on
             the performance of empirical chlorophyll models. This paper describes a study of these relation-
             ships in reservoirs and is derived from a nationwide research project conducted for the U.S. Ar-
             my Corps of Engineers. Effects of N/P ratios, flushing rate, turbidity, and impoundment mor-
             phometry on phosphorus/chlorophyll relationships  are systematically evaluated. Apparent
             lake/reservoir differences in average chlorophyll response to phosphorus are related more to the
             limited generality of phosphorus/chlorophyll regression models and to regional factors than to
             effects of impoundment type. The analysis has led to the development of a more general model
             which explicitly accounts for effects of algal growth limitation by phosphorus, nitrogen, light
             and flushing rate.
 Early attempts at empirical eutrophication modeling
 were based upon data from northern, natural lakes
 (Vollenwelder, 1968). Model structures have evolved
 considerably since  that time as larger data sets in-
 cluding both lakes and reservoirs have become avail-
 able. The basis of most existing models is that chloro-
 phyll a, the most practical and commonly employed
 measure of algal standing crop, is directly related to
 impoundment total phosphorus concentration, which,
 in turn, is related to external total phosphorus loading,
 mean depth, and hydraulic residence time. Given the
 origins of the modeling concept and the large array of
 published  models,  a  systematic analysis of model
 strengths and weakness is needed to provide guid-
 ance for applications to reservoirs.
   This paper presents empirical evidence that sug-
 gests that the generality of simple phosphorus/chloro-
 phyll relationships in reservoirs is rather low because
 of  systematic  effects  of  other factors, including
 nitrogen, turbidity, depth, and flushing rate.  Implica-
 tions of the results for assessing lake/reservoir d f-
 ferences in eutrophication response are also discuss-
 ed. The work is derived from  a series  of research
 reports prepared for the U.S. Army Corps of Engineers
 and aimed at the development and testing of simpli-
 fied predictive techniques for reservoir water quality
 (Walker, 1981, 1982b, 1983). The reports describe dala
 base development,  preliminary  model testing, ard
 model refinements, respectively.
   Figure 1  depicts the structure of a model network
that has  been  developed  for predicting reservoir-
average  conditions  (Walker,  1983).  Methods  for
simulating   spatial  gradients  in  phosphorus  and
related  trophic  state  indicators  have  also been
developed.  Nutrient  retention formulations have been
modified to account  for effects of  nonlinear sedi-
mentation kinetics, seasonal variations in inflow con-
ditions,  and inflow nutrient partitioning (ortho-P ver-
sus nonortho-P and  inorganic N versus organic N) en
nutrient mass balances. This paper is concerned with
the submodel which predicts mean chlorophyll a as a
function of pool  nutrients, nonalgal turbidity,  mixed-
layer depth, and  summer flushing rate. This is a con-
 siderable extension  beyond a  simple phosphorus/
 chlorophyll  regression  but provides advantages  in
 terms of generality and accuracy without substantial-
 ly increasing data requirements. The following analy-
 sis demonstrates the need for a  more complex model
 by showing that factors other than total phosphorus
 influence chlorophyll production in many reservoirs.
   A series  of  data sets  have  been assembled for
 testing empirical models in Corps reservoirs (Walker
 1981,  1982b, 1983). The approach  has been to build
 these  data  sets starting at the  most basic level, in-
 dividual water quality and hydrologic measurements.
 This permits using uniform data-reduction procedures
 in  developing  the  summary statistics required for
 model  testing,  including  nutrient  loadings  and
 average water  quality conditions. Uniform  screening
 criteria based upon sampling program design, sampl-
 ing frequency,  and statistical variability in the sum-
 mary values have also been applied to select those im-
 poundments with the best  information, within the con-
 straints of existing  data sources. Data from 65 Corps
 reservoirs are  analyzed, derived primarily from the
 EPA's National Eutrophication Survey (U.S. Environ.
 Prot. Agency, 1978). Reservoir water quality conditions
 are summarized by mean  concentrations  measured
 between April  and  October  at  depths less than 5
 meters and weighted across stations  based  upon
 relative surface area. Each reservoir was sampled at
 least three times between April and October.
  Within  certain ranges  of  nonalgal  turbidity, in-
organic N/P ratio, and flushing rate, chlorophyll a is
roughly proportional  to  total phosphorus  with  an
average proportionality  constant of  about .28, as
shown in  Figure 2.  A similar result (geometric mean
ratio  = .24) was obtained in the OECD Synthesis
Study (Organ. Econ. Coop. Dev., 1982) for annual-mean
values in 99 lakes and reservoirs with inorganic N/P
ratios greater than 10 and excluding impoundments in
which  algal  biomass was  "suppressed artificially or
by natural turbidily." Based upon  analysis of EPA Na-
tional Eutrophication Survey data from 757 lakes and
reservoirs, Hern et al. (1981) and  Lambou et al. (1982)
found  that the  chlorophyll a/total  P  response  ratio
                                                 292

-------
                                                                    COMPARATIVE ANALYSIS OF RESERVOIRS
 varied from .001 to 2.81, with average values of .24, .29,
 and .24 for spring, summer, and fall samples, respec-
 tively. An average response ratio of .5 is typical of nor-
 thern  lakes in Minnesota (Organ.  Econ. Coop. Dev.,
 1982) and Vermont (Vt. Dep. Water Resour., 1980).
   Figure 3 plots the chlorophyll/total P ratio in Corps
 reservoirs  against  various impoundment character-
 istics. The horizontal dashed lines are located at the
 average response ratio of .28 (or - .55 on a log scale).
 Deviations from the average  ratio are  apparent  at
 variable extremes. The vertical lines in the inorganic
 N/P, summer residence time, and nonalgal turbidity
 plots depict the approximate regions in which factors
 other than total phosphorus may influence chlorophyll
 response, based upon consistent deviations from the
 mean  ratio. As indicated in Figure 3, apparent devia-
 tions at low total N/P ratios and high total phosphorus
 concentrations are explained  by variations  in the
 other three factors listed.
   The strongest relationship is observed  in the case
 of nonalgal turbidity (inverse Secchi depth, corrected
 for light attenuation by chlorophyll a and chlorophyll a
 related substances). Computed in this way, the term
 turbidity is used loosely here because it would also in-
 clude  effects  of color,  which  may be important  in
 some  impoundments, and effects  of variations in the
 algal light attenuation coefficient. Lambou et al. (1982)
 also found that the response ratio  was  related to non-
 algal light  extinction, in this case computed as the
 residual  from Carlson's (1977) chlorophyll/transparen-
 cy regression for northern lakes.
  The shape of the turbidity plot suggests that there
 is no  absolute cutpoint  below which the response
 ratio is independent of turbidity, particularly when the
 ratio is plotted against the product of mean  mixed
 layer depth (volume/surface  area,  mid-summer) and
                                                    turbidity, which is inversely related to the effect of tur-
                                                    bidity on depth-averaged light intensity. The strongest
                                                    deviations occur, however, in impoundments with tur-
                                                    bidities exceeding about .9 l/m. The turbidity plot  is
                                                    also influenced to some extent by the propagation of
                                                    data errors in chlorophyll, since measured chlorophyll
                                                    values are used to compute both the response ratio
                                                    and turbidity. This is  a minor problem, however, be-
     1.8"


     1.5-



1  *'*
  •k
 "f  0.9'

 8
 -  0.6-


     0.3-


     0.0-'
                                                                                          -.55 + x
                                                                                          7A  (o)
                                                                                        5.
                                                                                      SE -.037
                                                          0.5  0.8  1.1   1.4  1.7   2.0  2.3   2.6

                                                                    LOG  [  TOTAL P, MG/M3 ]


                                                    Figure 2.—Chlorophyll a versus total phosphorus. Footnote.
                                                    (•) inorganic N/P< 7, non-algal turbidity > 9 l/m or summer
                                                    hydraulic residence time < .04 years; (o) other reservoirs.
                            MODEL
                                            NETWORK
Inflow Total  P
Inflow Ortho  P
Mean Depth
Hyd. Res.  Time
Inflow Total N
Inflow Inorg N
Mean Depth
Hyd. Res.  Time
                                                                        Hypol. Oxygen
                                                                        Depletion Rate
Mixed Layer Depth
Summer  Res. Time
Region
Latitude
Total P
Summer  Res.  Time
Mean Depth


Figure 1.—Empirical model network.
                                            Composite Nutrient
                                            Concentration
                              Non-Algal
                              Turbidity
                                                         7>—»Particulate P
                                                293

-------
LAKE AND RESERVOIR MANAGEMENT
cause chlorophyll and turbidity are weakly correlated
(r = .22)  and  because  at high  nonalgal turbidities
(where the apparent  effects are strongest),  the com-
putation of turbidity is controlled primarily by the Sec-
chi measurement and is highly insensitive to chlorc-
phyll.
   Nonalgal   turbidity   may   influence  chlorophyll
response through mechanisms related to light limita-
tion or nutrient availability. The former is likely to be
more important in these reservoirs, based upon (1) the
observation that reservoirs with high turbidity levels
also tend to  have relatively  high concentrations of
ortho-phosphorus  and  inorganic nitrogen  (Walker,
1983) and (2) the relationship between response  ratio
and the product of mixed layer  depth  and turbidity
(Fig. 3).  Preliminary testing of models for predicting
station-mean  chlorophyll concentrations from station-
mean  phosphorus values also indicated that, at N/P
ratios exceeding 8,  residuals were  negatively cor-
related (r= - .67) with the product of nonalgal turbidi-
ty and station total  depth; similarly, residuals from
models predicting reservoir-mean chlorophyll a  based
upon  normalized  phosphorus  loadings  were also
negatively correlated (r= -.67 to -.81) with the pro-
duct of turbidity and mean depth (Walker, 1982b). The
apparent  importance  of the depth-turbidity product
suggests  a light limitation mechanism, since the pro-
duct is related to depth-averaged light  intensity and
the depth  term would not be expected to be important
if the mechanism were related to nutrient availability.
In the network (Fig.  1), effects of inflow  nutrient
availability are considered in  the  nutrient  retention
submodels.
  At low values of the turbidity-mixed depth product,
the response ratio  approaches .5 (-.3 on log scale),
which, as  discussed previously, is typical of northern
lakes in Minnesota  and Vermont. While color is impor-
tant in some cases, northern lakes would be expected
0.0-

-0.3-
-0.6-
-0.9-
-1.2-
-1.5-

o
N/P - 7J-I ° 00

o a o J of ° o B •. „ 5 o> a . ... o o « • o '. • 0 . . • • 0.0- -0.3 -0.6' -0.9- -1.2- -1.5- o fo ° ° 0 ° °

0 ",.0 O O. . . 0 0 °. 0 « o • . * -1.0 -0.7 -0.4 -0.1 0.2 0.5 0.8 LOG [ NON-ALGAL TURBIDITY, 1/M ] -0.2 0.1 0.4 0.7 1.0 LOG [ TURBIDITY * MIXED DEPTH ] 1.3 Figure 3.—Chlorophyll a/Total P versus reservoir characteristics. Footnote: Horizontal line = average B/P ratio for P-limited impoundments, log 10 scales vertical lines = approximate outpoints for nitrogen, turbidity, and residence time effects on B/P ratio (•) Inorganic N/P < 7, nonalgal turbidity > .9 1/m or summer hydraulic residence time < .04 years (o) other reservoirs a = nonalgal turbidity (l/m) computed from: a = 1/S - .025 B, minimum = .08 l/m where, B = mean chlorophyll a (mg/m3), S = mean Secchi depth (m) 294


-------
                                                                     COMPARATIVE ANALYSIS OF RESERVOIRS
to have lower  nonalgal turbidity levels because of
geologic and  land-use  considerations. Analysis of
data from 20 Vermont lakes (Walker, 1982a) indicates
nonalgal turbidity levels ranging from 0. to .7 1/m, with
a median value of .1  1/m.  Applied  to mixed layer
depths ranging from 1.2 to  6 meters,  the turbidity-
depth product  ranges from  0 to  1.9, with a median
value of .4, or - .4 on a log scale. This value is below
the range  of  Corps reservoirs  shown in  Figure 3
(minimum -.15) and is consistent with the relatively
high response ratio observed for Vermont lakes.
  The dependencies noted in Figure 3 indicate that
some of the variance in the chlorophyll/total P ratio is
systematic and related to specific impoundment char-
acteristics. This suggests that the generality of the
model is limited and there is room for improvement.
Results provide an  approximate basis  for assessing
the applicability of the linear phosphorus/chlorophyll
model to reservoirs. Even after screening the data bas-
ed upon the criteria in Figure 3, appreciable variance
in the chlorophyll/total P  ratio remains (.037 on log
scales, which corresponds to a 90 percent confidence
factor of  2.43  for chlorophyll  predicted from total
phosphorus. Problems  remain with this approach to
determining model applicability based upon distinct
values of various factors:
  1. A  significant  percentage (42  percent)  of  the
Corps reservoirs is outside of the applicability range.
A model is lacking for these reservoirs.
  2. Distinct outpoints seem unrealistic and tend to
create  artificial classifications.  Actually,  effects of
these additional factors would be expected to vary
continuously over a range of values, but not necessari-
ly in a linear fashion.
  3. Potential problems would arise in model applica-
tions to reservoirs which are within but near the ap-
plicability margins. For example, the linear model
would  be  incapable of  simulating   responses  to
changes in  nutrient  levels involving  a change from
nitrogen to phosphorus limitation or vice versa.
  4. The apparent effects of  nonalgal turbidity appear
to be continuous and not completely  isolated by a
single outpoint. Systematic  deviations remain, par-
ticularly in relation to the product of mean mixed layer
depth  and  turbidity  (Fig.  3). Additional  analysis
(Walker, 1982b)  indicates that the slope of the chloro-
phyll a/total P regression increases (from 1.0 to about
1.45) when  more restrictive  turbidity  outpoints  are
used.
  These considerations have led to the development
of a more complex  model which explicitly considers
factors other than phosphorus (Fig.  1). Based upon
kinetic theories of algal growth, the  model provides
estimates of chlorophyll a which have two to threefold
less variance than a simple phosphorus/chlorophyll
regression when applied to Corps reservoirs and other
independent  data sets,  even  when  the data  are
restricted to P-limited impoundments, based upon the
criteria in Figure 3 (Walker, 1983).
  The question of whether lakes and  reservoirs are
different from an empirical modeling perspective can-
not be answered categorically because of the con-
tinuum of conditions both within and between  the
groups (Canfield and Bachman, 1981). In one respect,
spatial variations in water quality are likely to be im-
portant in a higher percentage of reservoirs because
of elongated morphometry and loading distributions.
Regional factors influencing  nonpoint source nutrient
loadings and inflow nutrient  partitioning are also im-
portant in making lake/reservoir comparisons because
the mid-latitudes of the United States contain a higher
percentage of reservoirs than the northern (glacial) or
southern (subtropical) latitudes (Walker, 1980).
  Table 1 presents a summary of lake and reservoir
data from the EPA National  Eutrophication Survey
Compendium  (U.S.  Environ.  Prot.  Agency,  1978).
Analyses of variance have been conducted on various
characteristics, with the data grouped  by  impound-
ment type (Corps reservoirs, non-Corps reservoirs, and
natural lakes). To eliminate highly eutrophic impound-
ments that are outside of the range of the Corps data
set previously analyzed, those with median total phos-
phorus  concentrations  exceeding  250  mg/m3  have
been excluded from the calculations. Significant dif-
ferences among means  are apparent for most of the
variables considered. Because of its association with
region, impoundment type is not necessarily the only
causal factor responsible for the among group dif-
ferences.
  Regardless of the causal factor, both the Corps and
the non-Corps reservoirs tend  to  differ  from  the
natural lakes  in many variables which are important to
empirical  chlorophyll models. Based  upon the  F
statistics, the groups differ most strongly with respect
to optical characteristics, including Secchi depth, tur-
bidity, the product of Secchi depth and chlorophyll a,
the ratio of Secchi depth to mean depth (roughly pro-
portional to depth-averaged light  intensity), and the
product of  nonalgal turbidity  and  mean depth.  The
highest F statistic (47.32) is observed for the product
of nonalgal turbidity and mean depth, which averages
4.57 in Corps reservoirs, 3.16 in non-Corps reservoirs,
and 1.48 in natural lakes. Based upon kinetic theories
of algal growth, the chlorophyll-Secchi product is pro-
portional to the fraction of light extinction attributed
to chlorophyll and to the areal photosynthetic  rate
under nutrient-saturated conditions  (Walker, 1982b,
1983). The mean values of the optical parameters and
residence time  suggest  that light limitation  and
flushing rate  are  potentially more important as con-
trolling factors for algal growth in reservoirs than in
lakes, on the  average.
  Conversely, the  inorganic and total N/P ratios sug-
gest that nitrogen limitation is somewhat more impor-
tant in lakes,  on the average. The nitrogen differences
may reflect regional factors, as well as criteria used in
selecting   impoundments  for  study  under  the
EPA/NES. The emphasis  during early years of the
survey was on systems affected by point sources;
these would  tend to be  nitrogen-limited.  The first
monitoring  year (1972) focused on northcentral  and
northeastern  States,  which contain relatively  high
percentages of natural lakes.
  Possibly as a result of differences in the average ef-
fects of algal growth limitation by nitrogen,  light, and
flushing rate, the three groups differ significantly with
respect to  the chlorophyll/total  P  ratio (F = 12.31).
When the data are screened  to permit a  focus on
primarily P-limited impoundments, (based  upon the
criteria in Figure  3, and assuming that the summer
residence time averages twice the annual value), the
mean chlorophyll/total P ratios tend to increase  and
differences among the groups become less signifi-
cant  (F = 3.21, p  < .04).  The group  means are not
significantly  different (F = 1.62, p<  .20) when im-
poundments  with  turbidity-mean   depth  products
greater than 5 are also excluded. Mean  depth is used
as a surrogate here in the absence of mixed depth in-
formation for  the NES data set. Figure 3 suggests that
effect of the turbidity-mixed depth product is more or
less continuous, however, and screening based upon
a single value would not be expected to completely
                                                 295

-------
 LAKE AND RESERVOIR MANAGEMENT
                       Table 1.—Lake/reservoir comparisons derived from NES Compendium.
  Variable
  Number
  Mean depth (m)
  Residence time (yrs)

  Overflow rate (m/yr)
  Total P (mg/m3)
  Chlorophyll a (mg/m3)
  Inorganic N/P
  Total N/P**
  Secchi depth  (m)
  Nonalgal turbidity (l/m)

  Chi a x Secchi (mg/m2)
  Secchi/mean depth

  Turbidity x mean depth
  Chi a/total P
  Number

  Percent of total
  Chi a/total P
  Number

  Percent of Total
  Chi a/total P
  Number

  Percent of total

  Chi a/total P
                                            Group Qeometric Means

                                               Reservoirs     Natural

                                             CE   Non-CE    Lakes
91
7.6.
.46
16.6
36
8.9
28.E.
21 .£>
1.1
.60
9.E
.14
4.57
277
6.3
.24
26.3
45
9.3
18.2
18.6
1.2
.50
10.7
.19
3.16
163
5.6
.72
7.8
35
10.2
13.8
25.1
1.7
.26
17.4
.30
1.48
                                               .24       .21

                                                Inorg. N/P > 7
3      223
91 %    81 %
                    .30
                 122
                 75%
                                               .25       .23       .34

                                   Inorg. N/P > 7, Turbidity < .9 1/m, T >.02 yrs
  5-
4      143
59%    52%


 .33      .30
                 110
                 67%


                   .35
                                   Inorg. N/P >7, Turbidity < .9 1/m, T> .02 yrs
                                          Turbidity x Mean Depth V 5
  4-
0      107
44%    39%
                                               .38
          .33
                 103
                 63%


                   .36
                                             Within-Group Mean
                                                      Square
                                               3.16*

                                              18.14*
                                              26.54*
                                               4.79*
                                               0.94
                                              12.66*
                                               5.07*
                                              17.79*

                                              36.76*
                                              25.69*

                                              36.94*
                                              47.32*


                                              12.31*
.145

.679
.535
.152
.158
.234
.107
.095

.145
.109

.095
.178


.102
                                              12.06*
                                                        .088
                                                                                             3.21*
                                                        .059
                                                                                             1.62
                                                                                                      .054
 Notes:
 based upon data from EPA National Eutrophication Survey Compendium with nutrient budget data, non-missing values for above variables, and median Total
 P< 250 mq/m3
 • F statistic for analysis of variance on log scales significant at p< 05

  *' based upon 88, 245, and 83 impoundments, respectively
eliminate its effects within each group. The suc-
cessive reductions (.102  to  .054) in the within-group
mean square reflect the increased applicability of the
chlorophyll/phosphorus model when nitrogen-limited,
turbid,   and  rapidly-flushed   impoundments  a-e
eliminated.   Under  P-limited  conditions,  response
ratios calculated from the EPA/NES data set tend "o
be somewhat higher than those calculated from the
Corps data set (Fig. 2 and 3), possibly as a result of d f-
ferences in  data reduction procedures; median total
phosphorus  concentrations  are  reported  in  the
EPA/NES Compendium and would tend to be lower
than the mean values used in the Corps data set be-
cause of positive skewness.
  Most of the apparent differences between lakes and
reservoirs  with respect  to  chlorophyll/phosphorus
ratio are explained by simultaneous variations in other
factors which influence algal growth, not by impound-
         ment type. These  factors, in  turn, can  be traced to
         watershed  characteristics which  control  the export
         and partitioning of nutrients, and the generation and
         transport of sediment. If one attempts to apply a given
         phosphorus/chlorophyll regression to a collection of
         lakes and/or reservoirs, the model  would be biased in
         certain systems, based upon N/P ratio, turbidity, and
         flushing rate, because it would not  incorporate effects
         of  limiting  factors other than phosphorus.  A more
         complex model is  needed if it is to be generally ap-
         plicable over the wide range of lake and reservoir con-
         ditions.

         REFERENCES

         Canfield, D.E., and R.W. Bachman. 1981. Prediction of total
           phosphorus concentrations,  chlorophyll  a, and Secchi
           depths in natural and artificial lakes.  Can. J. Fish Aquat
           Sci. 38(4):414-23.
                                                    296

-------
Carlson, R.E. 1977. A trophic state index for lakes. Limnol.
  Oceanogr. 22(2):361-9.
Hern, S.C., V.W. Lambou, LR. Williams, and W.D. Taylor.
  1981. Modifications of models predicting trophic state of
  lakes: adjustment of models to account for the biological
  manifestation of  nutrients.  EPA-600/3-81-001.  Environ.
  Monitor. Sys. Lab. U.S. Environ.  Prot. Agency, Las Vegas,
  Nev.
Lambou,  V.W.,  S.C. Hern, W.D. Taylor, and  LR. Williams.
  1982. Chlorophyll, phosphorus,  Secchi disk, and trophic
  state. Water Resour. Bull. 18(5):807-14.
Organization for Economic Cooperation and Development.
  1982. Eutrophication of Waters:  Monitoring, Assessment,
  and Control. Synthesis rep. OCED Coop. Progr. Eutrophi-
  cation. Paris.
U.S.  Environmental  Protection  Agency.  1978.   National
  Eutrophication Survey Compendium. Work.  Pap. 474-477.
  Corvallis Environ. Res.  Lab., Las Vegas Environ. Monitor.
  Support Lab.
Vermont  Department of Water Resources.  1980.  Vermont
  Lake Classification Survey. Water Qual.  Div. Lakes Progr.
  Montpelier.
Vollenweider, R.A.  1968. The Scientific  Basis of Lake and
  Stream Eutrophication, with Particular Reference to Phos-
  phorus and Nitrogen as Eutrophication Factors. Tech. rep.
  DAS/DSI/68. Organ. Econ. Coop.  Dev.,  Paris.
                  COMPARATIVE ANALYSIS OF RESERVOIRS


Walker, W.W. 1980. Variability of trophic state indicators" in
  reservoirs.  In  Restoration of Lakes and  Inland Waters.
  Proc. Int. Symp. Inland  Waters and  Lake Restoration,
  Portland,  Maine. EPA-440/5-81-010. U.S. Environ.  Prot.
  Agency, Washington, D.C.

	1981. Empirical methods for predicting eutrophica-
  tion in impoundments—phase I: data base development.
  Prepared for Office of the Chief, Army Corps of Engineers.
  Tech. rep. E-81-9. Waterways Exp. Sta., Vicksburg, Miss.

	1982a.  Calibration and testing of a eutrophication
  analysis  procedure  for Vermont lakes. Prepared for Vt.
  Agency Environ. Conserv. Dep. Water Resour. Environ.
  Eng. Lakes Prog. Final rep.

	1982b.  Empirical methods for predicting eutrophi-
  cation  in  impoundments—phase II:  model  testing.
  Prepared for Office of the Chief, Army Corps of Engineers.
  Tech. rep. E-81-9. Waterways Exp. Sta., Vicksburg, Miss.
	1983. Empirical methods for predicting eutrophica-
  tion in impoundments—phase II extension: model refine-
  ments. Prepared for Office of the  Chief, Army Corps of
  Engineers.  Tech.  rep.  E-81-9. Waterways   Exp.  Sta.,
  Vicksburg, Miss. Draft.
                                                      297

-------
                                      Fishery  Management
EFFECTS OF FISH ATTRACTORS ON SPORT  FISHING  SUCCESS
ON NORRIS RESERVOIR, TENNESSEE
R. Glenn Thomas
J. Larry Wilson
Department of Forestry,
 Wildlife and  Fisheries
University of Tennessee
Knoxville, Tennessee


            ABSTRACT

            A creel survey was conducted in 1980 on Morris Reservoir, Tenn., from March through October to
            assess the effects of artificial fish attractors (also known as hides, havens, shelters, or reefs) on fisherman
            success. An average of 9.8 brush-pile attractors was installed in each of 259 coves in the lake by
            TVA/CETA between 1978 and 1980. The 7.5-month creel survey yielded 1,435 party interviews,
            separating those fishermen who had fished only attractor areas (115 individuals) from those who had
            fished other areas exclusively (2,582), and those who had fished both attractor and nonattractor sites.
            Ten species of fish were caught by attractor fishermen, whereas nonattractor fishermen caught
            specimens of 19 species, and those who had fished both area types on the same day accounted for
            15 species. Comparisons of attractor and nonattractor creels indicated that percent successful, mean
            number of fish caught, mean fish per man/hour, and mean kilograms per man/hour were all significantly
            higher for attractor fisherman. Analysis of variance showed that those fishermen angling primarily
            for crappie (Pomoxis spp.) caught significantly more fish per man/hour and kilograms per man/hour,
            contributing most to the higher overall success rates for attractor fishermen.
INTRODUCTION

Fisheries  biologists and  fishermen  have long
recognized that underwater structures tend to attract
fish. Where natural cover is sparse, artificial fish at-
tractors (shelters, hides,  havens, or artificial  reefs)
have been used to concentrate desirable species for
sport and commercial fishing, while contributing more
or less positively to  a number of other biotic para-
meters. Experimentation has been conducted on at-
tractors' effects  on primary production, fish produc-
tion, and fish spawning, survival, and condition. These
studies document the reasons that "fish seem to find
there  (in shelters)  something which meets their
instinctive environmental  needs,  and whrch  is prob-
ably conducive to their general well-being and growth"
(Hubbs and Eschmeyer, 1983).
  In  this country, the effects of manmade concen-
trators have been studied largely in the last 50 years
(Pierce and Hooper, 1979), though in the coastal
waters of Japan, artificial reefs were constructed and
studied as early as 1800 (Stone, 1978). The efficacy of
artificial structures in concentrating fish in any struc-
tureless water, be it  coastal marine areas or clearcut
multipurpose reservoirs,  has  since been well  docu-
mented (Stroud, 1975). This study examines the possi-
ble changes in sportfishing success attributable to
the placement of brush fish attractors in such a reser-
voir.
  Specific objectives were to evaluate: (1) the amount
and  nature  of  attractor use by fishermen, (2) the
relative success of  fishermen in attractor and non-
                                             299

-------
 LAKE AND RESERVOIR MANAGEMENT
 attractor areas; and (3) differences in the composition
 of the catches made  in attractor areas, nonattractor
 areas, and in both area types.
 DESCRIPTION OF STUDY AREA

 Morris Reservoir is a 13,840 hectare storage impound-
 ment located about 42 km northwest of Knoxvillo,
 Tenn. The bifurcate reservoir extends 116 km up the
 Clinch River basin, and 90  km up the Powell River
 basin, as well as backing up many smaller tributaries.
 It was completed in 1936,  the first Tennessee Valley
 Authority (TVA) dam to be  closed.  The floodplain was
 clearcut  prior to impoundment, as  was standard prac-
 tice.
   Total  area of the Norris watershed is 7,542  km^,
 largely woodlands. Normal pool is  at 310.9 m, but fluc-
 tuations  of over 9 m in a year are common. The lake's
 mean depth is 22.9 m. A relatively clear, mesotrophic
 impoundment, visibility reaches 3  m at times.
   Norris is  located  in the ridge  and  valley physio-
 graphic  province of East  Tennessee.  Its banks are
 largely composed of limestone and dolomite bluffs
 and weathered shale ridges,  with  some clay/si t
 residuum banks and flats.  Most coves are narrow and
 deep.
   Aquatic macrophytes and emergent vegetation are
 extremely rare, as  is  large allochthonous  detritus.
 Combined with the preimpoundment basin clearing,
 these factors result in a significant lack of substrate
 for  the  attachment  of  periphyton (Prince  and
 Maughan, 1979a).
   In 1977,  TVA, in  cooperation with the Tennessee
 Wildlife  Resources Agency (TWRA), began construc-
 ting fish  attractors in Norris. Crews of Comprehensive
 Employment Training  Act  (CETA) personnel  and
 Young Adult Conservation Corps (YACC) personnel
 worked  with TVA Water  Resources supervision to
 build and place attractors. Coves were used exclusive-
 ly;  each  site received  an average of 9.8 weighted
 brushpiles in a  line down the middle of the cove (Fig.
 1). Brush units were low-profile (generally less than  1
 m high) and  up to 7 m in diameter. Brush (mostly hard-
 wood) cut on-site supplied  almost all the attractcr
 material, with tires used only occasionally. A  tree at
 the head  of each attractor cove was painted with a 2 n
 white stripe to identify the  site. By  1980, 259 sites had
 been completed.

 METHODS

 In January and February 1980, attractor coves were ex-
 amined and  marked at the  position of the deepest at-
   Fish  Attractor  Site
   Cove  Area
Figure 1.—Schematic representation of typical Norris attrac-
tor site.
tractor unit. With winter drawdown, most of the attrac-
tors in  each cove were  exposed;  submerged units
were located with a flasher-type depthfinder. A shore-
line tree adjacent to the deepest unit was flagged with
surveying tape to delineate the extent of each attrac-
tor area.
   From March 1 through Oct. 15 a survey of fishermen
was taken by boat. The lake was divided into six areas
of approximately 2,307 full-pool  hectares, each about
as  large as could be covered  in 6 hours.  Morning
surveys were for 1 he 6 hours after dawn and afternoon
surveys included  the 6 hours  before sunset. Days
worked  (5 of 7), sample time (morning or  afternoon),
and area of the lake were  selected at random by com-
puter.
  The following information was recorded for each
interview:
  1. Area (one of six)
  2. Sub-area:
    a. Having  fished atlractor area(s) exclusively
    b. Having  fished nonattractor area(s) exclusively
    c. Having  fished both attractor and nonattractor
areas
  3. Time of interview
  4. Number of hours fished
  5. Boat or bank fishing
  6. Number of fishermen in the party
  7. Number successful: Where  success was defined
as the taking of at least one fish deemed worth keep-
ing by the fisherman
  8. Primary species (or groups; as "crappie," "wall-
eye/sauger," "catfish") sought
  9. Number and aggregate weight of each species in
a party's creel
  10. Knowledge and usage of attractor sites by the
fishermen.
  Species-sought designations included the following
groups and individual species: crappie (white crappie,
Pomoxis annularis, black crappie, P. nigromaculatus);
bass  (largemouth  bass. Micropterus  salmoides,
smallmouth bass,  M. dolomieui,  spotted  bass, M.
punctulatus); sunfish (bluegill, Lepomis macrochirus,
rock bass Ambloplites rupestris); catfish (channel cat-
fish, Ictalurus punctatus,  flathead catfish, Pylodictis
olivaris);  walleye/sauger  (Stizostedion  vitreum/S.
canadense); white  bass  (Morone  chrysops);  and
striped bass (M. saxatilis).
  The data were transcribed, keypunched, recorded
on disc, and analyzed using the Statistical Analysis
System (SAS).
  Four  values  were obtained  for  the comparison
testing of  the  catches made in attractor and  non-
attractor areas. The percentage successful was defin-
ed as the number of successful fishermen in  a party
divided by  the total number in  the party (x  100%).
Mean  fish per  man at the time  of interview was the
total number of fish in a party's creel divided by the
number  in the party. Mean fish per man-hour was the
total number of fish caught divided  by the number of
man-hours  fished  by  that party. Kilograms per man-
hour was the sum of weights of all fish caught divided
by the total number of man-hours fished by that party.
  The comparison of the catches of attractor and  non-
attractor fishermen was first made by computing the
means for the four  terms just described for both
groups, and testing by t-test (at least 95 percent prob-
ability level). Grouping fisherman according to species
group sought results in  a more precise analysis of
catch rates (Davis and Hughes,  1965; Lambou, 1966).
The catch rates (in fish per man-hour and kilograms
per man-hour) of these sub-groups were compared by
                                                300

-------
                                                                                   FISHERY MANAGEMENT
two-way analysis of variance, and tested at 99 percent
probability levels. The mean weights of each species
taken in attractor areas and elsewhere were compared
at the 95 percent level (t-test).
RESULTS AND  DISCUSSION

Fisherman  Use  of Attractors

During the study period, 1,435 fishermen interviews
were  taken from  parties totaling  3,055  fishermen.
Fifty-eight party interviews were taken from 115 attrac-
tor fishermen. Eighty-four percent  of the interviews
taken (1,201 parties, 2,582 fishermen) were from peo-
ple who had fished areas other than attractor sites. A
total of 176 parties (358 fishermen) had fished both at-
tractor sites and nonattractor areas.
   The low number of parties found to have fished only
attractor sites may be attributed to several factors, in-
cluding:  (1)  the  necessity  of excluding  from this
category any fishermen who may have fished outside
the attractor area(s), and (2) the fisherman's frequent
lack of knowledge of the attractors. Fifty-seven per-
cent of the interviewees reported no use of attractor
sites at any time.
   Boat/bank fishing. The highest percentage of bank
fisherman  parties  in the three area types was in the
nonattractor group (20.5 percent).

Percent of Each Group
  (Number of Parties)   Attractor  Nonattractor    Both
        Boat
        Bank
84 5 (49)
15.5 ( 9)
79 5 (599)
20.5 (246)
98.3(173)

 1.7 (  3)
   This would be expected, since attractor coves con-
stitute only a small fraction of total bank area, and
most are inaccessible from the bank. The bank fisher-
men who did fish attractor sites did so more because
the area was easily accessible than because it con-
tained brushpiles. Few bank fishermen (three parties)
were mobile enough to fish both attractor and nonat-
tractor areas. The percentages of fishermen who fish-
ed from boats were similar enough in attractor (84.5)
and nonattractor (79.5) areas to preclude the possibili-
ty that this was an important bias favoring one group
over the other.

Relative Fisherman Success

Overall success. Four measures of relative success
for attractor and nonattractor fishermen are compared
in Table 1. Attractor fishermen were significantly more
successful (t-test, a  - .05)  than  were nonattractor
fishermen, by each evaluation: (1) percent successful,
(2) mean  fish per man, (3) mean fish per man-hour, and
(4) mean  kilograms per man-hour. Indications are that
greater mean catch rates on attractors were not the
result of  a few inordinately successful fishermen, and
attractor fishermen had more fish in their creels. Fish
were taken at a faster rate in attractor areas than in
nonattractor areas, and the greater catch on attrac-
tors was not comprised of higher numbers of much
smaller fish.
  Catch  rates of species-sought groups. The results
of testing the primary-species-sought success rates of
attractor and nonattractor fishermen are presented in
Table 2. Analysis of variance indicated that the catch
rates for attractor-area crappie  fishermen were prin-
cipally responsible for the overall higher catch rates of
attractor fishermen. The only catch rates that differed
significantly between the two types were for the crap-
pie fishermen (2.64 fish per man-hour versus 1.01 fish
per man-hour; 0.45 kilograms per man-hour versus 0.19
kilograms per man-hour),  and catfish, which showed
4.1  kilograms per man-hour on attractors, because of
a few exceptional catches. Sample size was too low to
effectively evaluate significance between catch rates
for catfish fishermen, as well as for sunfish fishermen.
Fishermen seeking  fish of any species outside attrac-
                 Table 1.—Overall success, including all fishermen in the two discrete-area qroups.
         Evaluation
                         Attractor Fishermen
                              (N = 1
                                                                                       Nonattractor
                                                                                    Fishermen (N=2582)
 Percent successful
 Mean fish per man
 Mean fish per man-hour
 Mean kiloarams per man-hour
                                                    56.75a
                                044^
                                                                     33.69

                                                                      1 32
                                                                      063
                                                                      0 16
"Significantly higher  a - 05
                       Table 2.—Catch rates for primarv-species-souqht fisherman qroups.
Primary Species
Group Sought
Any fisn

Crappie

Bass

Sunfish

Catfish

Variable3
FMH
KMH
FMH
KMH
FMH
KMH
FMH
KMH
FMH
KMH
Attractor
Catch Rate N
u.35
003
2.64f
0.45t>
022
015
050
0.04
080
4 fib
26
26
74
74
10
10
2
2
1
1
Nonattractor
Catch Rate N
u51
008
101
019
015
009
260
027
056
029
83J
833
539
539
403
403
202
202
35
35
 dFMH - mean fish per man hour
 KMH = mean kilograms per man hour

 ^Significantly greater a  = 01 where ANOVA shows a probabi'itv of greater F  - 0001
                                                  301

-------
 LAKE AND RESERVOIR MANAGEMENT
 tor areas caught fish (and tallied kilograms) slightly
 more often than did that group of fishermen in attrac-
 tor coves. Bass fishermen may have been  slightly
 more successful on attractors.
   Mean trip length at time of interview. Since signifi-
 cant variation in mean  trip length between attractor,
 nonattractor,  and  both-area  fishermen  could con-
 tribute  bias in catch-rate determination,  a Duncan's
 multiple range test was computed for those mean:;.
 Mean trip length was 2.09 hours for attractor fisher-
 men, 2.39 hours for nonattractor fishermen, and 3.77
 hours for those fishermen who had fished both area
 types. There was no significant difference (a = .05)
 between the means for attractor and non-attractor pa--
 ties, but both-area fishermen  averaged significantly
 longer times on the water at the time of  interview. It
 would be expected that, on the average, it would taka
 somewhat longer to fish both area types.  Since eaten
 rate testing was conducted only on the two nonover-
 lapping  groups of  fishermen,  the significant dr-
 ference for the mean trip length of both-area fisher-
 men was not considered relevant to the analysis.

 Catch  Composition

 Overall species diversity. Attractor fishermen took 310
 fish of 10 species, while nonattractor fishermen took
 3,400 fish of 19 species. Both-area fishermen  caught
 543 fish of 15 species.
  The large species diversity in the creel from non-
 attractor areas was likely a result both of  heavie-
 pressure in these areas and the  use of more varied
 fishing techniques there.  These variables would ex-
 plain the intermediate range species diversity (15) in
 the creel of both-area fishermen, the group which ac-
 counted for the intermediate amount and variability (in
 type) of effort.  Rotenone  sampling (Chapman, 1975;
 Brown, unpubl.) has indicated about the same number
 of species in attractor and nonattractor areas. Prince
 (1977) found that "the frequency of sport fishes on the*
 reefs generally coincided with their overall  abundance1
 in the impoundment."
  In  this study,  as  in others (Manges, 1959;  Pierce
 1967;  White,  1974), crappie comprised  by  far the
 largest  part of  the catch on  attractors (Table 3)
Seventy-four percent of the catch from attractors was
crappie, of which 14 percent were black crappie. Only
8 percent of the crappie taken  elsewhere  were black
crappie.
   Bluegills made  up the largest  part  of  the non-
 attractor catch, at almost 45 percent, and the second
 largest part of the attractor catch (17.7 percent), even
 though  only  one party was angling on  attractors
 primarily for  sunfish. As  with  crappie, the bluegill
 catch of both-area fishermen was intermediate to the
 other groups. Greater effort toward crappie  in attrac-
 tor coves (mainly with minnows for bait) contributed to
 the  higher crappie and lower bluegill catches there.
   The centrarchid basses comprised a larger percen-
 tage of the catch of both-area fishermen than of the
 other two groups;  most successful bass fishermen
 moved along  more or  less constantly.  Smallmouth
 bass were the principal bass taken by non-attractor
 and both-area fishermen, and largemouth bass were
 taken more often than other basses on attractor sites.
 Spotted  bass appeared least  frequently in the creel
 from all three area types. These results are in line with
 the reported tendencies for largemouth bass to orient
 to  underwater  structures  (Prince  and  Maughan,
 1979b),  for smallmouth bass to  stay  near  rocky
 substrates and open water, and for spotted bass to be
 distributed over various types  of habitat (Eschmeyer
 1944).
  Walleye were  seen nine times more often than
 sauger in the overall creel. Walleye and sauger were
 an important part (7.2 percent) of the catch of both-
 area fishermen, as were channel  catfish (7.2 percent).
 Many walleye fishermen were quite successful trolling
 in and out of deep attractor coves, which amounted to
 almost all of the fisherman usage of the deeper attrac-
 tor units. Walleye/sauger fishermen also took many of
 the both-area  channel catfish with  this trolling pat-
 tern.
  A  few  excellent catches of  large flathead catfish
 (mean weight  3.6 kg)  were made in  attractor areas,
 which accounted for 3.6 percent of the total number of
 fish caught there, and  45.0 percent of the total weight.
 It might  be expected that attractor areas would pro-
 vide  excellent forage  for the  large,  piscivorous  flat-
 heads, since attractor areas likely held many times
 the  number of small,  non-pelagic fishes (up  to 2,540
 bluegills  under 125 mm  per hectare (Brown,  unpubl.)
 that  nonattractor coves held.
  Catch composition by primary species sought. Sub-
dividing fisherman effort into species-sought groups
(and  a group of those fishermen seeking any species)
(Table 4) provided detail in success analysis (Lambou,
 1966).  The large numbers  of  harvestable bluegills
Table 3.—Catch composition of each area type expressed as the percentage of the total number of fish caught in that area
                                                type.


Crappie
Bluegill
Largemouth bass
Smallmouth bass
Spotted bass
Walleye & sauger
Flathead catfish
Channel catfish
White bass
Striped bass
Others

Lonqnose gar carp redhorse.

Attractor Area(s)
74.2
17.7
1.3
1.0
0.6
0.3
36
0.6
0.0
00
0.7
100.0
drum other sunfishes, rock bass, turtles
Percent of Catch
Nonattractor Area(s)
39.0
44.8
1.4
3.2
0.8
1.4
1.1
3.0
2.2
0.7
2.4
100.0


Both Area(s)
48.6
25.6
2.0
35
1 7
7.2
0.4
7.2
1.3
0.0
29
1000

                                                302

-------
                                                                                   FISHERY MANAGEMENT
found in attractor areas by rotenone inventory (Brown,
unpubl.) were not equally represented in the overall at-
tractor creel. However, they did appear among the fish
caught by fishermen seeking any species of fish near
attractors (85.7 percent bluegills).
  Crappie anglers in  all areas averaged about nine
crappie for every 10 fish kept. The methods employed
by crappie  fishermen were  demonstrably species-
selective.
  Bass fishermen who had not ventured from attrac-
tor areas took largemouths exclusively. Smallmouths
were especially prevalent (49.6 percent) in the creel of
bass  fishermen  who  had not fished attractor sites,
and were somewhat less common (30.4 percent) in the
creel  of fishermen who had fished both attractor and
nonattractor areas.
  Walleye and sauger and white bass were not sought
by any attractor fishermen. Walleye did appear as a
small fraction of the total catch of attractor fishermen
as incidental catch  in a crappie angler's creel.
  Mean weights. Average weights were computed for
each  species adequately represented in the creels
from attractor and nonattractor areas (Table 5). Slight
differences were seen between the groups, but none
were significant (t-test, a = .05). Fisherman selectivity
would be expected to negate any differences which
may have existed in the populations. Significant size
differences for attractor and nonattractor fishes were
seldom seen in fishermen's creels (Pierce, 1967), but
                         may be  demonstrated  by less  biased sampling
                         methods (Prince, 1977).

                         SUMMARY

                         1. During the study period, 1,435 party interviews were
                         taken. Of these, 84 percent (1,201) parties had fished
                         only nonattractor areas at the time of interview. The
                         greater total amount and variability (in type)  of effort
                         expended  by these fishermen probably explained the
                         larger species diversity (19 species) taken from non-
                         attractor areas. Attractor-only fishermen (58 parties)
                         took specimens of 10  species, and those fishermen
                         who had fished both  attractor and nonattractor areas
                         (176 parties) expended  the intermediate total amount
                         of effort  and  collected the intermediate number of
                         species (15).
                           2. Comparisons of  attractor and nonattractor creels
                         indicated  that percentage successful, mean number
                         of fish caught, mean fish per man-hour, and mean kilo-
                         grams  per man-hour were all significantly higher for
                         attractor  fishermen.  Crappie  anglers'   catch  con-
                         tributed most  to the  higher overall success rates for
                         attractor fishermen.
                           3. Public fishing pressure on the attractors was low
                         because fishermen were unaware of their installation,
                         attractor  markers disappeared, or fishermen  used
                         non-cove  fishing sites.  Deep  attractor  units  were
                         almost completely unused.
        Table 4.—Primary species groups sought expressed as percentage of all parties for each area type and
    principal species caught (percentage of total number caught by each species-sought-sub-group in each area type).
Primary Species
Group Sought
Any fish
Crappie
Bass
Sunfish
Catfish
Walleye/sauger
White bass
Miscellaneous
Attractor
Areas
17.3
67.3
10.3
1.7
1.7

—
1.7
100.0
Species
% Caught
Bluegill
85.7
Crappie
90.2
Larqemouth
'100
Bluegill
100
Flathead
100

—
Channel
100
Nonattractor
Areas
27.7
208
17.9
7.7
1.4
11.4
1.1
02
100.0
Species
% Caught
Bluegill
75.0
Crappie
90.2
Smallmouth
49.6
Bluegill
96.5
Channel
72.0
Walleye
257
Crappie
257
Channel
221
White Bass
71.1
Turtles
500
Both
Areas
13.8
28.7
32.8
3.5
1 7
19.5
—
—
100.0
Species
% Caught
Bluegill
43.0
Crappie
90.0
Smallmouth
30.4
Bluegill
100
Channel
87.5
Walleye
44.7
Channel
23.7
—
~
         Table 5.—Mean weights of species adequately represented in both attractor and nonattractor creel.
   Species
Area: Number of Samples
 Containing This Species
Mean Weight
    (kg)
Probability
  of >/t/
White crappie
Black crappie
Bluegill
Attractor.
Nonattractor:
Attractor:
Nonattractor:
Attractor:
Nonattractor.
il
163
10
42
10
220
U.19
0.18
0.15
016
013
0 11
u.ao
0.68
048
                                                  303

-------
LAKE AND RESERVOIR MANAGEMENT
REFERENCES

Brown, A.M. Unpubl. Fish inventory data comparing areas
  with and without fish attractors, Morris Reservoir,  1979.
  Tenn. Valley Author, rep. 1980.

Chapman, W.R. 1975. Small lake management: fish attractor
  evaluation. N.C. Wildlife Resour. Comm., N C. Federal Aid
  Proj. F-21. Raleigh, N.C.

Davis, J.T., and J.S. Hughes. 1965. Creel census on  Busse/
  Brake Reservoir for the first three years. Proc. Annu. Conl
  Southeastern Game and Fish Comm  17:495-506.
Eschmeyer, R.W. 1944. Norris Lake fishing, 1944. Tenn. Div
  Game and Fish, Nashville.

Hubbs, C.L, and R.W. Eschmeyer 1938. The improvement of
  lakes for fishing, a method of fish management Bull. Inst
  Fish. Res. 2:1-233.

Manges, D.E. 1959. Large impoundment investigations, study
  of the  construction and  value of brush shelters. Tenn.
  Federal  Aid  Proj.  F-1. Tenn.  Wildl.  Resour. Agency,
  Nashville.

Lambou,  V.W.  1966. Recommended  method of  reporting
  creel survey data for reservoirs Okla. Dep. Wildl. Conserv
  Bull. 4:1-33.

Pierce, B.E. 1967. Brush  shelter studies.  W Va.  Federal Aid
  Proj. F-10-R-8. West Va Dep.  Nat. Resour. Div. Game and
  Fish, Charleston.
Pierce, B.E., and G.R. Hooper. 1979. Fish standing crop com-
   parisons of tire and brush fish attractors in Barkley Lake,
   Ky.  Proc.  Annu. Conf.  Southeastern  Game  and  Fish
   Comm. 33:688-91.

Prince, E.D. 1977. The biological effects of artificial reefs in
   Smith  Mountain Lake, Va. Ph.D. diss. Virginia Polytechnic
   Inst. State Univ., Blacksburg.

Prince, E.D., and O.E. Maughan. 1979a. Attraction of fishes to
   tire reefs in Smith Mountain Lake, Va. Pages 19-25 in D.L
   Johnson and R.A. Stein, eds. Response of Fish to Habitat
   Structure in Standing Water. North Central Div. Am Fish
   Soc. Spec. Publ. 6.

	•  1979b. Telemetric  observations  of largemouth
   bass near underwater structures in Smith Mountain Lake,
   Va. Pages 26-32 in  D.L Johnson and R.A. Stein,  eds.
   Response  of  Fish  and  Habitat  Structure in  Standing
   Water.  North Central Div. Am. Fish. Soc. Spec. Publ. 6.
Stone,  R.B.  1978. Artificial  reefs and fishery management.
   Fisheries: Bull. Arn.  Fish. Soc. 3(1):2-4.

Stroud, R.H.  1975. Artificial reef construction. Sport Fish
   Inst  Bull. 264:4-5.

White,  M.G., III. 1974.  Installation and evaluation of fish
   attractors. S.C. Federal Aid Proj. F-16-4.  S.C. Wildl. Marine
   Resour. Dep  Div. Game and Freshw. Fish. Columbia, S.C.
                                                     304

-------
RECENT  APPLICATIONS OF  HYDROACOUSTICS TO ASSESSMENT
OF LIMNETIC  FISH ABUNDANCE AND  BEHAVIOR
RICHARD E.  THORNE
GARY L. THOMAS
Fisheries Research Institute
University of Washington
Seattle,  Washington
            ABSTRACT

            Since 1969, the authors have been involved in over 200 hydroacoustic surveys of fish populations
            in more than 25 lakes. These studies have included a variety of different species assemblages and
            objectives, although most, such as Lakes Washington and Ozette in Washington and Tustumena in
            Alaska, are sockeye salmon nursery lakes. The objectives of these studies have included fisheries
            management, evaluation of lake enhancement programs, or environmental impact. During the 14 years
            of these investigations, both the equipment and procedures have evolved and improved considerably.
            Earlier techniques were very limited in their ability to detect fish near surface or in shallow water and
            had very limited capabilty for size discrimination. Current technology has solved most of these pro-
            blems. These developments and their capabilities are presented along with the results of surveys on
            lakes with a variety of biological and physical characteristics. The results include a considerable amount
            of "ground truth" data from other assessment techniques. In many cases these data are obtained
            from various net sampling techniques. However, some comparisons have revealed considerable biases
            with net sampling techniques which are associated with changes in light intensity, turbidity,  or fish
            behavior.
INTRODUCTION

Hydroacoustic techniques are gaining widespread ac-
ceptance in the field of fisheries science. The increase
in importance of this tool is largely the result of its
ability  to  avoid the limitations of traditional tech-
niques, such as catch per unit effort (c/f) and mark and
recapture. In general,  c/f data are subject to large
sources of error from factors that affect the activity of
fish and the efficiency of the gear, whereas, the mark
and recapture techniques are often too time consum-
ing and expensive to be feasible.  In addition, the fac-
tors that affect c/f data are often the same factors that
characterize different fish habitats, i.e., current, depth,
water clarity, etc.
  While hydroacoustic techniques escape many of
these limitations, they are subject to several others.
The purpose of this paper is to review some recent ap-
plications of hydroacoustic techniques for  assess-
ment of limnetic fishes and to report on progress in
overcoming these limitations.
Validity of Hydroacoustic Techniques

It   is  appropriate  to  establish  the  status  of
hydroacoustics as a valid technique for assessing fish
abundance. This will be approached in two ways: (1) by
reviewing the extensive use of hydroacoustic techni-
ques, and (2) comparing its results with those of tradi-
tional techniques.
   Hydroacoustics has been used for over 50 years to
detect fish and has long been a major tool for  com-
mercial fisheries. Application to fisheries manage-
ment has been more recent, but has increased con-
siderably since the development of the echo integrator
and supporting theory in the late 1960's. International-
ly,  emphasis  on  the  application  of  hydroacoustic
techniques has been overwhelmingly on marine  com-
mercial fisheries management. For example, at the re-
cent Symposium of Fisheries Acoustics (Bergen, Nor-
way, 1982) only two of 114 papers dealt with applica-
tions in lakes. The reasons for this imbalance are pro-
bably the historical use of hydroacoustics in marine
commercial fisheries and the considerable economic
value of marine fisheries.
  An exception to this trend is the experience of the
Fisheries  Research Institute  of the  University  of
Washington. However, this has resulted from interest
in the limnetic  phase of a  fish subject  to a major
marine commercial fishery, the sockeye salmon (On-
corhynchus nerka). The Institute was established to
study limnetic factors  affecting sockeye salmon pro-
duction  in  Alaska. Hydroacoustic techniques were
first used as part of this research in 1961 (Rogers,
1967). This application was limited to echogram count-
ing from commercial echosounders and was not very
successful because of the near surface distribution of
the fish and limitations of the  equipment. In 1968, as
part of the newly-organized Sea Grant Program, the In-
stitute began research on the echo integration tech-
nique, including its application to Lake Washington in
1969 (Thome and Woodey, 1970). In the subsequent 13
years, the  Institute has been involved in over  230
surveys on 35 different lakes in 6 States and 3 coun-
tries using echo integration  and echo counting tech-
niques (Table 1).
  Establishing the validity of hydroacoustics as  an
assessment technique is complicated by the lack of
adequate alternative means of assessment for com-
parison.  However, several  comparisons  have been
made using a variety of techniques. Examples in the
marine environment are described in Thorne (1973,
1977a) and Trumble et  al. (1982). In  the freshwater en-
vironment one of  the  most  extensive data bases is
from Lake Washington where the  Institute has con-
ducted 65 hydroacoustic surveys over 12 years. Com-
parisons of hydroacoustic estimates of adult sockeye
                                                305

-------
LAKE AND RESERVOIR MANAGEMENT

salmon  with  two other techniques  (Thorne, 197(3)
showed less than a 2 percent difference between the
three methods over a 4-year period.
   Comparisons  of  hydroacoustic  estimates  of
juvenile sockeye  salmon with estimates for net  cat-
ches are also extensive in Lake Washington. This lake
is unusual in that it is accessible to large vessels so
that efficient net sampling techniques can be used. A
3 m Isaacs-Kidd midwater trawl  is towed at over 2.5
m/sec.  Figure  1  shows  a comparison  of  the
hydroacoustic populations  estimates  with the mean
catch per unit effort for corresponding surveys. Each
mean is the result of over  20 tows made at several
depths and five areas in a stratified sampling design.
The two techniques compare very well except for two
occasions when analysis of the acoustic data demon-
strated  that the  five net sampling areas were  not
representative of the total lake. Excluding those two
points, the correlation coefficient (r2) is 0.73.
  While  such  comparisons are valuable and  inspire
confidence in  the acoustic  techniques, they must be
made cautiously  because  of biases  in the various
techniques. For example,  similar comparisons  be-
tween acoustic estimates and net catches of sockeye
in Great Central Lake, British  Columbia,  by Robinson
and Barraclough (1978) showed that the relationshia
between acoustic estimates and net catches changed
with changes in  moonlight and cloud  cover. The/
determined that catching efficiency for clear sky and
dark of moon was only 43 percent of that for overcast
                Acoustic F'opulation Estimate (millions)
Figure 1.—Comparison of hydroacoustic estimates of pre-
smolt sockeye salmon with weighted mean midwater trawl
catches from corresponding surveys in Lake Washington
1970-1981.                                        '
    Table 1.—Locations and numbers of hydroacoustic surveys involving personnel of the University of Washington's
                                       Fisheries Fiesearch Institute.
Location
Washington









Alaska












British Columbia





Colorado
Idaho
Montana

New Mexico
Tanzania
Total
Lake
Banks
Chelan
Chester Morse
Osoyoos
Ozette
Quinault
Ross
Sammamish
Washington
Wenatchee
Auke
Becharof
Blue
Crescent
Euchemy
Hugh Smith
Iliamna
McDonald
Nunavaugaluk
Orton
Osprey
Tustumena
Ugashik
Babine
Cultus
Frazer
Harrison
Shuswap
Pitt
Twin Lakes
Pend Oreille
Flathead
Ashley
Morgan
Tanganyika
35 lakes
Year
1973
1971
1974
1973-74
1979-80
1971-75
1970-73
1972-73
1969-80
1972-75
1975
1974
1975
1982-83
1982
1982-83
1971-80
1982-83
1973
1975
1975
1981-83
1974
1975
1977,1980
1980
1980
1971-74,80
1980
1980
1975
1980
1980
1973
1975-76

Major Species
Various
Various
Various
Sockeye salmon
Sockeye salmon
Sockeye salmon
Rainbow trout
Sockeye salmon
Sockeye salmon
Sockeye salmon
Dolly Varden
Sockeye salmon
Dolly Varden
Sockeye salmon
Sockeye salmon
Sockeye salmon
Sockeye salmon
Sockeye salmon
Sockeye salmon
Dolly Varden
Doly Varden
Sockeye salmon
Sockeye salmon
Sockeye salmon
Sockeye salmon
Sockeye salmon
Sockeye salmon
Sockeye salmon
Sockeye salmon
Lake trout
Kokanee
Various
Various
Various
Various

No. Surveys
2
9
2
6
4
49
14
9
65
7
1
1
1
3
1
4
10
2
1
1
2
8
1
1
2
1
3
5
5
3
2
1
1
2
2
231
                                                306

-------
sky and dark of moon, and catching efficiency for
clear sky and full moon was only 6.4 percent. Other
studies  have shown  changes in catching efficiency
with changes in environmental conditions (Thorne, in
press).
Progress in Minimizing Technique Limitations

Hydroacoustic  techniques similarly  have specific
limitations that have to be taken into account. Limita-
tions described by Thorne (1977b, in press) include: (1)
limits to detection  near surface  and bottom  boun-
daries,  (2)  lack of  biological  information including
species and  size, and (3) potential bias associated
with target strength.
   Detection of  fish  near boundaries is limited by the
finite width of the pulse  of sound transmitted  by an
acoustic system. The  problem  is aggravated for near
surface fishes  by potential avoidance  of the  boat
(Olsen et al. 1982). Approaches to overcoming this
limitation include:  (1) high frequency,  short  pulse
length  and  towed  transducers;  (2)  stationary
transducer deployment; (3) doppler sonar; and (4) diel
stratification. Short  pulse acoustic  systems  with
transducers in towed bodies are being used in marine
assessments to achieve greater detection of fish near
bottom  (Dickie  et al. 1983). In lakes, however, the use
of smaller vessels may preclude such operation. Us-
ing hydroacoustic techniques to assess juvenile sock-
eye salmon in  Lake Tustumena, Alaska, during 1982
and 1983 illustrates an approach that was taken for
limnetic fishes.
                              FISHERY MANAGEMENT

  Previous tow net surveys at Lake Tustumena have
shown that fish densities are often highest near sur-
face. Even using a small boat (7 m length) and towing
the transducer just below  the surface from a  bow
mounted boom, the technique  was only effective
below 2.5 m. To obtain fish abundance information
from the upper 2.5 m, transducers were placed on the
bottom  in several locations, oriented toward the sur-
face. Such a deployment had been used previously to
study the abundance and behavior of fish under the
ice in the nearshore Beaufort Sea (Thorne, 1980) and
could detect fish within 0.1  m of the surface.
  The disadvantage of such stationary deployments
is the small  sampling power, since the transducer is
fixed in  a single location. However, in this application
the purpose was not to obtain information on the den-
sity of fish in the upper 2.5 m for extrapolation to the
lake as a whole, but to determine the near surface ver-
tical distribution, or  more specifically the  relative
abundance in the upper 2.5 m compared to the 2.5 to 5
m depth strata that could be measured with the usual
procedures.
  The results from several stations demonstrated that
the vertical distribution varies much less from station
to  station than the actual fish abundance. Therefore,
the necessary correction for the upper depth strata
could be obtained with a reasonably small number of
stations.
   A complicating problem at Lake Tustumena was the
fact that much of the lake was too deep for a bottom
mounted transducer. To avoid potential bias from dif-
ferent vertical distributions in deeper water, it was
necessary to devise a mode of uplooking deployment

              BOAT AND ELECTRONICS
                  ANCHOR


Figure 2.—Diagram of up-looking transducer deployment for deep water.
                                                 307

-------
 LAKE AND RESERVOIR MANAGEMENT
 that could  be used in the deeper water. This was,
 achieved by suspending a transducer at depth bet-
 ween an anchored surface buoy and the boat in such a
 way  that  the surface above the transducer  was.
 unobstructed (Fig. 2).  Initially, problems were  en-
 countered with stability and  orientation,  but these
 were solved by adding a gimbaled mount to the trans
 ducer framework.
   Doppler sonars  are another  way to detect fish near
 boundaries (Acker and Hendershot, 1982), but they are-
 presently in the developmental stage. Often problems
 with distribution near boundaries can be overcome by
 taking advantage of diel changes in vertical migration
 This is  particularly true for near-bottom fish that ver
 tically migrate (Thorne et al. 1982), but the reverse was
 true for Lake Tustumena, where we discovered thai
 the fish were near  the surface during day, but dispers-
 ed downward at dark. This unusual reverse  vertical
 migration  may be an  adaption to limited light  for
 feeding in the turbid, glacial lake.
   Since hydroacoustic techniques do not directly cap-
 ture fish, information on biological  characteristics, in-
 cluding species and size, is not directly obtained. Con-
 sequently, hydroacoustic  techniques  are  virtually
 always  used in conjunction with direct capture techni-
 ques. However, often the techniques are not well coor-
 dinated, leading to inefficiency. We investigated how
 best to use direct  capture techniques in conjunction
 with hydroacoustics.  Our philosophy is that hydro-
 acoustic techniques alleviate the need for direct cap-
 ture techniques to determine  fish  abundance. Thus,
 the  focus  on their use  in conjunction with hydro-
 acoustics should  be on how  best to determine the
 composition   of   the  hydroacoustically  measured
 population.
   This philosophy  affects the operation of direct cap-
 ture techniques in  two ways. First, to achieve an un-
 biased  mean, certain  limitations,  such as random-
 ness, are imposed  on the allocation of direct capture
 effort for abundance information. In contrast, with
 hydroacoustics, allocation of direct capture effort  is
 directed by the hydroacoustic observations,  that is,
 they are deployed when hydroacoustics detects a con-
 centration of fish that needs identification. The mean
 catch is not the concern, only  the composition.
   Second,  when direct capture techniques are  used
 for abundance, the underlying  assumption is that the
 catching efficiency is virtually constant. In contrast,
 advocates of hydroacoustics, as well as many others,
 contend that catching efficiency is  extremely variable
 in a manner that is  difficult, if not impossible, to deter-
 mine. Therefore, investigation  of net performance in
 support of hydroacoustics focuses on selectivity and
 relative  changes in efficiency  among species, rather
 than on overall net efficiency. The  hydroacoustic
 capability to know  the number, location, and behavior
 of fish has already improved understanding of some
 gear types (Thorne, in press) and will provide a firm
 basis for evaluating such techniques  as they are
 employed together in future studies.
  While  research  on  hydroacoustic  techniques for
 direct identification of fish has been relatively minor,
 considerable attention has been given to the problem
of directly determining fish size, since information on
 fish  target  strength  is  needed  to  convert  hydro-
acoustic data  to  absolute abundance, as well  as
whatever size data  is obtainable. The literature in this
area is extensive (Ehrenberg, 1982). Research has con-
centrated in two areas: (1) development of techniques
to extract fish target information directly from conven-
tional single  beam transducer systems  (Clay, 1983),
 and (2) further development of the dual beam target
 strength measurement system (Ehrenberg, 1974). An
 important recent development is a  microprocessor-
 based dual  beam data processor, developed at the
 University of  Washington and being marketed by
 Biosonics, Inc., of Seattle.
   While   instrumentation  to  measure  the  target
 strength of fish is becoming more powerful and more
 available, the amount of size information available in
 the acoustic signal is still limited by  the fact that the
 range of target strengths associated with a fish of a
 given size covers about 30  dB, whereas the mean
 target strength changes only about 7 dB per doubling
 of fish length. Thus information about size requires
 large samples. While it was possible  to readily distin-
 guish adult sockeye salmon (over 60 cm) from resident
 fish (less than 20 cm) in Lake Washington  as des-
 cribed in Thorne (1979), the effectiveness of the tech-
 niques  to extract information about  smaller dif-
 ferences in a mixed population is still open to ques-
 tion.
   However,  while direct discrimination between two
 size groups in a mixed population is very difficult with
 present techniques, detection of different size groups
 in different portions of a lake is frequent. Usually the
 separation is by depth, and often by a thermocline. An
 example is  the  surveys conducted  in  Twin  Lakes,
 Colo., in  1980. Figure 3 shows the vertical distribution
 of two size groups of fish at  night. The deeper group
 were adult lake trout (Salvelinus namaycush) and were
 below the thermocline. Their acoustic target strengths
 were greater than -37 dB. The shallower fish group,
 above the thermocline, had target strengths less than
 - 55 dB and were probably juvenile lake trout. Similar
 stratification of different size groups has been noted
 in sockeye salmon nursery lakes, where fry and yearl-
 ings overlap  briefly in time,  but often not  in  space.
 Other investigators have noted similar stratification
 using  hydracouslic  techniques  (Burczynski,  pers.
 comm.).

 Other Problem Areas

 Two other disadvantages noted by Thorne (1977b,  in
 press)  were  (1) high cost of equipment, and (2) com-
                    RELATIVE ABUNDANCE
Figure 3.—Vertical distribution of two size groups of fish in
Twin Lakes, Colo., during September 1980.
                                                 308

-------
                                                                                        FISHERY MANAGEMENT
plexity. Many types of echosounders are available, but
most were designed for commercial fishing and are
not adequate for scientific purposes. Only one scien-
tific quality echosounder has been available for under
$12,000, the  Simrad EY-M, but this situation will be
helped by the scheduled production of the Biosonics
Minisounder in 1984. Commercial availability of micro-
processor-based data analysis systems  is also in-
creasing  rapidly.  The seeming complexity  of hydro-
acoustic  techniques results  primarily from  a lack of
training in these concepts in our  professional  and
academic institutions, and remains one of the major
hindrances to their  more widespread use.
REFERENCES

Acker,  W.,  and R. Hendershot. 1982.  New  concepts for
  riverine  sonar.  Paper  107. Symp. Fisheries Acoustics.
  Bergen, Norway.
Burczynski, J. Pers. comm. 1983. Bionsonics, Inc., Seattle,
  Wash.
Clay, C. 1983. Deconvolution of the fish scattering PDF from
  the echo PDF for a single transducer sonar. J. Acoust. Soc.
  Am. 73(6): 1989-94.
Dickie, L, R. Dowd, and P. Boudreau. 1983. An echo counting
  and logging system (Ecolog) for demersal fish size distri-
  butions  and densities. Can. J.  Fish. Aquat. Sci. 40(4):
  487-98.
Ehrenberg,  J.  1974.  Two applications for  a dual-beam
transducer  in hydroacoustic assessment systems. Oceans
74.
	1982. A review of in situ target strength estimation
  techniques. • Paper  104.  Symp.  Fisheries Acoustics.
  Bergen, Norway.
Olsen,  K., J. Angell, and F. Pettersen. 1982.  Observed fish
  reactions to a surveying vessel with special reference to
  herring,  cod, capelin, and  polar  cod.  Pap. 48. Symp.
  Fisheries Acoustics. Bergen, Norway.
Robinson,  D., and W.  Barraclough.  1978. Population  esti-
  mates of sockeye salmon (Oncorhynchus nerka) in a fer-
  tilized oligotrophic lake.  J. Fish. Res. Board Can. 35(6):
  851-60.
Rogers, D.  1967. Estimation of pelagic fish populations in the
  Wood River lakes, Alaska, from tow net catches and echo-
  gram marks. Ph.D. Thesis. Univ. Washington.
Thorne,  R. 1973. Acoustic assessment of hake,  1968-1973.
  Oceans  73.
	1977a. Acoustic assessment of hake and herring
  stocks in Puget Sound, Wash., and southeastern Alaska.
  ICES Rapp. et Proces-verbaux 170: 265-78.
	. 1977b. Acoustic surveys. Pages 20-38 in A. Saville,
  ed. Survey Methods of  Appraising Fishery  Resources.
  Fish. Tech. Pap. 171. Food Agric. Organ.
	1979. Hydroacoustic estimates of adult sockeye
  salmon  in Lake Washington, 1972-5. J. Fish. Res. Board
  Can. 36: 1145-9.
	1980. Application  of stationary hydroacoustic sys-
  tems for studies of fish abundance and behavior. Oceans
  80.
	In press. Hydroacoustics. Chapt. 12 in L. Nielsen
  and D. Johnson, eds. Fisheries Techniques. Am. Fish. Soc.
Thorne, R., and J. Woodey.  1970. Stock assessment by  echo
  integration and its application to juvenile sockeye salmon
  in Lake  Washington.  Fish. Res. Inst. Circ. 70-2.  Univ.
  Washington.
Thorne,  R., R. Trumble, N. Lemberg, and D. Blankenbecker.
  1982. Hydroacoustic assessment and management of her-
  ring  fisheries in Washington and  southeastern Alaska.
  Pap. 98. Symp. Fisheries Acoustics. Bergen, Norway.
Trumble, R., R. Thorne, and N. Lemberg. 1982. The Strait of
  Georgia herring fishery: A case history of timely manage-
  ment  aided by hydroacoustic surveys. Fish.  Bull.  80(2):
  381-8.
                                                     309

-------
  USE OF COLUMBIA RIVER RESERVOIRS
  FOR  REARING BY JUVENILE FALL CHINOOK SALMON AND SOME
  MANAGEMENT IMPLICATIONS
  GERARD A. GRAY
  DENNIS W.  RONDORF
  U.S. Fish and Wildlife Service
  National Fisheries Research Center
  Willard  Field Station
  Cook, Washington

             ABSTRACT
                rohr,hR    wh-         f fa" Chin°0k Salmon have been altered bV impoundment of
             the Columbia River. While growth rates remain high, riverine survival has been reduced by im-
             poundment, passage mortality at dams, predation, disease, thermal stress, and possibly late
             ocean entry. Late ocean entry may reduce fall Chinook salmon survival by delaying passage
             through a sequence of habitats (spatial  windows) at a time, size, or physiological state (temporal
             windows) when they are best adapted  to that habitat. Extended freshwater residence retards
             higher growth rates normally associated with entry into the estuary and ocean, and reduces the
             higher survival associated with increased size. Managers of hatchery and wild stocks must take
             these spatio-temporal windows into account when evaluating management techniques such as
             flow manipulation or transportation.
 INTRODUCTION

 Man-induced changes in the Columbia River basin
 have produced environmental changes that adverse y
 affect the survival of fall chinook salmon (Oncorhyn-
 chus tshawytscha). Logging, grazing, road construc-
 tion, and mining have polluted tributaries once used
 for spawning and rearing, and dams constructed on
 tributaries and the main stem have replaced produc-
 tive spawning and  rearing  habitat  with slack-water
 reservoirs. These developments have led to decreased
 survival of juvenile and adult migrants through turbine
 mortality and other passage problems. The resulting
 decline in productivity and survival, coupled with in-
 creased commercial and sport fishing pressure, has
 caused a significant reduction  in  returns  of adul
 salmon to the river. Twenty-two State and Federal hat-
 cheries currently produce juvenile fall chinook salmon
 each year to aid in the mitigation of these losses. In-
 creasing numbers of hatchery fish are being released
 in main stem reservoirs where  the habitat differs
 radically from the riverine environment typically used
 by fall chinook salmon during their early life history. If
 production of fish released  in reservoirs is to be ir-
 creased, it is necessary to assess critical factors of
 the freshwater life history of  wild juvenile fall chinook
 salmon and  relate  these factors to management of
 hatchery fish released in reservoir habitats.
  Our objectives in the present analysis were to: (1)
 determine the relation between early life history of
 wild fall chinook salmon in  the Columbia River and
 survival to the adult state, (2) identify changes in life
 history that have occurred after impoundment, and (3)
 outline management alternatives to increase the use
 of reservoirs in the Columbia  basin to rear fall chinook
 salmon.

 ORIGINAL  LIFE HISTORY

 Historically, fall chinook salmon spawned in the main
stem Columbia River and in lower portions of its large
 tributaries (Fulton, 1968). Eggs hatched in spring and
 fry emerged from the gravel soon thereafter. Most fall
 chinook  salmon  fry  drifted  downstream   after
 emergence,  and within  a short period began  to
 migrate actively, reaching the estuary during their first
 year of life. Migration rates depend largely on water
 velocity; they exceeded 40 km per day in the Snake
 River (Krcma and Raleigh, 1970), but were  generally
 much slower. Growth during the downstream migra-
 tion was often substantial; Rich (1922) estimated that
 the  mean length  of fish entering the Columbia  River
 estuary increased from about 55 mm in June to about
 95 mm in October.
   Columbia River fall chinook salmon may  have ex-
 hibited other rearing and migration tendencies under
 free-flowing conditions, but we are not aware of any
 pertinent studies,  because differentiating of stocks in
 the Columbia River without marking was difficult and
 the marking of fry was generally impractical.  In the
 Sixes River, Ore., a system with less fish stock diversi-
 ty than the Columbia  River,  Reimers (1973)  used
 scales to identify  six major life history types among
 juvenile fall chinook salmon, ranging from immediate
 migration  to  sea  after emergence to remaining in
 freshwater for a full year before emigration; most com-
 monly fish remained  in freshwater through early sum-
 mer and emigrated to the estuary  for short-term rear-
 ing prior to ocean entry. Survival was  highest among
 fish  that remained in streams until  early summer,
 emigrated to and  remained in  the estuary until  fall'
 and then emigrated to sea. this was the second most
 abundant life history type. Highest adult returns came
 from juveniles that entered the estuary at a length of
 about 90 mm and emigrated to the ocean at about 130
 mm;  juveniles less than 120 mm at ocean entry  pro-
duced few returning adults (Reimers, 1973;  Reimers
and Downey, 1982). Fall chinook salmon survival in the
Sixes River was directly related to the attainment of a
certain minimum length before ocean entry, and there
                                               310

-------
                                                                                 FISHERY MANAGEMENT
was no apparent survival advantage to growth in ex-
cess of this length (Reimers and Downey, 1982). The
minimum length was most often reached by juveniles
that remained in freshwater and estuarine habitats for
extended periods before entering the ocean.
  The "minimum length theory" was corroborated by
using hatchery-reared salmon in both the Sixes River
(Reimers and Concannon, 1977)  and the Columbia
River (Fowler and Banks, 1980), and was consistent
with mean lengths of fall Chinook salmon found in the
Columbia River estuary before dam construction. Rich
(1922) who sampled  juvenile Chinook  salmon  in
1914-16 in  the Columbia River estuary with a beach
seine, estimated that mean length of fish reached  a
fall maximum of about 110 mm in October. However,
recent sampling by Dawley et al. (1982) indicated that
beach seine samples underestimate the mean length
of fish found offshore by 5-20 mm, suggesting that
mean length of fish in the earlier study by Rich (1922)
was probably 110  to 130  mm. As  judged by these
observations, it is likely that mean length of fish enter-
ing the ocean was  about 130 mm.
CURRENT LIFE HISTORY

The life history of fall Chinook salmon in the main
stem Columbia River and in large tributaries such as
the Snake River has been radically altered by changes
in spawning, rearing, and migrating conditions that
accompanied or followed impoundment of the river.
Passage to major spawning grounds was blocked by
the construction of Hells Canyon Dam (river kilometer
395) on the Snake River and Chief Joseph Dam (river
kilometer 872) on the Columbia River. Of the spawning
habitat now accessible, 65 percent on the Snake River
and 88 percent on the Columbia River have been inun-
dated by the construction of 13 hydropower dams. Fall
chinook salmon produced in these spawning areas, as
well as many hatchery fish, must migrate in reaches
impounded by these dams.
   Studies to determine the effect of impoundments
on the freshwater life history of juvenile salmon have
indicated that although growth is sometimes substan-
tial  (Parente and Smith, 1981), migration  rates are
reduced (Thorpe and Morgan, 1978;  Raymond, 1979).
Growth of fall chinook salmon in impounded sections
of the Snake River exceeded that  in the river (Krcma
and Raleigh, 1970); growth can exceed 1  mm/day in
Columbia River impoundment  backwaters (Parente
and Smith, 1981). The principal dietary components of
fall  chinook  salmon  included terrestrial insects,
microcrustaceans, and aquatic insects in impounded
reaches of the Columbia  River  (U.S. Fish Wildl. Serv.,
unpubl. data) as opposed to largely aquatic insects in
riverine habitats (Becker,  1973).
   Migration rates of juvenile salmonids are slower in
reservoirs than in free-flowing  environments (Thorpe
and Morgan,  1978;  Raymond,  1979).  For example,
mean migration rate  for hatchery reared fall chinook
salmon has been estimated to be 5.8 km/day in a main
stem reservoir (Sims and Miller, 1982), but was 27 km/
day in a reach that was 41 percent free-flowing. Sims
(1970) observed a maximum migration rate of 10.9 km/
day among wild fall chinook salmon in an upper Snake
River impoundment. Other studies have indicated that
subyearling chinook  salmon travel  the 168  km be-
tween Priest Rapids Dam and McNary Dam  in about 9
days, and pass through the John Day pool in an addi-
tional 22 days (Sims et al. 1977;  Sims and Miller, 1982).
If fall chinook salmon were released at Priest Rapids
Dam at a mean length of 85 mm, grew at 1 mm/day,
and migrated downstream  at the same rate as esti-
mated for subyearling chinook salmon,  mean length
at arrival at John Day Dam would be about 116 mm.
This length  is within the range of 100 to 120 mm
minimum  length at  ocean entry  noted in Oregon
coastal rivers; however, Columbia River fish  still must
travel 345 km and pass two dams before reaching the
estuary. Thus,  juvenile fall  chinook salmon continue
to live  in  a  reservoir environment when, in a more
natural life history situation, they would  have already
entered the estuary or ocean. Delays in migration of
juvenile salmonids through reservoirs may reduce sur-
vival (Sims,  1970;  Raymond, 1979) as a result of  in-
creased  predation,  thermal  stress, and  disease,
especially during  low flow years (Park, 1969;  Ray-
mond, 1979).
  Survival of juvenile fall chinook salmon migrating
through  reservoirs can be relatively high  if migrations
are short in distance and duration. Estimated survival
of fish traveling 100  km  through Brownlee  Reservoir
on  the Snake  River ranged from 85 to 100 percent
when the  migrations  were  not delayed by low flows
(Sims, 1970). In the Columbia River, the survival of
juvenile fall chinook salmon migrating 292 km through
two reservoirs  and  past McNary Dam was estimated
at 45 percent during 1977, a low-flow year (Sims et al.
1978). If 11 percent passage mortality was assumed
for McNary  Dam (Schoeneman et al. 1961), survival
was about 80 percent per 100 km of travel, or about 20
to 30 percentage points higher than the 50-60 percent
per 100 km reported in the free-flowing Columbia River
below Bonneville Dam (Dawley et al. 1980). Since low
survival  is associated with prolonged  residence in
reservoirs, either  dam passage mortality  is higher
than 11  percent or some other mortality factors are
decreasing survival. We agree that the mortality fac-
tors identified by  other investigators  contribute to
lower survival, but an examination of natural life
history patterns may further explain the lower survival.


RELATION OF CURRENT LIFE
HISTORY TO SURVIVAL

Survival may  be  linked to life history patterns of
juvenile salmonids, since they  must pass through a
series of spatio-temporal "windows" during their life
history to maximize fitness (Miller and Brannon, 1981).
The emigration of juvenile chinook salmon from fresh-
water to the ocean can be viewed as a sequential use
of habitats (spatial  windows) at a time, size, or physio-
logical state (temporal windows) when they are best
adapted for that habitat. Juvenile fall chinook salmon
use four different habitats during  their  first year: (1)
habitats for dispersal from spawning areas, (2) fresh-
water riverine habitat, (3) estuarine habitat, and (4) the
ocean. Historically, entry into each of these habitats
occurred primarily  during the time window when the
likelihood of survival  was highest.
  The sequential use of habitat enables juvenile fall
chinook salmon to enter the next habitat and grow at a
faster rate. Dispersal  into freshwater rearing areas at
lengths of about 30-40 mm reduces the population
density  and competition during  freshwater rearing
(Reimers, 1973). Freshwater rearing enables juveniles
to mature  and  develop the osmoregulatory properties
needed  for saltwater entry.  Hypoosmoregulatory
capacity  is highest at lengths of about 80 mm, the
same approximate  size (90 mm) observed for estuary
entry among fall chinook salmon having the most suc-
                                                311

-------
 LAKE AND RESERVOIR MANAGEMENT
 cessful life history pattern (Clarke, 1982). Ability :o
 osmoregulate in seawater  is associated  with high
 Na+-K+  ATPase levels, but the ability  to  rapidly
 regenerate enzyme levels results  in high survival of
 fish 65 to 100 mm long when they are exposed to sea-
 water (Wagner et al. 1969; Zaugg, 1982; Gould et al.
 1983).  Since  hypoosmoregulatory capacity and  SLT-
 vival at estuary entry are optimal in fish about 80 to 90
 mm  long  these sizes  may  define  the  time window
 when survival at estuary entry is near the maximum. If
 fish enter the estuary at these lengths, they can begin
 growing at  an  increased rate. Growth of juveni e
 salmonids  in  freshwater  is  limited  by  food and
 temperature to  about one  third of their  potential;
 therefore,  the fish must  migrate to the estuary and
 ocean  to maximize growth  and,  by increasing size,
 reduce mortality (Brett, 1983). However, if freshwater
 rearing continues beyond the time window for estuaiy
 entrance,  growth  and survival  are reduced. The
 minimum length theory suggests that ocean entry can
 be successful when lengths  reach 100 to 120 mm,  but
 if freshwater rearing continues beyond  the time win-
 dow  for ocean  entrance, growth rate, length, and
 subsequent survival  are  further reduced, compared
 with fish entering the ocean.
   Fall  Chinook  salmon are adapted to enter each
 habitat within a specific time window of their  life
 history, but changes  in  the Columbia basin have
 modified  life history patterns. Life histories   under
 pristine conditions  suggest  that survival  of  fall
 chinook salmon was highest for fish  entering  the
 estuary at mean lengths of 80-90 mm and the ocean at
 a minimum length of 100-120 mm. Reservoir rearing of
 small juvenile salmon is beneficial, since freshwater
 or estuarine growth is prerequisite to the attainment
 of a minimum length necessary for high ocean sur-
 vival. However, the slow migration rates of juvenile  fall
 chinook salmon in reservoirs prohibits the sequential
 entry to each habitat during the appropriate time win-
 dow.  As a result, residence  in reservoirs beyond the
 time required  to grow to lengths of 100-120 mm may
 increase mortality in  freshwater and metabolic  costs
 of seawater adaptation, without the survival benefits
 of increased growth in the ocean.
MANAGEMENT ALTERNATIVES

To effect changes in the critical relation of time, sizei,
and physiological windows, managers can manipulate
time and size at release of hatchery fish; travel rates
may be increased  by flow manipulation and transpor-
tation of juveniles. Three alternatives combining these
factors should be considered: (1) release subyearlings
in spring, provide  flows to  either speed  movement
through the entire system  or until fish enter a collec-
tion facility and can be transported; (2) extend rearing
of subyearlings in hatcheries from spring to summer
or fall; at release, supply  either adequate flows  and
transportation or only adequate  flows; or (3) extend
hatchery rearing to the yearling stage and release fish
before the spring freshet.
  The  first  alternative  (short-term  rearing  of sub-
yearlings) would  minimize hatchery costs by using
reservoirs to rear fall chinook salmon until they reach
a minimum  size for saltwater  entry and  positioning
them for entry into this new  habitat. Wild fish  would
also benefit from increased flows.
  The  second alternative  (long-term rearing of sub-
yearlings) should benefit hatchery fall chinook salmon
by increasing survival  benefits associated  with  in-
 creased  size.  By  providing  adquate  flow,  and  if
 necessary  transportation,  fish would move  quickly
 from  freshwater  to the estuary.  Wild  fall  chinook
 salmon would  benefit from decreased passage and
 reservoir mortality associated with higher flows and
 transportation.  However, providing  high  flows during
 summer and fall  may be prohibitively expensive. In-
 creased hatchery costs resulting from extending rear-
 ing would also  be undesirable.
   Benefits associated with the third alternative (rear-
 ing to yearling stage), include higher survival resulting
 from the larger size of migrants and increased travel
 rates associated with larger fish and higher flows dur-
 ing spring.  Increased flows during spring may benefit
 early wild migrants but special provisions would be
 necessary for later migrants. Negative factors  include
 a significant increase in  costs and  logistic problems
 for hatcheries.
REFERENCES

Becker, D.C. 1973. Food and growth parameters of juvenile
  chinook salmon, Oncorhynchus tshawytscha, in  central
  Columbia River. U.S. Natl. Mar. Fish. Serv. Fish. Bull 71-
  387-400.

Brett, J.R. 1983. Life energetics of sockeye salmon, Oncor-
  hynchus nerka. Pages 29-63 in W.P. Aspey and S.I. Lustick,
  eds. Proc. Behavioral Energetics: The Cost of Survival in
  Vertebrates. Ohio State Univ.  Press. Columbus.

Clarke, W.C. 1982. Evaluation of  the seawater challenge test
  as an index of marine survival. Aquaculture 28: 177-83.
Dawley, E.M., R.D. Ledgerwood, T.H. Blahm, and A.L Jensen.
  1982. Migrational characteristics and  survival of juvenile
  salmonids entering the Columbia River estuary. U.S. Natl.
  Mar. Fish. Serv., Northw.  and  Alaska Fish. Cent.  Seattle
  Wash.

Dawley, E.M., et  al.  1980. A study to define the migration
  characteristics of chinook and  coho salmon and steelhead
  in the Columbia River estuary. U.S. Natl. Mar. Fish. Serv.,
  Northw. and Alaska Fish. Cent., Seattle, Wash.

Fowler, L.G., and J.L. Banks. 1980. Survival rates of three
  sizes of hatchery reared fall  chinook salmon. U  S Fish
  Wildl. Serv. Tech. Ser. 80-1.

Fulton, LA. 1968. Spawning areas and abundance of chinook
  salmon (Oncorhynchus  tshawytscha) in the Columbia
  River basin—past and present. U.S. Fish Wildl. Serv. Spec.
  Sci. Rep. Fish. 571.

Gould, R.W., et al. 1983. Seawater acclimation of premigra-
  tory  (presmolt)  fall chinook salmon: a  possible new
  management strategy? U.S. Natl. Mar. Fish. Serv., Northw.
  and Alaska Fish. Cent., Seattle, Wash.

Krcma, R.F., and  R.F. Raleigh.  1970. Migration of juvenile
  salmon and trout into Brownlee Reservoir, 1962-65. U.S.
  Fish Wildl. Serv., Fish. Bull 68: 203-17.

Miller, R.J., and E.L. Brannon 1981. The origin and develop-
  ment of life history patterns in pacific salmonids. Pages
  296-309 in E.L. Brannon and E.G. Salo, eds. Proc. Salmon
  and Trout Migratory Behavior Symp. Univ. Washington,
  Seattle.

Parente, W.D., and J.G. Smith. 1981. Columbia River back-
  water study: Phase II. U.S. Fish. Wildl. Serv. and Columbia
  River Inter-Tribal Fish Comm., Vancouver, Wash.

Park, D.L.  1969. Seasonal changes in downstream migration
  of age-group 0 chinook  salmon in the upper Columbia
  River. Trans. Am. Fish. Soc. 98: 315-17.

Raymond, H.L 1979. Effects of dams and impoundments on
  migrations of juvenile chinook salmon  and steelhead from
  the Snake River,  1966 to  1975. Trans. Am. Fish Soc 108"
  505-29.
                                                  312

-------
Reimers, P.E. 1973. The length of residence of juvenile fall
  Chinook salmon in Sixes River, Ore. Res. Rep. Fish Comm.
  Ore. 4: 2-43.
Reimers, P.E., and G.L Concannon. 1977. Extended resi-
  dence of hatchery-released juvenile fall Chinook salmon in
  Elk River, Ore. Ore. Dep. Fish Wildl., Res. Sect. Info. Rep.
  Sen, Fish. No. 77-2.
Reimers, P.E., and T.W. Downey. 1982. Population dynamics
  of fall Chinook salmon in Sixes River. Ore. Dep. Fish Wildl.,
  Res. Develop. Sect. Annu. Prog. Rept., Oct. 1, 1979-Sept.
  30, 1980.
Rich, W.H. 1922. Early history and seaward migration  of
  Chinook salmon  in the  Columbia and Sacramento rivers.
  Bull. U.S. Bur. Fish. 37: 1-74.
Schoeneman, D.E., R.T. Pressey, and C.O. Junge,  Jr. 1961.
  Mortalities of downstream  migrant  salmon  at  McNary
  Dam. Trans. Am. Fish. Soc. 90: 58-72.

Sims, C.W. 1970. Emigration  of  juvenile salmon and trout
  from  Brownlee Reservoir, 1963-65. U.S. Fish Wildl. Serv.,
  Fish.  Bull. 68: 245-59.

Sims, C.W., and D.R. Miller.  1982. Effects of flow  on the
  migratory behavior and  survival of juvenile fall and sum-
  mer Chinook salmon in John Day Reservoir. U.S. Natl. Mar.
  Fish.  Serv., Northw. and Alaska Fish. Cent., Seattle, Wash.
                                 FISHERY MANAGEMENT

Sims, C.W., W.W. Bentley, and R.C. Johnsen. 1977. Effects of
  power peaking operations on juvenile salmon and steel-
  head  trout migrations—progress 1976. U.S.  Natl. Mar.
  Fish. Serv., Northw. and Alaska Fish. Cent., Seattle, Wash.
	.  1978. Effects of power peaking operations  on
  juvenile salmon and steelhead trout migrations—progress
  1977.  U.S. Natl. Mar. Fish. Serv., Northw. and Alaska Fish.
  Cent., Seattle, Wash.
Thorpe,  J.E., and R.I.G. Morgan. 1978. Periodicity in Atlantic
  salmon Salmo salar L. smolt migration. J. Fish Biol.  12:
  541-8.
Wagner, H.H., P.P. Conte, and J.L Fessler. 1969. Develop-
  ment  of osmotic  and ionic regulation  in two  races of
  Chinook salmon  Oncorhynchus  tshawytscha.  Comp.
  Biochem. Physiol.  29: 325-41.
Zaugg, W.S. 1982. Some changes in smoltification and sea-
  water  adaptability of salmonids resulting from environ-
  mental and other factors. Aquaculture 28: 143-51.
                                                        313

-------
 THE EXPANSION OF THE WHITE PERCH, MORONE AMERICANA,
 POPULATION IN LAKE ANNA RESERVOIR, VIRGINIA
 A. C.  COOKE
 Environmental Laboratory
 Virginia Electric  and  Power Company
 Mineral,  Virginia
             ABSTRACT

             The expansion of the white perch, Morone americana, population in Lake Anna Reservoir, Va., is
             documented by examining data resulting from several collection methods. White perch data were derived
             from annual cove rotenone studies, bimonthly gill net and shoreline electroshock surveys, semiweek-
             ly intake screen impingement studies and seasonal entrainment and ichthyoplankton tow surveys.
             The length of the sampling period ranged from 2 to 7 years depending on how long each method
             had been employed on the Reservoir and whisn white perch appeared in each collection. An attempt
             is made to determine the impact of this expansion upon other species of fish in the reservoir and
             to address future implications.
 The objective of the Lake Anna fish survey program is
 to monitor fish populations to determine if changes
 occur (percent of standing crop, species composition,
 etc.), and, if so, reasons for these changes. These in-
 tensive studies will continue at least through 1985.
   The white perch, Morone americana (Gmelin), is a
 euryhaline species commonly found along the Atlan-
 tic Coast.  Where it has been introduced into fresh-
 water,  population  explosions have  been common,
 often at the expense of native species (Dence, 1952;
 Scott and Christie, 1963; Hergenrader and Bliss, 1971;
 Christie, 1972). White perch were first documented n
 Lake Anna, a power station cooling reservoir, in 1973,
 just 1 year after the lake was filled. Although fish
 population studies  continued during the following
 years, white perch were not collected again until 1976.
 It is not known how this species entered the lake.
  The white perch population has increased dramati-
 cally in the Reservoir since 1976 (monitored by  on-
 going  adult fish  and  ichthyoplankton  survey  pro-
 grams). Annual cove rotenone  surveys, bimonthly  gill
 net and shoreline electrofishing surveys and biweekly
 screen impingement  samples  provide estimates  for
 juvenile and  adult  fishes while  seasonal  screen
 entrainment and ichthyoplankton tow surveys provide
 data for larval and post larval fish estimates.
STUDY SITE

Lake Anna is a 5,264 ha manmade lake formed to pro-
vide condenser cooling  water  for the North Anna
Nuclear Power Station, a two-unit facility with a total
rated  capacity  of  1,755  megawatts. This  lake is
located in Louisa, Spotsylvania, and Orange Counties
in the Piedmont area of Virginia and was formed h
1972 by the impoundment of a portion of the North An-
na River.  Lake Anna consists of two waterbodies, a
3,887 ha reservoir and a 1,377 ha Waste Heat Treat-
ment  Facility (WHTF).  The WHTF  receives  and
dissipates the heated discharge from the power sta-
tion and is separated from the reservoir by a series of
three earthen dikes, eventually returning the cooled
water to the reservoir through a submerged opening in
the last dike.
  This study will address only conditions in the reser-
voir. The reservoir is approximately 27 km long with
438 km of shoreline. It has a volume of approximately
3.0  x  108  m3 at normal pool  elevation. The mean
depth of the reservoir is approximately 8 m with a max-
imum depth of 24 m.
MATERIALS AND METHODS

The population increase of white perch in Lake Anna
reservoir was  monitored  by several  techniques at
various stations throughout the reservoir (Fig. 1).

Rotenone

Cove rotenone samples have been taken during the
month of August each year since 1976. Four coves are
sampled  in the reservoir, two upper, one middle, and
one lower reservoir. The coves range in size from 0.631
ha to 0.874 ha. They are blocked with a 75 m x  7.5 m
x 6.2 mm block net secured on the cove bottom and
checked  by divers.  A mark-recapture procedure tests
the effectiveness of the  rotenone and efficiency of
pickup. One hundred  to  150 fish are identified to
species,  measured, fin-clipped, and  released  inside
the block net prior to the addition of rotenone to each
cove. Shortly after the marked fish are released (5 to
10 minutes) liquid rotenone is mixed with lake water
and pumped into the cove through a 6 m long hose in
an attempt to achieve a concentration of one part per
million. A curtain of rotenone is laid down, surface to
bottom, by raising and lowering the hose as the boat
proceeds slowly throughout the cove, special atten-
tion being given to feeder streams and beaver lodges.
  Fish are collected and transported  to the laboratory
where they are separated by species and length class,
enumerated and weighed. Marked fish are measured
individually and recorded. Collection and laboratory
treatment are repeated on the second day with the ex-
ception of weights which are not recorded  because of
decomposition, but, estimated from the previous day's
results. On the afternoon  of the second day, after a
                                               314

-------
                                                                                 FISHERY MANAGEMENT
final cove cleanup, the net is transported to the next
cove.
  Previous rotenone studies on Lake Anna indicated 2
days are sufficient for completion of a rotenone cove
survey  during  the  month  of  August as higher
temperatures increase the toxicity of rotenone to fish
and also hasten detoxification.
  Rotenone  standing crop estimates are based on a
tag  return percentage  calculated   from  the mark-
recapture technique and cove surface area. Species
weights and  numbers are multiplied  by this correction
factor to obtain estimated standing crops (kg/ha).
Gill Net

Experimental gill nets have been used in Lake Anna on
a regular basis since 1976. Nets are set bimonthly at
five reservoir stations on 2 consecutive days, two and
three stations each day. Samples are taken in shallow
water during spring and above the thermocline during
summer. Experimental gill nets are 91.44 m in length
and 1.83 m deep. Nets  consist of  six panels,  each
15.24 m long with bar mesh sizes ranging from 1.27 cm
to 7.62 cm in 1.27 cm increments. Nets are set during
the afternoon  of one day and fished approximately 20
hours later. Data are compared between years on the
basis of catch per net day.
Electrofishing

Bimonthly shoreline electrofishing samples have been
collected in the reservoir at seven stations since 1981.
During 1980, samples were collected  sporadically
throughout the reservoir. Fish are collected during
daylight hours by shocking a 100 m stretch of shore-
line  using  a  Type  VI-A  Smith-Root  electrofisher
operating at 1,000 volts, 3-4 amps and 60 D.C. pulses
per second.  This unit is operated from a 4.9 m boat.
Data are compared between years on  the basis of
catch per electrofishing station day.
Impingement

North Anna Power Station condenser cooling water is
withdrawn from Lake Anna by a series of circulating
water pumps (4 C.W.P./unit) each rated at a capacity of
15 m3/sec. The cooling water is filtered by a single
rotating traveling screen (9.5 mm bar mesh)  in front of
each C.W.P. to prevent clogging of pumps and con-
denser tubes by fish and miscellaneous debris. During
C.W.P. operation, fishes too large to pass through the
intake screens are trapped  (impinged)  against the
screens and subsequently removed by a spray wash
system.
  The travelling screens are sampled on a 4 week cy-
cle with two 24-hour samples being collected on non-
consecutive days each week for the first 3 weeks. Dur-
ing the fourth week, a composite sample is taken con-
sisting of 12 continuous 2-hour samples. Each screen
is washed for a minimum of 10 minutes to insure com-
plete fish removal. All operable screens are washed
when  the corresponding C.W.P. is in operation. The
fish are washed into a catch basket at  the end of a
sluiceway, removed and transported to the laboratory
for analysis. Impingement rates have been monitored
since the station began generating electricity in 1978.
To compare yearly results the total number of fish im-
pinged, by species, is divided by the average  number
of  screens operating that year to obtain per-screen im-
pingement rates. Total numbers impinged are also ex-
amined between years.

Entrainment

Entrainment samples are samples  of  the ichthyo-
plankton which pass through the traveling screens
and eventually out through  the station discharge.
Samples have been collected once per week during
the spring  and early  summer since  1976  (March
through July). Samples are taken four times per day
(0600, 1200, 1800, and 2400 hours) at the  surface, mid-
depth, and bottom. Samples are collected by a set of
paired conical nets, placed in a predetermined intake
forebay in front of the traveling screens for 10 minutes
                                    RAM UN KEY ARM  (RAM)
                                                                         •  <3ILL NET  (7)

                                                                         *  ELECTROFISH (9)

                                                                         -  ROTENONE  (4)

                                                                         *  ICHTHYOPLANKTON TOW (6)
  NORTH ANNA ARM (NAR)
                        NORTH ANNA POWER STATIC
                               WASTE HEAT TREATMENT FACILITY

Figure 1.—Lake Anna adult fish and ichthyoplankton sampling stations.
                                                315

-------
 LAKE AND RESERVOIR MANAGEMENT
 per depth. The conical nets measure 0.5 m x  1.5 m
 and are of 505 \i mesh. The volume of water filtered is
 determined by use of large-vaned, low-velocity-sensi-
 tive digital flowmeters. Results are noted as number
 of larvae and/or eggs per 1,000 m3 water entrained.

 Ichthyoplankton Tows

 Since 1979,  ichthyoplankton  collection tows  have
 been taken during the spring and  summer months.
 Samples are collected weekly at three stations (upper,
 middle, and lower resevoir). At each station an oblique
 tow is made  in open water, 6-18 m deep, and a cove
 tow along the adacent shoreline. The oblique sample
 consists of a 5-minute stepped-oblique tow (4 m to sur-
 face; 1 minute per meter) while a cove tow sample con-
 sists of a 2-minute surface tow.
   Samples are collected by means of a side-towed net
 (0.38 m square, 1.5 m long, 505 n mesh). A digital flow-
 meter is  used to  determine  the volume  of  water
 filtered. Results are  recorded as  number  of  larvae
 and/or eggs  per  1,000  m3, and compared  between
 years.


 RESULTS AND  DISCUSSION

 The validity of rotenone as a method of determining
 fish populations in lakes has been questioned in re-
 cent years. It appears to work well for some species
 and poorly for others (Carter, 1958; Hayne et al. 1968;
 Aggus et al. 1980). Annual cove rotenone studies do,
 however, provide valuable  data on changes in fish
 populations when conducted in the  same coves over
 several years as repopulation of fished areas is rapid
                             (Sandoz, 1959;  King et  al. 1981). Lake Anna annual
                             cove rotenone results have shown an increasing white
                             perch population with the largest increase occurring
                             between 1981 and 1982 (Table 1).
                                During the 1976 study,  primarily  young-of-the-year
                             white perch were collected as no individuals were>
                             127.0 mm total  length.  The white  perch  population
                             made up only 0.02 percent of the estimated reservoir
                             standing crop. In 1977 white perch comprised 0.17 per-
                             cent of the estimated total standing crop and 0.3 per-
                             cent of the estimated total number. The  percent of
                             total weight has gradually  increased  through 1981
                             while the percent of total  number remained fairly
                             steady.  Both  of  these  percentages  increased
                             significantly during 1982 studies. Estimated percent
                             of total  standing crop during 1982 (18.2 percent) was
                             three times the 1981 value (5.26 percent)  and much
                             greater than found in 1977  (0.17 percent). Mean weight
                             decreased  somewhat during 1982  as  more smaller
                             (younger) fishes were collected. This is indicative of
                             an expanding population with increasing recruitment.
                               Sampling by impingement on power station intake
                             screens also has its inherent biases as it samples
                             primarily pelagic, planktivorous species. However, it is
                             a valuable sampling tool when yearly results are com-
                             pared for the principally sampled species to determine
                             changes in their population  levels. Results from this
                             sampling method at  North Anna also indicate an in-
                             creasing white perch population with the greatest in-
                             crease occurring during 1981 and 1982 (Table 2).
                               Impingement results for  1978 are incomplete as the
                             station did not begin operation until June and,  al-
                             though screen sampling began well  before that date,
                             the first several months of 1978 were missed. These
 Table 1.—White perch collected by cove rotenone sampling in Lake Anna 1976-82: length-frequencies, estimated number
	collected per hectare, percent of total fkihes collected, percent of total weight collected.

__	Percent collected lor each size class by date

 Length frequencies         1976        1977        1978        1979        1980         1981         1982
 (mm) T.L.
0-101.6
101.7-127.0
127.1 - 152.4
152.5-177.8
177.9 - 203.2
203.3 - 228.6
228.7 - 254.0
254.1 - 279.4
77.6 84.30
22.6 1.30

8.20
3.10
3.10


88.80

1.80
6.60
1.90
0.60
0.30

71.80

14.10
6.50
5.20
2.00
0.40

56.30
2.70
13.30
21.00
5.90
0.80


12.00
28.10
18.20
26.40
13.00
1.50

0.80
15.0
16.5
36.3
22.7
9.2
0.3


   Estimated total
 number per hectare

   Percent of total
  number collected

   Percent of total
  weight collected
31.0
159.00
            0.30
            0.17
                      725.00       894.00
             2.50
                        0.73
                                   2.40
                                   1.68
                                             1028.00
                                      1.80
                                                                         1.78
                                                1290.00      3209.0
                                                             2.20
                                                             5.26
                                                               7.9
                                                                                                 18.0
 Table 2.—White perch collected by screen impingement at North Anna power station 1978-82: length frequencies, actual
     number collected, percent of total fishes collected and estimated number impinged per screen during the year.

                                Percent Length Frequencies (mm total length)
Year
1978
1979
1980
1981
1982
0-99
12.2
40.3
14.4
4.3
100 • 149
27.3
15.5
19.3
22.2
150 - 199
45.0
30.5
58.2
67.5
200 • 249
13.5
12.5
8.0
6.0
250 +
1.9
1.2
0.1
0.0
Actual
Total
8
311
174
613
1312
%of
Total
0.2
0.6
1.9
7.7
Est. No. Per
Screen Year
240
118
412
1234
                                                 316

-------
                                                                                   FISHERY MANAGEMENT
late winter months (January, February, and March) are
when traditionally large numbers of white perch are
impinged.
  During 1979, white perch comprised 0.2 percent of
the fishes  impinged  with  a total estimated  mean
number per screen of 240 fish. This mean number per
screen decreased during 1980 to 118 but the percent
of white perch in the total catch was 0.6 percent in
1980 compared to 0.2 percent in 1979.
  During 1981 white perch increased to 1.9 percent
and the estimated per screen total increased to 412.
These numbers increased again during 1982 to 7.7 per-
cent and 1,234 respectively.
  There was a drop in the number of small (0-99 mm
total  length) white perch collected during 1982 imp-
ingement sampling which would appear to contradict
the increase in this size class suggested by rotenone
results. This discrepancy may be due to the fact that
August rotenone samples collect young-of-the-year
white perch  spawned in April while the peak impinge-
ment for white  perch  is in January, February, and
March.  A good  growth year should boost  most of
these young-of-the-year fish into the next size class,
which did, in fact, show an  increase in 1982 over 1981
data.
               The entrainment of fish eggs and larvae is selective
             for broadcast spawning species whose larvae normal-
             ly frequent open water. Again, over a period of years,
             changes in certain populations should be reflected in
             entrainment  data. At North Anna these data show in-
             creasing numbers of white perch larvae  per 1,000 m3
             entrainment  since  1978 with the  largest  increase oc-
             curring between 1980 and 1981 (Table 3). Entrainment
             sampling during 1978 collected only a few white perch
             larvae and then only during 2 weeks. In 1979, both
             numbers of larvae collected per 1,000 m3 and numbers
             of weeks during which they were collected increased.
             In 1980, both of these parameters increased again.
               The largest increase in number per 1,000 m3 occur-
             red during 1981  when the mean total was four times
             greater  than in  1980 although larvae were  collected
             during 9 weeks, 1 week less than in 1980. During 1982,
             white perch  larvae comprised 31.5 percent of the total
             larvae collected and the number per 1,000 m3  also in-
             creased over 1981. Larvae were  primarily  collected
             during 8 weeks  since the June collection consisted of
             one larger individual. Also, 1982 data show a single
             peak distribution rather than the  bimodal  peaks of
             1981, perhaps indicating a better sampling of existing
             larvae.
   Table 3.—White perch larvae collected by entrainment at North Anna power station 1978-82: percent of white perch
                                   larvae collected per 1,000 cubic meters.
      Total

    % of total
  larvae collected
                         Week
1978
                                                      1979
                                         7.0
 0.2
                                                      142.0
                4.3
                            1980
                                                                    217.0
                              7.2
                                                                                  1981
                                          883.0
                                           22.8
                                                         1982
April -



May -




June -



July-
1
2
3
4
5
1
2
3
4
5
1
2
3
4
1


2.8
2.1
6.3
42.9 33.8
18.3
8.5
28.2
57.1






1.4
7.8
19.4
19.4
13.8
20.3
8.3
6.9
1.8
0.9



0.3
7.9
16.1
30.2
16.6
19.7
6.4
1.5
1.3






7.5
7.5
26.3
43.1
11.6
0.3
*
3.1
0.3


0.3
                                                                                                970.0
                                                         31.5
 Table 4.—White perch larvae collected by ichthyoplankton tows in Lake Anna 1979-82: percent of total number white perch
                   larvae collected per 1,000 cubic meters. Cove and open water tows combined.

April -



May -



June -



Week
1
2
3
4
1
2
3
4
1
2
3
4
1979


67
17.4
13.0
22.1
25.8
10.2
3.2
1.6


1980


13.0
43.0
32.9
8.6
2.5





1981


0.5
57.9
6.5
32.1
2.0
0.6
0.3
0.1


1982
0.2
0.0
1.1
21.8
19.0
36.8
18.4
1.7
0.8
0.2


      Total per
  1,000 cubic meters

      % of total
   larvae collected
     575.0
       1.5
                     3670.0
                        7.4
                                      16745.0
                                         15.0
6957.0
   7.0
                                                  317

-------
 LAKE AND RESERVOIR MANAGEMENT
   Ichthyoplankton tow data also suggest a major in-
 crease between 1980 and 1981 (Table 4). The 1981 data
 set contains averages of three stations per week while
 1980 data represent only one station per week. Small
 variations would, therefore, be normal. However, the
 between-year differences of more than a fourfold in-
 crease (1979-80 and 1980-81) should represent a real
 increase of considerable magnitude.
   Larvae were also collected during 3  more weeks in
 1981 than in 1980. During 1982 less than one half as
 many white perch  larvae were  collected and they
 represented about 50 percent of the 1981 total.  The
 1981 data show a large number of larvae collected dur-
 ing one early week (fourth week of April) followed by a
 rapid decline. This probably represents the collection
 of one or more  schools of white  perch  larvae  and
 therefore is not truly indicative of actual abundance  If
 this peak  is eliminated, 1981 results are more com-
 parable to 1982 results. The 1982 tow data establish a
 more normal curve and therefore, perhaps, a more
 representative sampling result. In any case, tow data
 also  show an increasing population of white perch
 since 1979.
  Although shoreline  electrofishing has been  con-
 ducted in the reservoir only since 1980, the threefold
 increase in 1982 data over  1981  appears significant
 (Table 5).  During  sporadic  1980  sampling no white
 perch were collected by this method. During 1981, five
 were collected from 42  sampling station days. These
 fish  represented 0.1 percent of the total number  col-
 lected and 0.2 percent of the total weight. In 1982, the
 actual numbers of white perch collected more than
 doubled (to 13)  from  39  sampling  station  days,
 representing 0.3 percent of the total number and 0.7
 percent of the total weight collected. White perch do
 not normally frequent shoreline areas unless structure
 is evident (Jones et al. 1978), so it is normal that this
 species would comprise only a small  percentage of
 the total electrofishing  catch and still  be a predomi-
 nant species in the lake.
  The gill net survey results are the only data that do
 not support the other suggestions of  an  increasing
 white perch population. Gill  net catches instead sug-
 gest a steady state situation since 1979 and especial-
 ly do not show an increase between 1981  and 1982
 (Table 6). White perch were first collected by gill  not
 during 1976. The percent of total number figure in-
 creased  through  1979 and  then  decreased  slightly
 each  year through 1982. The  percent total weight
 figure increased through 1981 and decreased during
                                1982. Both values showed a marked drop during 1982
                                in contrast with rotenone results and other data.
                                   Gill net results have been reliable indicators of ex-
                                panding  white  perch populations  in  other  areas
                                (Hergenrader and Bliss, 1971). However, in  Lake Anna,
                                gill net results do not represent the real situation, in
                                view of the other sampling method results.
                                   The increase in the white perch population can be
                                better  shown graphically  (Fig. 2). The  four major
                                sampling methods graphed are cove rotenone, screen
                                impingement,  screen entrainment, and ichthyoplank-
                                ton tow. This figure indicates the large increase in lar-
                                vae collected during 1981 and the subsequent large in-
                                crease in  1982 of intermediate size perch collected.
                                This increasing trend should continue through 1983,
                                although probably less dramatically  in view of 1982
                                ichthyoplankton results.
                                  It  should appear  from  these  results  that white
                                perch, originally almost nonexistent  in Lake  Anna,
                                have expanded  to  become one  of  the dominant
                                species present. This type of increase after planned or
                                accidental introduction of this species has also occur-
                                red in other areas of the country. An introduction of
                                white perch in  Cross Lake, N.Y., led to a population of
                                considerable size in a short time and it was suggested
                                that, because of its adaptability, introduction of white
                                perch into lakes may produce undesirable changes in
                                native fish populations (Dence, 1952). This actually oc-
                                curred  in  Wagon Train Reservoir, Neb.,  where the
                                white perch expansion was concomitant with the
                                decline  of a  previously dominant  species,  black
                                bullhead (Ictalurus nebulosus). White perch increased
                                from an insignificant to a dominant species in 2 years
                                (Hergenrader and Bliss, 1971).
                                  Since 1977, the increase in Lake Anna's white perch
                                population has been  accompanied by a decrease in
                                the black crappie (Pomoxis nigromaculatus) popula-
                                tion. Black crappie comprised  15.0  percent  of the
                                reservoir standing crop  in 1976 (from rotenone survey)
                                but declined to 0.4 percent by 1981. This decrease is
                                probably not directly related to white perch  as the ma-
                                jor decreases  occurred during 1976 and  1977 when
                                white perch still comprised an insignificant portion of
                                the standing crop.
                                  It is possible, however, that white perch filled the
                                void created by natural fluctuations of the black crap-
                                pie population, which occur frequently (Swingle and
                                Swingle, 1968) and cannot now be easily displaced. An
                                increase in the black crappie percentage of standing
                                crop occurred during 1982 (1.9 percent) but this may be
 Table 5.-White perch collected by shoreline electroshocking in Lake Anna 1980-82, catch per shocking day per station
	Number, percent of total fishes collected. Weight, percent of total weight collected.
Year
                      Number
                                             % Total
                                                                     Weight (g)
                                                                        % Total
1980
1981
1982
  0.0
  0.1
  0.3
  0.0
  0.1
  0.3
    0.0
    9.5
   18.3
  0.0
  0.2
  0.7
   Table 6.—White perch collected by gill net in Lake Anna 1976-82, catch per net day per station. Number percent of
  	total fishes collected. Weight, percent of total weight collected.
Year

1976
1977
1978
1979
1980
1981
1982
Number


   0.2
   2.6
   8.6
   9.5
   9.4
   6.7
% Total

  D.1
  3.7
  3.5
  33.5
  23.2
  22.3
  115.0
Weight (g)


    0.1
    0.2
    0.5
    0.6
    0.5
    0.3
% Total

  0.1
  0.5
  7.8
  9.3
  9.7
  5.0
                                                 318

-------
                                                                                       FISHERY MANAGEMENT
PERCENT
1100
1000 -
900 -
800 -
700
600 -
500
400
300 -
200
100
0 -
- 1 0 0 -
19
_,B
.--"'
-^
^xa-~- -B-
^
_.'-lr~"
.--•'
.---"



y^
—-^
•j- —
!
] rr-i-i T-I i i i ["'! i i i ii i i r; \ i i r Tf i i i j i i i i i i i'"i"i j ' i i i t \ i 'i 'i [ i i i" < i i > • < \ '• ' "'"'-t -< i . • 7 ' i ' • • ' ' ' T; • . i ......
77 1978 1979 1980 1981 1 982 1979 1930 1981 19E
TEAR
LEGEND: GEAR * * * IMPINGEMENT » o o ENTPA INMEtiT
a a o ROTE NONE i i i TOW
Figure 2.—Cumulative percent of yearly changes in collection numbers of white perch from Lake Anna Reservoir 1977-1982.
a  result of a  decrease in fishing pressure on this
species (which was indicated by creel surveys) rather
than  an actual increase in the population. Average
weight was the same  both years and an  increasing
population should show a decreasing average weight
from  year to  year as more small  fish  are  being
recruited into the population.
   There are several reasons why this replacement of
black crappie by white  perch may be feasible. Studies
have  indicated that black crappie larvae feed on the
same organisms as newly hatched larvae of yellow
perch, pumpkinseed,  and  bluegill, but sequentially,
with the resource being used by different  species in
turn as the larvae appear (Keast, 1980).
   White perch larvae would presumably feed on the
same prey items as black crappie and at the  same
time,  as their spawning seasons overlap at  Lake Anna
with white perch spawning earlier. Larger white perch,
up to 200  mm total length may compete  with  black
crappie for food, as insects are the primary prey for
both (Reid, 1972; laboratory stomach analysis). As fish
make up a large percentage of the diet of white  perch
>200 mm  total length (Reid, 1972) it is also possible
this fish would prey on crappie fry.
   Should white perch continue to replace  black crap-
pie in Lake Anna the outlook may not be entirely bad.
Fishing pressure is very light on white perch at this
time as they have not yet been recognized  as a  game
fish by local anglers. This species is well regarded by
anglers in other waters  associated with  power sta-
tions  (Moore and Fisbie, 1972) and may replace  black
crappie also as a pan fish if white perch achieve their
growth  potential and  do not stunt from overpopula-
tion.
REFERENCES

Aggus, L.R., et al. 1980. Barkley Lake Symposium, evaluation
  of standing crops of fishes in Crooked Creek Bay, Barkley
  Lake, Ky.  Proc. Ann. Conf. Ass. Fish Wildl.  Agencies
  33:710-22.
Dence, W.A. 1952. Establishment of white perch, Morone
  americana, in central New York. Copeia 3:200-1.

Carter, B.T. 1958. What significant information can be gained
  from rotenone population studies in impoundments. Proc.
  11th Annu. Conf. S.E. Ass. Game Fish Comm. (1957). 83-4.

Christie, W.J. 1972. Lake Ontario: Effects of exploitation,
  introductions,  and entraphication on the salmonid com-
  munity. J. Fish. Res. Board Can. 29:913-29.
Hayne, D.W., G.E. Hall, and H.M. Nichols. 1968. An evaluation
  of cove sampling of fish populations in Douglas Reservoir,
  Tenn. Pages 244-97 in Am. Fish. Soc. Reservoir Fish.
  Resour. Symp., Athens, Ga.
Hergenrader, G.L., and Q.P. Bliss. 1971. The white perch in
  Nebraska. Trans. Am. Fish. Soc. 100:734-8.

Jones, P.W., F.D. Martin, and J.D. Hardy, Jr. 1978. Develop-
  ment of fishes of the mid-Atlantic bight. Center Environ.
  Stud. Univ. Maryland. #783 U.S. Fish Wildl. Serv.
Keast, A. 1980. Food and feeding relationships of young fish
  in the first weeks after the beginning of exogenous feeding
  in LakeOpinicon, Ontario. Environ. Biol. Fishes 5(4):305-14.
King,  T.A., J.C. Williams, W.D. Davies, and  W.L Shelton.
  1981. Fixed versus random sampling of fishes in  a large
  reservoir.  Trans. Am. Fish. Soc. 110:563-8.
Moore, C.J.  and  C.M.  Frisbie. 1972. A winter sport  fishing
  survey in a warm water discharge of a stream electric sta-
  tion on the Patuxent  River,  Maryland.  Chesapeake Sci.
  13(2):110-15.
Reid,  W.F., Jr. 1972. Utilization of the crayfish Orconectes
  Limosus as forage by white perch (Morone americana) in a
  Maine Lake. Trans. Am. Fish. Soc. 101:608-12.
Sandoz, V.  1959. Changes in  the fish  population of Lake
  Murray following the reduction of gizzard shad numbers.
  Proc. Okla. Acad.  Sci. 37:174-81.
Scott, W.B.,  and W.J. Christie. 1963. The invasion of the lower
  Great Lakes by the white  perch,  Roccus  americanus
  (Gmelin). J. Fish. Res. Board Can. 20(5):1189-95.
Swingle, H.S., and W.E. Swingle. 1968. Problems in dynamics
  of fish populations  in reservoirs.  In Am. Fish. Soc. Res.
  Fish. Resour. Symp., Athens, Ga.
                                                    319

-------
  CATCH  COMPOSITION AND POTENTIAL IMPACT OF BAITED AND
  UNBAITED COMMERCIALLY FISHED HOOP NETS IN THREE CENTRAL
  FLORIDA LAKES
 MARTY M. HALE
 JOE E. CRUMPTON
 DENNIS J. RENFRO
 Florida Game and Fresh Water Fish Commission
 Fisheries Research Laboratory
 Eustis, Florida
            ABSTRACT

            From June 1980 through March 1983, 528 loop nets fished in Lake George, Little Lake George
            and Crescent Lake, Fla. were observed, representing 4,356 net-days of fishing. Commercially im-
            portant species comprised 66.1 percent ol the total catch by number while game fish species

             «H !fnrllf    P6frCK  ° L^ t?t2l CatCh and 23'1 percent of a" 9arne fish Cau9ht. ln"ial mor-
            tality for all game fish caught in lake hoop nets was 0.4 percent. Hoop nets baited with blueback
            herring, Alosa aestivalis, or soybean chips caught more commercially important species (4 5/net-
            n.ty> /i w utnbai'ed nets (2.3/net-day) and 'ewer game fish species (0.5/net-day) than unbaited
            nets (1.6/net-day). Commercially important species comprised 57.7 percent of the total catch in
            unbaited nets and 89.3 percent of the catch in baited nets. Game fish species comprised 41 5
            percent of the catch in unbaited nets and 10.5 percent of the total catch in baited nets Juvenile
            black crappie Pomoxis nigromaculatus (< 22.9 cm) comprised 83.0 percent of all game fish
            caught in unbaited hoop nets while bluegill comprised 74.6 percent in baited nets. The catch rate
              ,   ^IZf Hn°nn9ame  Sn  Caugnt in ^'^ and unbaited nets combined was  1.4/net-day  An
            estimated 10-20 game fish/hectare/year were caught in commercial nets during the study Based
            on our knowledge of standing crop data frcm Florida lakes, legally fished hoop nets set for cat-
            fish apparently had little impact on the game fish populations of these lakes
 INTRODUCTION

 Commercial fishing is an important industry in many
 of the counties bordering the St. Johns River. Based
 on  1981-82 commercial landings data  from  major
 commercial fish houses on the river, approximately
 4.14 million kg of commercially important species
 were harvested with an ex-vessel value of 2.87 million
 dollars (Hale et al. 1982). Total revenue returned to the
 local economy by this fishery amounted to nearly six
 million dollars. Hoop nets  caught approximately 80
 percent of  the 1,238,575 kg of catfish  harvested in
 1981-82.
  For a number of years, commercial fishing has been
 a source of controversy between  commercial and
 sport  fishermen along the St. Johns River (Dequine,
 1952; Hale et al. in press). Public pressure from sport
 fishermen resulted in the elimination of all types of
 commercial  fishing  devices with the exception of
 trotlines in 1946. In response to concerned sportsmen,
 commercial fishermen, and  fishery scientists, a staff
 of fishery biologists was appointed to determine the
 proper place of commercial fishing in fresh water. In
 early 1948,  game  fish by-catches were  observed by
 Florida Game and Fresh Water Fish Commission per-
 sonnel in commercially fished hoop nets, pound nets,
 wire traps and  haul seines to determine possible ef-
 fects these commercial fishing devices might have on
game  fish  populations. Data  indicated  that these
devices could be used to take catfish and other non-
game fish with  little or no impact on game fish popu-
 lations (Dequine, 1952).
  Although studies were conducted in the late 1940's
and in 1980 to observe catch composition of hoop nets
fished in riverine habitats (Dequine, 1948,1950; Hale et
al. in press), more specific  data were needed  to
answer questions concerning the present impact of
hoop nets fished in lake habitats. The main objectives
of this study were to determine the present catch com-
position of lake hoop nets and what effects, if any,
those nets had on game fish populations.
MATERIALS AND METHODS

To document game fish by-catch in lake hoop nets,
project  personnel  accompanied  commercial
fishermen during normal fishing operations in three
central Florida lakes. These three lakes (Lake George,
Crescent Lake, and Little Lake George) are a part of or
connected indirectly to the St. Johns River and repre-
sent a total surface area of approximately 25,500 hec-
tares. Fishing effort was reported in net-days with one
net-day equalling one hoop net fished for a 24-hour
period.
  Observed hoop nets usually consisted of four hoops
varying from 0.9 m to 1.8 m in diameter with a funnel at
each of the front two hoops. The front funnel and net
wall were constructed of 5.1  -7.6 cm stretch mesh
nylon netting, with the rear funnel and wall of 5.7 cm
stretch mesh.  Because  many  fishermen build their
own nets, some variation in design and construction
                                               320

-------
was  observed. Nets were anchored to the lake bed
with  1.2 m sections of 1.3 cm diameter reinforcement
rod and nylon rope and were set with funnel openings
facing downstream. Observed nets were fished empty
(unbaited) or were baited with soybean chips or blue-
back herring, Alosa aestivalis. An attempt was made
by the fishermen to fish the nets on a weekly basis.
  Harvestable-size game fish were defined as: bluegill
Lepomis macrochirus;  warmouth,  L gulosus;  red-
breast  sunfish,  L  auritis; and  redear  sunfish,  L.
microlophus ^  15.2 cm;  black crappie, Pomoxis
nigromaculatus  ^ 22.9  cm;   largemouth  bass,
Micropterus salmoides; and striped bass, Morone sax-
atilis^25.4 cm. Hybrid striped bass (White bass, M.
chrysops, X striped  bass),  an introduced  game  fish,
were considered harvestable at a total  length of ^25.4
cm.
  Commercial catch information including location of
the  set,  amount   of   time  fished, numbers  of
harvestable  and  nonharvestable  size  game  fish
caught  and  initial mortality of netted  fish was docu-
mented for each hoop net observed. Initial mortality in
noncommercial species was assumed when the fish
could not swim away under its own power before the
fishing  boat moved to another site. Because commer-
                               FISHERY MANAGEMENT

cially important species were kept and not returned,
initial mortality was assumed when the fish was either
dead  or considered unfit to keep. Delayed mortality
was not addressed in this study. All game fish caught
in observed hoop nets were immediately returned to
the water in compliance with Florida Game and Fresh
Water Fish Commission regulations.
   Estimates of game  fish harvest, initial  mortality,
and number caught per hectare during 1 year were
determined for all hoop nets fished in the three lakes.
These parameters were determined by using known
fishing pressure, area of the three lakes, and the game
fish catch rate.
RESULTS AND DISCUSSION

From June 1980 through March 1983, commission per-
sonnel accompanied commercial fishermen during 23
trips on Lake George,  Little Lake George, and Cres-
cent Lake to record hoop net catches. Hoop nets
caught a variety of fish, crustaceans and amphibians.
Twenty-eight species of fish were captured along with
blue crabs Callinectes sapidus and  one  American
alligator, Alligator mississippiensis.
  Table 1.—Catch composition of 528 commercially fished hoop nets observed in Lake George, Little Lake George, and
                           Crescent Lake, Fla., from June 1980 through March 1983.

                                                                                              Percent
                                           Number          Percent           Number            Initial
Species                                    Caught          of Total            Dead            Mortality

Commercially Important Species

Catfish (white catfish, channel catfish,
  brown bullhead)                              10,919              59.9               116                1.1
Blue crab                                       667              3.7                0                0.0
Striped mullet                                    231              1.3                0                0.0
Gizzard shad                                     174              1.0                25               14.4
Golden shiner                                     20              0.1                1                5.0
American eel                                      17              0.1                0                0.0
Southern flounder                                  6            <0.1                1               16.7
Blueback herring                                   3            <0.1                1               33.3
SUM                                         12,037              66.1               145                —
MEAN                                           —               —                —                1.2

Game Fish Species

Black crappie                                   5,366              29.4                25                0.5
Bluegill                                         655              3.6                4                0.6
Warmouth                                       19              0.1                0                0.0
Striped bass                                      14              0.1                0                0.0
Largemouth bass                                   6            < 0.1                0                0.0
Redear sunfish                                     2            < 0.1                0                0.0
Redbreast sunfish                                  1            <0.1                0                0.0
Hybrid striped bass                                 1            < 0.1                0                0.0
SUM                                          6,064              33.3                29                —
MEAN                                           —               —                —                0.5

Other Nongame Fish Species

Hogchoaker                                       68              0.4                0                0.0
Atlantic croaker                                    18              0.1                1                5.6
Atlantic stingray                                   10            < 0.1                0                0.0
Bowfin                                            5            
-------
 LAKE AND RESERVOIR MANAGEMENT
 Table 2.—Catch data from 528 commercially fished hoop nets observed from June 1980 through March 1983 in Lake George,
                                  Little Lake Georcie, and Crescent Lake, Fla.
                                                         Unbaited nets
                                                                                            Baited nets
 Number hoop nets observed
 Number net-days
 Number commercially important species
 Number game fish species
 Number game tish/net-day
 Number black crappie/net-day
 Number bluegill/net-day
 Percent game fish species
 Number commercial species/net-day
 Number catfish/net-day
 Percent catfish
 Percent commercially important species

   Commercially  important  species comprised  613.1
 percent of all fish harvested by observed hoop  nets
 (Table 1) as compared to 94.0 percent in wire traps and
 78.7 percent in river hoop nets (Hale et al., 1982). White
 catfish, Ictaluris catus, channel catfish, /. punctatus;
 and brown bullhead, /. nebulosis combined  made up
 90.7 percent of all commercially  important species
 caught. Blue crab comprised 5.5 percent of  the com-
 mercially  important species,  while striped  mullet,
 Mugil  cephalus,  and gizzard shad, Dorosoma cepe-
 dianum, comprised 1.9 percent and 1.4 percent respec-
 tively.  Catfish, the most economically important com-
 mercial species, exhibited 1.1 percent initial mortality
 while  the overall initial  mortality  of all  commercial
 species was 1.2 percent. This higher percent mortality
 could  be the result of overcrowding, gilling  of small
 catfish, and predation of gilled or injured catfish by
 blue crabs.
   Game fish species comprised 33.3 percent  of the
 total hoop net  catch as compared to 5.9 percent in
 wire traps and 19.7 percent in river hoop nets (Hale et
 al., 1982).  In 4,356 net-days, 6,064 game fish were
 caught, or 1.4 game fish/net-day (much higher than the
 0.8 game  fish/net day recorded in river hoop nets.)
 Harvestable size game fish represented 7.7 percent of
 the total net harvest and 23.1 percent of all game fish
 caught. Catch rates of harvestable size game  fish in
 lake hoop  nets (0.2/net-day) were  identical to catch
 rates in river hoop nets (0.2/net-day). Black crappie
 was the dominant game  fish species  caught, com-
 prising 88.5 percent of all game fish, followed by blue-
 gill at  10.8 percent. All  other game fish  species
 totalled <1.0 percent of all game fish caught. Bluegill
 exhibited the highest initial mortality of all game fish
 species caught (0.6 percent) while black crappie ex-
 hibited 0.5 percent initial mortality.  No initial mortality
 was observed in any other game fish species. Overall
 initial  mortality for all  game fish was 0.5 percent or
 0.007 fish/net-day. This was  much  lower than the 2.0
 percent initial mortality for game fish observed in  river
 hoop nets (Hale et al. in press). One hoop net  of the
 type observed in this study would have to be fished for
 150 net-days to cause the initial mortality of one game
 fish.
   Nongame fish of no commercial value comprised
 only 0.6  percent  of  the  total  catch  (Table  I).
 Hogchoakers,  Trinectes maculatus,  comprised  59.6
 percent of  this  category,  while  Atlantic  croaker,
 Micropogon  undulatus,  and  Atlantic stingray,
Dasyatis sabina, comprised  15.8 percent and 8.8  per-
 cent, respectively. Overall initial mortality of fish from
 this category was 5.3 percent.
   One  factor  that  influenced  species  composition
 was whether or not hoop  nets  were fished baited or
 unbaited (Hale  et al. 1981, 1982; Pierce et al.  1981).
         385
       3,391
       7,732
       5,557
         1.6
         1.5
         0.1
        41.5
         2.3
         2.0
        49.9
        57.7
  143
  965
4,305
  507
  0.5
  0.1
  0.4
 10.5
  4.5
  4.2
 83.7
 89.3
Hoop nets observed from  1980-82 were  not  baited
while almost half of the observed nets in 1982-83 were
baited with soybean chips or blueback herring. Com-
mercially important species comprised 89.5 percent of
all fish caught in baited nets and 57.7 percent in un-
baited  nets. Catfish comprised 49.9 percent  of  the
total catch in unbaited hoop nets and 83.7 percent of
the total catch in baited nets (Table 2). A higher catfish
catch rate was also observed in baited nets (4.2 cat-
fish/net-day)  than   in   unbaited   nets   (2.0
catfish/net-day).   Pierce  et al. (1981)   reported  a
significantly higher catch rate  of channel catfish in
hoop nets baited with soybean cake than in unbaited
nets in  the upper Mississippi  River. The increased
revenue return from the higher catfish catch rate in
baited nets was partially offset by the additional cost
of bait.
  The presence or absence of bait in lake hoop nets
also affected game fish catch rate and composition.
Game fish comprised 41.5 percent of all fish caught in
unbaited nets and 10.5 percent in baited nets.  Un-
baited  nets averaged 1.6  game fish/net-day while
baited  nets averaged  05  game fish/net-day. Black
crappie and bluegill were the two most frequently cap-
tured game fish  species. They combined for 99.2 per-
cent and 99.3 percent of all game fish caught in baited
and unbaited nets, respectively. Black crappie com-
prised 24.6 percent of all game fish caught in baited
nets and 94.3 percent in unbaited nets. Catch rates of
black crappie were significantly higher (P<0.05) in un-
baited  nets (1.5/net-day) than in baited nets (0.1/net-
day). Black crappie were obviously not attracted to
bait and may have been repelled by its presence.
   Bluegill comprised  74.6  percent  of all  game  fish
caught in baited nets and only 5.0 percent in unbaited
nets. However, no significant difference (P >0.05) in
the bluegill catch  rate between baited and unbaited
lake hoop  nets  was  observed. Pierce et al. (1981)
reported a  higher catch rate of black crappie in un-
baited hoop nets and a higher bluegill catch rate in
baited hoop nets.
  Another factor that could have influenced lake hoop
net catch rates and composition was the presence of
a strong year  class of fish. Commercially important
species comprised 89.6 percent of all fish caught in
1982-83,  up from 55.3 percent  in 1980-82, although
catch rates remained virtually the same during both
time periods. Game fish species comprised 9.6 per-
cent of all fish caught in 1982-83, down from 43.8 per-
cent caught in 1980-82. Most of the game fish caught
in 1980-82 were  sub-harvestable size black crappie
from a very strong 1979 year class. As this year class
suffered natural and fishing mortality over the next
three years, game fish composition declined from 43.8
percent  in  1980-82 to 9.6 percent  in  1982-83.  The
                                                 322

-------
                                                                                     FISHERY MANAGEMENT
absence of large numbers of juvenile black crappie in
1982-83 partially explained the 34.2 percent decrease
in game fish species.
   Observed baited hoop nets were fished or emptied
an average  of every 6.7 days in all  three lakes. The
catch composition of one baited hoop net  averaged
29.9 organisms  from  the  commercially  important
category (28.0 catfish) and 3.5 game fish (2.6 bluegill
and 0.9 black crappie).  Unbaited nets were fished an
average of every 8.8 days. The average catch composi-
tion of one unbaited lake hoop net was 20.1 organisms
from the commercially  important category  (17.9 cat-
fish, 1.5 blue crabs and 0.5 striped  mullet) and 14.4
game fish (13.6 black crappie and 0.7 bluegill).
   Using an estimated range of 500 to 1,000 hoop nets
for fishing pressure, surface area of the three  lakes
(25,500  hectares) game  fish catch rate (1.4/net-day),
the number of fish caught per  hectare  in 1  year was
determined. An estimated 10.0 to 20.0 game fish/hec-
tare were caught over a 1 year period, of which all were
required to be returned, and suffered only 0.5 percent
initial mortality. Using our knowledge of standing crop
data from Florida lakes, we determined that legally
fished hoop nets set  for catfish had little impact on
the game fish populations of these lakes.
 REFERENCES

 Dequine, J.F. 1948. Biennial Rep. 1947-48. Florida Game
   Fresh Water Fish Comm. Tallahassee.
       _. 1950. Biennial Rep. 1949-50. Florida Game Fresh
   Water Fish Comm. Tallahassee.
	1952. Florida's controlled seining program.  Fla.
   Game Fresh Water Fish Comm. Fish Manage. Bull. No. 1,
   Tallahassee
Hale, M.M., J.E. Crumpton, and D.J. Renfro. 1981. Commer-
   cial Fisheries Investigations Rep., Fla. Game Fresh Water
   Fish Comm., 1980-81 Annu. Rep. Eustis.
	1982. Commercial  Fisheries Investigations Rep.
   Fla. Game Fresh Water Fish Comm., 1981-82 Annu. Rep.
   Eustis.

	In press. Game fish by-catch in commercially fished
  hoop nets in the St.  John's River,  Fla.  Proc. SE Ass.
Fish Wildl. Agencies.

Pierce, R.B., D.W. Coble, and S. Corley. 1981.  Fish catches
   in baited and unbaited hoop nets in the upper Mississippi
   River. N. Am. J. Fish. Manage. 1:204-06.
                                                  323

-------
 DEVELOPMENT OF  FISH POPULATIONS AND MANAGEMENT
 STRATEGIES FOR THE BLENHEIM-GILBOA
 PUMPED STORAGE  RESERVOIRS
 DAVID L THOMAS
 Charles T. Main
 Boston, Massachusetts

 QUENTIN ROSS
 ALAN MILTON
 New York Power Authority

 JAMES M. LYNCH
 Charles T. Main
            ABSTRACT
            Fish populations in the Blenheim-Gilboa (B-G) pumped storage reservoirs in the Catskills in
            New York have been studied since the reservoirs were completed in 1973. The reservoir popula-
            tions developed entirely from the fishes present in Schoharie Creek and from emigrants from
            Schoharie Reservoir located 2.5 miles upstream. The fish populations of Schoharie Creek were
            composed primarily of pumpkinseed, rock bass, white sucker, and brown bullhead. Thirty-four
            species of fish have been collected in the' reservoirs. The more common species in the upper
            reservoir included yellow perch, pumpkinseed and redbreast sunfish. In the lower reservoir,
            white sucker, carp, brown bullhead and pumpkinseed were common. Differences between the
            populations in the two reservoirs are attributed to differences in substrate and to loss of shallow
            water caused by water level fluctuations. Management techniques employed to date include
            removal of rough fish to enhance  gamefish, construction of constant level ponds for sunfish
            spawning, and stocking of trout for trout fishing. A fourth technique which appears attractive is
            the stocking of young walleye fry to enhance the walleye fishery.
 INTRODUCTION

 Pumped storage projects increase the  reliability of
 electrical  generation  systems by rapidly providing
 energy during times of peak demand. Water is releas-
 ed from an upper reservoir and runs through turbines
 into a lower  reservoir to generate electricity whon
 peak power is needed. During periods of low power de-
 mand, water from the lower reservoir is pumped to and
 stored in the  upper reservoir until the next period of
 peak demand.
  Usually, the development of a  pumped  storage
 facility requires the creation of an upper reservoir with
 modifications to an existing adjacent body of water to
 function as the lower reservoir. Therefore, the  fish
 community that  develops in  the upper reservoir
 reflects the well-established community  in the lower
 reservoir.
  Most ecological studies of pumped storage reser-
 voirs have examined the effects of project operation
 on established fish populations (Miracle and Gardner,
 1980). The development of the fish community  and
 fishery in newly constructed reservoirs has generally
 received less attention.  Questions such  as, What
 game species are compatible with  pumped storage
 operation? and How can these reservoirs be managed
 to provide a fishery? have been largely ignored.
  Both the   upper  and  lower reservoirs  of  the
 Blenheim-Gilboa  pumped storage project were con-
 structed at the same time and are similar in size. The
 purpose of this paper is to describe the fish popula-
tions  that developed  in  these two reservoirs,  d f-
 ferences between the two reservoirs, and future man-
 agement options.
THE BLENHEIM-GILBOA PROJECT

The  Project is a 1,000 MW pumped storage facility
located on Schoharie Creek in the northern Catskill
mountains of New York (Fig. 1). Lower B-G was form-
ed by impounding Schoharie Creek in the fall of 1971.
The  Project,  completed  in 1973,  is operated by the
New York Power Authority and uses a head differ-
ential of about 1,000 feet between two newly created
reservoirs.
  The Project is operated on a weekly cycle with Up-
per  B-G  full on Monday morning  and  sufficient
storage on other weekday mornings for at least 4
hours generation of rated output. No generation oc-
curs  on weekends. The usual daily schedule is genera-
tion from 8 a.m. to 9 p.m. and pumping from 11 p.m. to
7 a.m. Typically,  this cycle results in two  units
operating during the  generation cycle and four units
during the pump cycle. On weekends, water is pumped
to Upper B-G to achieve full storage capacity. Mean
water level fluctuations occurring in the reservoirs are
shown in Table 1.
  Schoharie Creek is the major source  of water for
Lower  B-G, with two additional small  creeks also
draining into the reservoir. Two miles upstream from
Lower B-G on Schoharie Creek is Gilboa Dam which
                                              324

-------
                                                                                 FISHERY MANAGEMENT
forms Schoharie Reservoir. Gilboa Dam usually spills
water from  late fall through  late  spring. Flow  in
Schoharie Creek between Lower B-G and Schoharie
Reservoir  is extremely variable ranging from 0  to
42,900 cfs during the years of the study.  In the sum-
mer, the creek  often becomes a series  of isolated
pools.
  The project  operation  supplies Upper B-G with
water. Other than precipitation and surface runoff, the
reservoir has no natural inflows.
  Upper and Lower B-G are similar in size and capaci-
ty (Table 1). However, Upper B-G is generally bowl-
shaped with steep banks, while Lower B-G is elongate
with a more gently sloping shoreline. Approximately
two thirds of the bottom of the littoral zone of Upper
B-G is a manmade embankment covered with riprap,
while the  bottom  in the Lower  B-G  is composed
primarily of silt and clay.
  Water  quality  sampling  has  been  conducted
throughout  the  10  years  of Project  operation.
Suspended sediment concentrations in  Lower B-G
are less than 10 mg/l more than 80 percent of the year
and  rarely exceed 30 mg/l. During the summer, con-
centrations are usually less than 5 mg/l. Ice  covers
part of the reservoir in winter and a maximum daily
average summer temperature of about  230-24°C  is
reached in July and August. Mixing caused by  project
operation ensures some open water in winter.  During
summer, lower water temperatures  are occasionally
found near  the  bottom  at  the  dam,  but thermal
stratification is minimal.  In Upper  B-G, suspended
sediment levels and temperatures are often slightly
lower than in the lower reservoir.
MATERIALS AND METHODS

Fish have been  sampled every year in Upper and
Lower B-G from  1973, the first year of Project opera-
tion, through 1982. Collection methods and periods
have varied; however, the two primary means of collec-
tion have been by gill nets and electrofishing. From
1973 through 1977, seines, trap nets, and block nets
were also used.
  Gill nets were 250 to 300 feet in length, 6 to 8 feet in
depth, and had various sized meshes of 1 to 4 inches.
They were set perpendicular to shore and were fished
overnight.
  Table 1.—Physical characteristics of the Blenheim-Gilboa
          Pumped Storage Project Reservoirs.
Characteristic
Watershed (square miles)
Gilboa Dam spilling
Gilboa Dam not spilling
Surface area (acres)
Full pool
Minimum pool
Capacity (acre feet)
Full pool
Minimum pool
Shoreline length (miles)
Depth (feet)
Maximum
full pool
Mean
Full pool
Minimum pool
Water level fluctuation (feet)
Daily
Weekly
Lower Reservoir

354
40

420
220

16,300
3,600
9.4


80

39
16

14.2
27.7
Upper Reservoir

<1
<1

390
260

18,400
5,700
3.5


95

47
22

10.2
22.7
  Electrofishing was conducted at night  along  the
shorelines with a pulsed direct current electrofisher.
  Trap nets were composed of a rectangular box with
a funnel-shaped entrance. Each net had two wings (15
ft x  20 in) and a main leader (25 ft  x 20 in) bisecting
the angle of  the  wings which directed fish to  the
funnel-shaped opening  of the  rectangular  box. Nets
were set  so that the box was generally in water less
than 8 feet deep.
RESULTS AND DISCUSSIONS

The sources of the fish populations that ultimately
developed in the B-G Reservoirs were the 5-mile sec-
tion of Schoharie  Creek  between Gilboa Dam and
Lower B-G, two small tributaries and spillover from
Schoharie Reservoir.  Schoharie Creek  above and
below the impoundment was sampled in 1973 prior to
the initiation of project operation. Fourteen species of
fish were collected between Gilboa Dam and Lower
B-G with pumpkinseed and rock bass the most abun-
dant. Game fishes  included walleye and smallmouth
bass. Fourteen species were also found downstream
of Lower B-G. Pumpkinseed, fallfish, and rock bass
were most abundant.
  During the 10 years of sampling conducted on the
B-G Reservoir, 35 species were collected (Table 2). In-
tensive collections of fish by trap net, seine, and boat
electrofishing from April through September 1977 in-
dicated the dominance of pan fishes (98 percent by
number, 48 percent by weight) in Upper B-G  and of
rough fishes (47 percent  by  number, 81 percent  by
weight)  in  Lower  B-G.  Numerically,  the common
fishes in Upper  B-G were yellow perch (63 percent of
                                                   Figure 1.—Location of the Blenheim-Gilboa pumped storage
                                                   project.
                                                325

-------
LAKE AND RESERVOIR MANAGEMENT
 the total catch), pumpkinseed (22 percent) redbreeist
 sunfish (7 percent), logperch (2 percent), and walleye (2
 percent). In Lower B-G, white sucker (31 percent), carp
 (12 percent),  brown bullhead (11  percent), pumpkin-
 seed  (10 percent), yellow  perch (9 percent),  aid
 smallmouth bass (7 percent) predominated.
   Early  changes in certain fish  populations are il-
 lustrated by trap net  collections  taken in the ressr-
 voirs from 1973 through 1977 (Table 3). Brown bullhead
 was consistently more abundant  in Lower B-G than
 Upper B-G but its population appeared to decline in
 the lower reservoir from 1974 through 1977. Pumpkin-
 seed was more abundant in Lower B-G in  1973 but ts
 numbers decreased so that it was more abundant in
 Upper  B-G from  1974 through  1977,  through  ts
 numbers fluctuated considerably.
   Redbreast sunfish and yellow perch were rarely col-
 lected by trap net in  Lower B-G  and no consistent
 trend  was  observed from 1973 through  1977.  Reid-
 breast sunfish was more abundant in Upper B-G but,
 again, its population did not exhibit any trend. Yellow

  Table 2.—Fishes collected in the Lower and Upper
     Reservoirs  Blenhelm-Gilboa Project from 1973
                  through 1982.
Scientific Name
Common Name
SALMONIDAE
  Coregonus artedii
  Salmo gairdneri
  Salmo trutta

ESOCIADE
  Esox niger

CYPRINIDAE
  Campostoma anomalum
  Cyprinus carpio
  Exoglossum maxillingua
  Notemigonus crysoleucas
  Notropis analostanus
  Notropis atherinoides
  Notropis cornutus
  Notropis hudsonius
  Notropis rubellus
  Pimephales notatus
  Rhinichthys cataractae
  Rhinichthys atratulus
  Semotilus atromaculatus
  Semotilus corporalis

CATOSTOMIDAE
  Catostomus commersoni
  Hypentelium nigricans
  Moxostoma macrolepidotum

ICTALURIDAE
  Ictalurus nebulosus
  Noturus flavus

CENTRARCHIDAE
  Ambloplites rupestris
  Lepomis auritus
  Lepomis cyanellus
  Lepomis gibbosus
  Lepomis macrochirus
  Micropterus dolomeiui
  Micropterus salmoides

PERCIDAE
  Etheostoma flabellare
  Etheostoma olmstedi
  Perca flavescens
  Percina caprodes
  Stizostedion vitreum
TROUTS
  Cisco
  Rainbow trout
  Brown trout

PIKES
  Chain pickerel

CARPS & MINNOWS
  Central stoneroller
  Common carp
  Cutlips minnow
  Golden shiner
  Satinfin shiner
  Emerald shiner
  Common shiner
  Spottail shiner
  Rosyface shiner
  Bluntnose minnow
  Longnose dace
  Blacknose dace
  Creek chub
  Fallfish

SUCKERS
  White sucker
  Northern hogsucker
  Shorthead redhorso

CATFISH ES
  Brown bullhead
  Stonecat

SUNFISHES
  Rock bass
  Redbreast sunfish
  Green sunfish
  Pumpkinseed
  Bluegill
  Smallmouth bass
  Largemouth bass

PERCHES
  Fantail darter
  Tesselated darter
  Yellow perch
  Logperch
  Walleye
 perch  was  more  abundant  in  Upper  B-G  and  its
 population appeared to increase after 1974.
   Gill  net and electrofishing catches (Table 4) cannot
 be used to make comparisons between years because
 of the differences in gear, effort, and time of deploy-
 ment.  However, they do provide valid comparisons be-
 tween fish populations in  Upper and Lower B-G at
 several different times during the operation of the
 pumped storage project. White sucker was consistent-
 ly more abundant in Lower B-G. Redbreast sunfish,
 pumpkinseed, yellow perch, and walleye  were con-
 sistently more abundant in Upper B-G. Rock bass and
 smallmouth bass were  more abundant in Lower B-G
 during the first half of the  study but they were more
 abundant in Upper  B-G during the second half of the
 study. Except for 1974  through  1975 carp was more
 abundant in Lower  B-G.
   The major factors affecting the development of the
 fish  communities in the Blenheim-Gilboa reservoirs
 appear to be the morphology and substrates of the
 two  reservoirs and their interaction with water  level
 fluctuations. Lower B-G has a gently sloped basin,
 and  the shoreline area subject to water level  fluctua-
 tions is greater than that in Upper B-G.  Two thirds of
 the shoreline in Upper B-G is a steeply sloped dike,
 thus less area is affected by fluctuating water levels.
 The substrate of the unexposed bottom  of both reser-
 voirs also differ. Rocky substrates predominate in Up-
 per B-G; clay and silt in Lower B-G.
   The low abundance of carp, brown bullhead, and
 white  sucker  in Upper B-G probably  reflects  the
 demersal habitats  of  these  species  and   their
 preference for shallow  water habitat. The decline of
 carp and brown  bullhead in Lower B-G  was probably
 caused by  water  level fluctuations  which  expose
 much  of the shallow water spawning habitat. White
 sucker which  spawn  in Schoharie Creek were  con-
 sistently more abundant in Lower B-G and did not ex-
 hibit any decline in abundance.
   The  greater and continued abundance of  yellow
 perch, redbreast sunfish, and walleye in Upper  B-G
 probably results from  the  fact that yellow perch is
 pelagic and walleye and redbreast sunfish prefer the
 rocky  substrate which  predominate  in  Upper B-G.
 Smallmouth bass and rock  bass  are fairly common in
 both reservoirs. Both  species spawn  in Schoharie
 Creek  and,  particularly in  the case of smallmouth,
 spawning may also occur in the  B-G reservoirs.

 Table 3.—Annual catch per unit effort (number per hour) of
   dominant species in trap net collections taken in the
   Upper and Lower Blenheim-Gilboa Reservoirs from
	 1973 through 1977.

                                Year
 Fishes	1973  1974   1975    1976   1977
 Carp
  Upper Reservoir
  Lower Reservoir
 Brown Bullhead
  Upper Reservoir
  Lower Reservoir
 Redbreast Sunfish
  Upper Reservoir
  Lower Reservoir
 Pumpkinseed
  Upper Reservoir
  Lower Reservoir
Yellow Perch
  Upper Reservoir
  Lower Reservoir
0.02  <0.01
0.48    0.01
0.01
0.01
  0    0.05     0  < 0.01     0
0.49    0.50   0.21    0.04   0.02


0.06    0.09   0.03    0.05   0.01
  0  < 0.01 < 0.01      0     0


0.09    0.50   0.28    0.59   0.14
2.53    0.36   0.13    0.05   0.04


0.14    0.07   0.18    0.45   0.26
0.08  <0.01   0.02    0.06   0.05
                                                 326

-------
                                                                                   FISHERY MANAGEMENT
  Pumpkinseed also spawn in Schoharie Creek but
were more abundant in Upper B-G and appeared to
decline in Lower B-G. This may be the result of habitat
preference. The riprap  in Upper B-G apparently pro-
vides more cover than found in Lower B-G.
MANAGEMENT OF THE RESERVOIR
FISHERIES

Early management of the B-G reservoirs entailed the
removal  of  rough fish. Over  4,000 carp and white
sucker each were removed from 1973 through 1979,
primarily by use of  block nets in Lower B-G. This
removal may have contributed to the apparent decline
in carp (Table 4). However, the disappearance of carp
in the trap nets (which caught primarily small fish)
suggests that  the absence  of suitable spawning
habitat due to water level fluctuations was the primary
cause of the decline.
   Three  experimental 1-acre  constant  level ponds
were constructed along the natural shorelines of Up-
per B-G  to evaluate their potential to supply centrar-
chids to  the reservoir. They have been stocked since
1979 with  largemouth bass  and pumpkinseed and
both species have spawned  in these ponds. These
observations suggest that on a larger scale, constant
level ponds could probably  make a  significant con-
tribution  to reservoir fish populations and could be in-
tensively managed to increase the production of both
game and forage species.
  The B-G reservoirs have also been managed for a
put-and-take trout fishery. Rainbow  and brown trout
were stocked in the reservoirs from 1977 through 1982.
Rainbow  trout, in particular, have provided a put-and-
take fishery and many individuals have held over to
the next year.
  Walleye  is  the dominant game  fish  in  nearby
Schoharie Reservoir  and yellow perch  is its primary
prey. Both species are common  in the B-G system,
particularly in the upper reservoir. Although there ap-
pears to be some walleye spawning within the B-G
reservoirs based on the presence of larvae, most
recruitment  is  probably from Schoharie  Reservoir.
Schoharie Creek above  Lower B-G does not provide
suitable spawning habitat for walleye. The abundance
of prey species (particularly yellow  perch in  Upper
B-G) and the relative abundance of small zooplankton
in the  B-G  reservoirs  suggests that walleye abun-
dance  could be increased  without unbalancing the
fish community.
  Mills and Schiavone (1982) maintain that the relative
abundance of small zooplankton species is a useful
index  of  the  predator-prey  balance  in  fish com-
munities  in small  warmwater   lakes  containing
predominately  centrarchids  and  percids.  These
authors observed that the relative abundance of small
zooplankton  species  increased  as   the  ratio  of
predators to prey decreased.
  In the B-G reservoirs, small zooplankton comprised
37 percent of the zooplankton community in Lower
B-G and 42 percent in Upper B-G (Gulp et al. 1978). In
nearby Schoharie Reservoir, where the fish communi-
ty is  also  dominated  by  a walleye-yellow  perch
association  and good spawning habitat is abundant,
small zooplankton comprise only 10 percent of the
zooplankton community.
  Based on the presence of catch-sized walleye in the
B-G reservoirs  and the means to  increase  their
population through stocking, we  conclude that there
 Table 4.—Comparison between fish abundance in the Upper and Lower Blenheim-Gilboa Reservoirs based on catch per
         unit effort of dominant species in gill netting and electrofishing collections from 1974 through 1982.

                                                             Period»
Fishes
White Sucker
Upper Reservoir
Lower Reservoir
Redbreast Sunfish
Upper Reservoir
Lower Reservoir
Pumpkinseed
Upper Reservoir
Lower Reservoir
Yellow Perch
Upper Reservoir
Lower Reservoir
Walleye
Upper Reservoir
Lower Reservoir
Carp
Upper Reservoir
Lower Reservoir
Rock Bass
Upper Reservoir
Lower Reservoir
Smallmouth Bass
Upper Reservoir
Lower Reservoir
1974-1975
N/H
0.02
0.04
0.06
0
0.18
0.05
0.19
0.02
0.02
<0.01
0.08
0.03
0.01
<0.01
0.01
<0.01
1976-1977
N/MIN
0.03
0.99
0.35
0.01
0.89
0.34
4.48
0.17
0.14
0.02
0.05
0.21
0.06
0.13
0.05
0.16
1979-1980
N/MIN
0.01
0.54
0.03
0
0.13
0.05
0.23
0.05
0.09
0.02
0.01
0.04
0.08
0.04
0.09
0.07
1981-1982
N/MIN
<0.01
0.21
0.04
0
0.29
0.02
1.19
0.29
0.12
0.01
<0.01
0.02
0.59
0.07
0.13
0.11
a1974-1975 data based on gill net collections from April-September 1974 and 1975,
 1976-1977 data based on electrofishing collections from April-October 1976 and April-September 1977;
 1979-1980 data base on electrofishing collections from May, June, October and November 1979 and May and August 1980.
 1980-1981 data based on electrofishing collections from October 1980 and 1981.
                                                 327

-------
LAKE AND RESERVOIR MANAGEMENT
is good potential for  the development of a  percid
fishery. The condition of walleye in the B-G system is
good—specimens 3 years and older are about  the
same size and weight as comparably aged specimens
from Schoharie  Reservoir.  The  absence of  good
spawning habitat in the B-G system can be compen-
sated for by periodic stocking of walleye fry. The con-
struction of larger constant  level ponds  would  en-
hance centrarchid populations and the stocking of
pelagic forage species could be beneficial if predator
populations greatly increased. Given the array of pos-
sible management options, it is likely that the recrea-
tional fishery in these reservoirs can be improved.
Miracle, R.D., and J.A. Gardner. 1980. Review of the literature
  on the  effects of pumped storage operations on ichthyc-
  fauna. Pages 40-53 in Proc. Clemson Workshop on Environ.
  Impacts of Pumped Storage Hydroelectric Operations. U.S.
  Fish Wildl. Serv.

Mills, EL, and A. Schiavone, Jr. 1982. Evaluation of fish com-
  munities through  assessment of zooplankton populations
  and measures of lake productivity. N. Am. J. Fish. Manage.
  2:14-27.
REFERENCES

Gulp, T.R., and Associates. 1978. Studies of the aquatic ecology
  of the Blenheim-Gilboa Pumped Storage Reservoirs and of
  the Prattsville Pumped Storage Site: Schoharie Reservoir and
  tributaries, Schoharie Creek, Ashokan Reservoir, and Esopus
  Creek and tributaries. Prog. rep. for the period Jan. 1-Dec. 31,
  1977. Ichthyological Associates, Inc.
                                                   328

-------
                                            Urban  Lake   Quality
FATE OF HEAVY METALS IN STORMWATER MANAGEMENT  SYSTEMS
HARVEY  H. HARPER
YOUSEF  A. YOUSEF
MARTIN  P. WANIELISTA
Department of Civil Engineering and  Environmental  Sciences
University of Central  Florida
Orlando,  Florida
            ABSTRACT

            The State of Florida requires that stormwater originating within a new project or development be manag-
            ed and treated within the boundaries of the development to protect surface waters. Retention and
            exfiltration systems are the most common management practices. Theoretically these provide com-
            plete removal efficiency up to the design capacity since none of the stormwater reaches the receiving
            water body by direct inflow. The fate of various pollutants once entering these systems is not known,
            particularly whether heavy metals remain in them locked tightly by a chemical or physical association
            or slowly disperse outward over a much larger area to other water bodies. Investigations on an 8-year-old
            stormwater retention facility (1.5 ha) in Orlando, Fla., to define movement of heavy metals into and
            out of the basin seek to answer these questions. Stormwater runoff has been collected and analyzed
            from the input pipe for approximately 1 year. In addition, both wet and dry bulk precipitation are being
            collected to estimate the relative significance of each input source. Forty-five separate 3 cm core samples
            were collected from within the pond, divided into four sections and analyzed for zinc, cadmium, cop-
            per, aluminum, iron, lead, nickel, chromium, and phosphorus as well as moisture and organic con-
            tents. Movements of heavy metals from the inlet were estimated using the top 1 cm values. While
            zinc and lead were removed rapidly from solution near the outfall, other metals such as copper and
            aluminum were mobile Deposition of metals correlated highly with the chemical speciation of the
            metal at the time of input. The role of plants in trapping and removing heavy metals is also under
            investigation. Although a large portion of the metals seem to remain within  the basin, a certain frac-
            tion may leave the pond through percolation and groundwater movement. Five multilevel  wells in-
            stalled from the edge of the pond outward monitor downward or horizontal movement. Groundwater
            monitoring will continue throughout the typically wet summer season.
INTRODUCTION

Within the past decade  a substantial amount of
research has accumulated on water pollution caused
by the operation of motor vehicles, mainly on the
potential  aquatic toxicity  of  heavy metals  such as
lead, zinc, and chromium.  Heavy metals  have  been
proposed by several researchers as the major toxicant
present in highway runoff  samples (Shaheen,  1975;
Winters  and Gidley, 1980). Many heavy metals are
known to be toxic in  high concentrations to a  wide
variety of aquatic plants and animals (Wilber  and
Hunter, 1977).
  Two of the most popular techniques for manage-
ment  of pollution from highway runoff are roadside
swales and detention/retention facilities. Many States
now require that specified amounts of excess rainfall
from developed areas be collected and treated in such
systems. However, with   continual  inputs of toxic
elements, especially heavy metals, the resulting ac-
                                                329

-------
LAKE AND RESERVOIR MANAGEMENT
cumulations may begin to present a toxicity or pollu-
tion potential to surrounding surface waters or ground
water,  particularly if metal species begin to migrate
out of or away from the stormwater management sys-
tems.
  No definitive studies have been conducted to deter-
mine  the  fate of  toxic  species,  especially  heavy
metals, in these stormwater management system!;.
This   paper  presents  the   preliminary   results  of
research conducted on a stormwater retention facility
receiving highway runoff  near Orlando, Fla. Concen-
trations of heavy metals in stormwater runoff entering
this facility were compared with average concentra-
tions within the retention  pond and also with ground-
water concentrations beneath the pond to aid in deter-
mining the fate of heavy metals  in these systems.
SITE DESCRIPTION

The site selected for this investigation is the Maitland
Interchange on Interstate 4. This interchange, located
north of the city of Orlando, was constructed in 1976
(Fig. 1). Three borrow pits dug to provide fill for con-
structing the overpass remain to serve as stormwater
detention/retention facilities. The  ponds are inter-
connected  by large culverts  so that when the north-
eastern pond (Pond A) exceeds the design level it can
discharge to the  northwestern  pond (Pond B).  The
northwestern pond has the capability to discharge to
the southwestern pond (referred to  hereafter as the
West Pond) when design  elevations are exceeded.
However, since the volume of both Pond A and Pond B
are quite large relative to their receiving watersheds, it
is anticipated that a discharge from Ponds A or B to
the West Pond would occur only as a result of an ex-
treme rainfall event. In the 2 years in which these in-
vestigations have been conducted no surface ex-
change of  waters between Ponds A and  B and the
Figure 1.—Study site at Maitland Interchange
West Pond has been observed. Therefore, under nor-
mal conditions, the only input into the West Pond is by
way of a 45 cm concrete culvert that drains much of
the Maitland Boulevard overpass. Discharge from the
West Pond  travels  to  Lake Lucien through a large
culvert. A flashboard riser system regulates the water
level in the West Pond, and a discharge rarely occurs
to Lake Lucien. Because of the well defined nature of
both the inputs and outputs to the West Pond, this
system was chosen for investigation.
  The West Pond has an approximate surface area of
1.3 ha and an average depth of 1.5 m. The pond main-
tains a large standing crop of filamentous algae, par-
ticularly  Chara, virtually year round.  Because of  the
shallow water depth and large amount of algal produc-
tion, the pond waters  remain  in a well oxygenated
state. The sediment material is predominately sand
which is covered by a 1 cm layer of organic  matter.
  Maitland  Boulevard  crosses over  Interstate 4 by
means of a  bridge overpass created during construc-
tion of the interchange. The Maitland Boulevard bridge
consists of  two sections, one carrying two lanes of
eastbound traffic and an  exit lane, the other carrying
two lanes of westbound traffic and another exit lane.
Traffic volume on Maitland Boulevard is approximate-
ly 12,000 average daily traffic (ADT) eastbound and
11,000 westbound. Traffic volume on I-4 through  the
Maitland Interchange is  approximately 42,000 ADT
eastbound and westbound.


FIELD INVESTIGATIONS

Field investigations  conducted during 1982 and 1983
at the West Pond  were  divided into the following
tasks: (1) determination  of the  quantity of heavy
metals entering the West  Pond by way of stormwater
runoff; (2) determination of the average heavy metal
concentrations  in   the  retention  basin  water;  (3)
assessment  of the accumulation of heavy metals in
the sediments of the pond; and (4) monitoring of heavy
metal concentrations in ground waters beneath the
retention basin. To  determine the quantity of heavy
metals entering the West  Pond by way of stormwater
runoff an Isco automatic sampler was installed on the
45 cm stormsewer line.  Flow-weighted  composite
samples  were collected over a 1 year period for 16
separate storm events  representing a wide  range of
rainfall  intensities  and   antecedent dry  periods.
Samples were analyzed for Cd, Cu, Zn, Pb, Ni, Cr, and
Fe using argon plasma emission spectroscopy, and
an average  concentration was  calculated for each
metal for the year.
  To determine the  average concentrations of heavy
metals in the West Pond water, samples were col-
lected on a biweekly basis for 1 year. Each of the five
samples was analyzed separately for the heavy metals
listed, and an average value was calculated  for each
metal on each sampling date.
  To determine the  accumulation  and vertical distri-
bution of heavy metals in the sediments, a series of
2.5 cm diameter core samples were collected to a
depth of 6.8 cm. Forty-three separate core  samples
were collected in the 1.3 ha West Pond, and metal con-
centrations  in sediment layers 0-0.8 cm, 0.8-2.8 cm,
2.8-4.8 cm,  and 4.8-6.8 cm were measured  for each
core sample. Melal concentrations in the 0-0.8  cm
layer were used to investigate horizontal movement of
heavy metals from  the point of discharge  into  the
pond. Average metal concentrations in  each of  the
core sections were  used  to determine the extent of
vertical migration.
                                                330

-------
                                                                                   URBAN LAKE QUALITY
  To investigate the possibility of groundwater con-
tamination by leaching of heavy metals from storm-
water management systems, five multiport monitoring
wells were installed at locations indicated in Figure 1.
Two of the monitoring  wells  were  installed at the
edges of the West Pond with the  remaining three in-
stalled at various  locations surrounding the storm-
water management  system. The wells were designed
so that all of the sample ports were housed in a single
casing  to minimize  soil disturbance  and  reduce
recovery time  for obtaining  representative ground
water samples compared to other  monitoring well
designs such as cluster wells.
  A schematic of the monitoring well design is shown
in Figure  2. All wells were installed to a depth of 6
meters with sample ports at 0.1 m, 0.5 m, 1.0 m, 3.0 m,
and 6.0 m below the average water table depth in the
area of the well. Ground water samples were collected
from each sample port on a monthly basis using a
peristaltic pump. Approximately 10 I of  ground water
were pumped and discarded from each  port before a
sample was  collected. Samples  were  analyzed for
heavy metals as described previously.
EXPERIMENTAL RESULTS AND
DISCUSSION

Removal of Paniculate Metal Species

A comparison of average heavy metal concentrations
in stormwater runoff and in the West Pond is given in
Table 1. Concentrations of heavy metals measured in
the incoming stormwater appear to exist predominate-
ly in associations with particulate matter. Paniculate
fractions accounted for 42 percent  of the total cad-
mium, 86 percent of total zinc, 47 percent of total cop-
per, 94 percent of total lead, 89 percent of total nickel,
67  percent of total chromium, and  96  percnt of the
total  iron.  To  determine the  horizontal  mobility of
these particulate fractions  the metal concentrations
measured in the top layer of the core samples were
plotted as a function of distance  from the stormwater
inlet into the pond. The results  of these determina-
tions are shown in Figure 3. It can be seen that both
lead and zinc reached a peak concentration in the up-
per sediments at a point very near the stormwater in-
let. For these metals  peak deposition occurred at a
distance of approximately 15 meters from the point of
  discharge. Concentrations of both lead and zinc ap-
  peared to peak and  decline quickly with increasing
  distance. This pattern suggests that a large portion of
  the particulate forms of these metals is associated
  with relatively heavy particles which  tend  to  settle
  rapidly.
    As seen  in  Figure  3,  chromium  concentrations
  reached a peak at a distance of 30 meters from the
  discharge point. The decline in chromium concentra-
  tions with increasing distance was not so pronounced
  as  for lead and  zinc.  This behavior suggests that
  chromium in stormwater runoff is  associated with a
  wider range of particulate sizes and  densities than
  lead or zinc. The behavior of copper and nickle species
  in stormwater runoff appears to be quite similar. Both
  of these metals did not reach a definite peak, but tend-
  ed to settle out fairly uniformly over a relatively large
  distance.  This  suggests  that  the  particle   sizes
  characteristic of particulate copper and zinc in storm-
  water runoff are much smaller and more mobile than
  those for lead, zinc, or chromium. It appears,  there-
  fore, that lead and zinc in highway stormwater  runoff
                                                                  -End Cap

                                                                  r- Coupling
        WATER
    5cm PVC Pipe -
                                         5cm PVC Pipe
                 End Cop

       FIELD INSTALLATION
CROSS SECTION OF SAMPLING
PORT
  Figure 2.—Schematic diagram of the multiport groundwater
  sampling device.
Table 1.—Comparison of average heavy metal concentrations in stormwater runoff and in the retention basin (West Pond) at
                                         Maitland Interchange.
                 Average Values in Incoming
                       Stormwater
Average Values in Retention
      Basin Water
     Percent Change
   Through Retention
         Basin
Parameter
pH
Cadmium
Zinc
Copper
Lead
Nickel
Chromium
Iron
Number of
samples
Dissolved
5.90
1.1*
50
32
43
3.2
3.3
48
16
Total
5.90
1.9
347
60
723
28
10
1176
16
Percent
Dissolved
58
14
53
6
11
33
4
—
Dissolved
5.97
0.8
5.8
14
16
1.8
2.3
20
34
Total
5.97
1.0
6.4
16
22
2.3
3.4
61
34
Percent
Dissolved
80
91
88
73
78
68
33
—
Dissolved
+ 15
-27
-88
-56
-63
-44
-30
-58
—
Total
+ 15
-47
-98
-73
-97
-92
-66
-95
—
 •All metal concentrations listed as ng'l
                                                 331

-------
 LAKE AND RESERVOIR MANAGEMENT
 are  associated with  larger,  more dense particles
 which will settle quickly from  the point of discharge.
 Chromium appears to be associated with  both  a
 smaller average particle size and a wider range of par-
 ticles than lead or zinc. The very mobile characteris-
 tics of copper and nickel indicate a relatively small
 particle size.
   The results  shown  in Figure 3 can be useful in
 designing stormwater retention facilities aimed at the
 removal of heavy metals. By examining the metal con-
 centrations in  the sediments  of existing stormwater
 management  facilities, the distance  from the inlet
 source at which metal concentrations began to ap-
 proach background levels can be determined. If  this
 distance is assumed to be the maximum distance over
 which heavy metals will settle out then  the outfall
 structure can be placed at a distance from the inputs
 that equals or exceeds the zone of settling in order to
 optimize metal particle removal.
   It should be noted that the triangular design of the
 West Pond (see Fig. 1)  produced a situation that caus-
 ed storm water coming in through the 45 cm inlet to
 spread out quickly  in a large,  wide area. This design
 caused the inlet velocity to decrease rapidly and aided
 in settling of particulate species. A narrow design with
 a relatively rapid flowthrough velocity would decrease
 metal removal  efficiencies.

 Removal of  Dissolved Species of
 Heavy Metals

 As seen in the average West Pond water quality listed
 in Table 1, total metal concentrations are reduced sub-
 stantially when compared  with the incoming storm
 water. With the exception of cadmium, removal of par-
 ticulate forms of heavy metals exceeded 70 percent in
                                                  the retention facility. As a result, the ratio of dissolved
                                                  forms of heavy metals to total metal concentrations
                                                  has increased  considerably  in  the  retention pond
                                                  water. Whereas heavy metals associated with the in-
                                                  coming storm water were largely particulate in nature,
                                                  the average heavy metal species in the retention pond
                                                  water are largely dissolved  in nature. It is interesting
                                                  to  note,  however,  that concentrations of dissolved
                                                  species of heavy rnetals are also much lower in the
                                                  retention  pond water than in the incoming storm-
                                                  waters. It appears, therefore, that not only are partic-
                                                  ulate forms of various heavy metals readily removed
                                                  from the water column upon entering the retention
                                                  pond, but dissolved species are removed as well. Re-
                                                  moval of stormwater-generated dissolved species of
                                                  heavy metals in the West Pond averaged between 27
                                                  percent for cadmium and 88 percent for zinc.
                                                    It has been reported by numerous researchers that
                                                  the greater part of dissolved heavy metals entering or
                                                  being transported  by  natural  water systems can,
                                                  under normal physiochemical conditions,  be rapidly
                                                  removed  from the  water phase and concentrated in
                                                  the sediment phase (Wilber and Hunter, 1977; Guthrie
                                                  and Cherry, 1979; Hem and Durum, 1973). Although the
                                                  exact mechanisms responsible for removing dissolv-
                                                  ed  species in the stormwater runoff were not deter-
                                                  mined, it is believed that several of the following pro-
                                                  cesses  may  have  been involved at  this  site:  (1)
                                                  precipitation; (2) cation exchange and adsorption; (3)
                                                  co-precipitation of  hydrous Fe/mn oxides;  and  (4)
                                                  association with organic molecules.

                                                  Fate of  Heavy Metals in the Sediments

                                                  It appears from the results  presented previously that
                                                  the fate of a large portion of both the suspended and
5
Q
LJ
05

CC
a

a»
•v
en
   o
   o
   0)
   •s.
        350--
        300--
        250--
        200--
        IOO-
         50-
                               30
                                      45        60        75

                                  Distance  From Inlet (meters)
                                                                        90
I05
120
Figure 3.—Concentrations of selected heavy metals in the top 1 cm of sediments in the Maitland detention pond as a func-
tion of distance from the stormwater inlet.
                                                 332

-------
                                                                                   URBAN LAKE QUALITY
dissolved fractions of storm water associated heavy
metals is  ultimate deposition by a wide variety of
mechanisms into the bottom sediments of the receiv-
ing water body. After several years of this continual
deposition, a large accumulation of heavy metals may
develop in the sediments. This concentrated  layer of
heavy metals may present a potential pollution hazard
if leaching were to occur.
   To investigate the  potential movement of sediment
deposited heavy metals the vertical  distribution  of
heavy metals in the 43 sediment cores were examined,
and the average metal concentrations  in sediment
layers 0-0.8, 0.8-2.8, 2.8-4.8, and 4.8-6.8 cm were
calculated. Since  the heavy  metal  content in the
4.8-6.8 cm layer was very similar to heavy metal con-
centrations measured  in nearby soils unaffected  by
stormwater runoff, these concentrations were con-
sidered equal to background values and subtracted
from each of the others. The vertical distributions of
Zn, Pb, Cr, Ni, Cu, and Fe in sediment cores collected
in the West Pond are shown in Figure 4. It can be seen
from Figure 4 that the metal concentrations decreas-
ed in an exponential fashion with increasing sediment
depth and can be modeled as follows:
       C =  C0e-
                 kz
 where: C = metal concentration at a desired sediment
            depth (fXj/g dry wt)
       z = the sediment depth (cm)
       C0 = metal concentration of the sediment at
             the surface (/^g/g dry wt).
   and k =  metal  reduction constant (1/cm)

   This model was found to fit the heavy metals tested
 with  a  correlation coefficient of 0.99 or better.  The
 metal reduction constants (k values) were 1.36 for Zn,
 1.14 for Cr, 1.03 for Pb, 1.03 for Ni, 0.98 for Fe, and  0.92
 for Cu (Yousef et  al. 1983). It appears, therefore,  that
 accumulated heavy metals are attenuated very quickly
 during  movement through sediment  material. Atten-
                                tuation of metals was found to be essentially com-
                                plete at a depth of approximately 5.0 cm with normal
                                background  concentrations  below that  depth.
                                Although all  heavy metals  tested  are  attenuated
                                quickly,  it appears from the calculated metal reduc-
                                tion constants  that the vertical mobility  of heavy
                                metals can be arranged in the following order:

                                 least mobile: Zn < Cr < Pb = Ni < Fe < Cu:
                                 most mobile.

                                It can  be concluded,  therefore, that heavy metals
                                deposited within this pond, upon reaching the sedi-
                                ments, were transformed into very stable associations
                                that remained near the sediment surface and declined
                                rapidly in concentrations with increasing depth.

                                Potential for Groundwater  Contamination
                                by Heavy Metals
                                A comparison of average total heavy metal concentra-
                                tions in the retention basin water, in the top 0.8 cm
                                sediment layer, and in groundwater samples collected
                                beneath  the  retention basin  is given  in  Table 2.
                                Because of the 7-year  accumulation of heavy metals
                                from stormwater runoff in the sediments, the concen-
                                trations of  sediment-associated  metals have
                                magnified  considerably  when compared  to  the
                                average retention  basin concentrations. Magnifica-
                                tion  factors for sediment-associated metals range
                                from 287 g/ml for cadmium to 6,838 g/ml for chromium
                                when compared to average retention basin concentra-
                                tions. One of the potential problems associated with
                                this accumulation is the possibility for leaching  and
                                downward movement  of  heavy metals into ground
                                waters. If movement of heavy metals is found to occur,
                                then these stormwater management facilities could
                                have tremendous pollution potential since they are in
                                widespread use.
                                  In spite of the large accumulation of heavy metals
                                in the sediments of the retention pond, there is no
                                evidence to indicate that leeching of metals is occurr-
  E
  u
  a.
  CD
  D
  o
  o
  C
  
-------
  LAKE AND RESERVOIR MANAGEMENT
       Table 2.—Comparison of average total heavy metal concentrations in the retention basin water, the top 1 cm of
          sediments, and in groundwaler samples collectod beneath the retention basin at Maltland Interchange.

Parameter
PH
Cadmium
Zinc
Copper
Lead
Nickel
Chromium
Number of
samples
Average Total
Concentration
In Retention
Basin (^g/l)
5.97
1.0
6.4
16
22
2.3
3.4

34
Average Total Metal Concentration
Average Sediment
Concentration In
Top 0.8 cm (Mg/kg)
—
287
29,915
7,204
56,630
8,064
23,249

43
In
Groundwater Samples Collected Beneath
the Stormwater Retention Basin (^g/l)
0.1 m 0.5 m 1.0 m 3.0 m
6.01
1.3
11
7.6
26
2.2
4.3

8
6.17
1.3
14
8.8
27
3.2
6.2

8
5.79
1.6
10
8.5
15
1.5
2.0

8
4.91
1.3
11
8.7
13
2.3
1.5

8

6.0 m
5.02
1.0
12
10
12
2.3
1.5

8
 ing into ground waters. As seen in Table 2, the concen-
 trations of all heavy metals tested are near or below
 total concentrations measured in water within the
 retention  basin.  Metal concentrations  in  ground
 waters actually appear to decrease in some cases
 with increasing depth in spite of a decrease in pH. It
 appears, therefore, that the retention facility does not
 contribute measurable  increased  heavy  metal pollu-
 tion to underlying ground water.
 SUMMARY AND CONCLUSIONS

 During this research  investigating the fate of heavy
 metals  in  a stormwater retention facility,  both  the
 horizontal and vertical migrations were measured and
 modeled. Groundwater monitoring wells were also in-
 stalled on the edges of the retention pond to monitor
 heavy metal movement  out of  the  retention basin.
 From these studies the  following conclusions were
 reached:
   1. Heavy metals associated with stormwater runoff
 originating from highway surfaces are predominately
 particulate in form.
  2. Upon entering stormwater retention basins most
 particulate forms of metals settle near the point of in-
 put. Lead and zinc were found to reach peak concen-
trations in  sediments 15  m from the inlet, chromium
was found to reach a peak 30 m from the inlet, while
copper and  nickel  tended to settle out over a larger
area. Sediment  concentrations  of all  heavy metals
tested appeared to approach background concentra-
tions at a distance of 120 m.
  3. Dissolved species of heavy metals contained in
stormwater runoff were also removed in the retention
 basin. Apparent removal of dissolved species ranged
 from 27 percent for cadmium to 88 percent for zinc.
   4. Heavy metals tend to accumulate in  the  sedi-
 ments of the retention basin. The vertical migration of
 heavy metals in the sediments was found to observe
 an exponential decay with  a  rapid attenuation  rate.
 Most heavy metals were found to be present  in the top
 0.8  cm  with normal sediment background levels
 observed at a depth of 6.8 cm.
   5. In spite of the accumulation of heavy metals in
 the sediments of the retention pond no increases in
 groundwater concentrations of heavy metals  were
 observed in monitoring wells beneath the pond.
REFERENCES

Guthrie, R.K., and D.S. Cherry. 1979. Trophic level accumula-
  tion of heavy metals in a coal ash basin drainage system.
  Water Res. Bull. 15(1):244-8.

Hem, J.D., and W.H. Duram. 1973. Solubility and occurrence
  of lead in surface water. J. Am. Water Works Ass. 65:562-7.

Shaheen, D.G.  1975. Contributions of urban roadway usage
  to water pollution.  EPA-600/2-75-004.  U.S. Environ. Prot.
  Agency, Washington, D.C.
Wilber, W.G., and J.V. Hunter. 1977. Aquatic  transport of
  heavy meals in the urban environment. Water Res. Bull
  13(4): 721-34.

Winters, G.L, and J.L. Gidley. 1980. Effects of roadway run-
  off on algae. FHWA/CA/TL-80/24. Washington, D.C.
Yousef, Y.A., M.P. Wanielista, T. Hvitved-Jacobsen, and H.H.
  Harper. 1983. Fate of  heavy metals in  stormwater runoff
  from highway bridges. Proc. Int. Symp. Highway Pollution.
  Elsevier Sci.  Publ., Amsterdam, Holland.
                                                  334

-------
A PROBABILISTIC EVALUATION OF INSTABILITY
IN HYPEREUTROPHIC SYSTEMS
DANIEL G. BURDEN
RONALD  F. MALONE
Department of Civil  Engineering
Louisiana State University
Baton Rouge, Louisiana


             ABSTRACT

             Water quality data on six hypereutrophic urban lakes in Baton Rouge, La., have been collected
             on a regular basis over a 4-year period as part of a cooperative restoration effort by the U.S. En-
             vironmental Protection Agency and the city/parish of East Baton Rouge through the Clean Lakes
             Program. Thirteen water quality parameters were measured during most of the sampling events,
             including total phosphorus (TP) and dissolved oxygen (DO). Concentrations for TP and DO during
             the pre-restoration period ranged from 0.136  mg/l to 1.340 mg/l and 0.40 mg/l to 25.30 mg/l,
             respectively. This observed variability during pre-restoration illustrates the instability present
             with these systems.  Inherent with this instability  is the occurrence of water quality problems,
             such as summer fishkills. A probabilistic approach was developed to describe the inherent
             variability typical of these systems. The use of a probability density function to estimate overall
             summer kill risk is also discussed. This technique was applied to one of the smaller lakes in the
             system to evaluate restoration efforts. Overall projected fishkill frequency was reduced from ap-
             proximately eight kills per year to one kill in 7 years following restoration. Such approaches can
             be used in management to evaluate data more readily where standard procedures are time con-
             suming and funds unavailable.
 INTRODUCTION

 Hypereutrophic systems have been defined as ecosys-
 tems that are disturbed and unstable (Barica, 1980).
 These systems are characteristically shallow and un-
 stratified, and thus are subject to radical transforma-
 tions because of changes in light, wind, and temper-
 ature (Uhlmann, 1978). Inherent  with these systems
 are unbalanced nutrient and dissolved oxygen cycles
 resulting from these transformations (Barica, 1978).
   Stability   in  hypereutrophic systems  can  be
 measured as the variability of a critical parameter,
 such as total phosphorus, through time in relation to
 its mean value (Uhlmann, 1978). This variability is
 significant in the management of water quality, par-
 ticularly when dealing with lake system responses
 such as summer fishkills (Barica, 1978). A statistical
 approach for handling the inherent  variability ex-
 hibited by hypereutrophic systems has improved deci-
 sionmaking capabilities related to water quality prob-
 lems. The approach uses a probability density func-
 tion for describing lake quality in terms of a controll-
 ing  parameter. The resulting parameter distribution
 can be used to characterize changes in  lake condi-
 tions and provide a quantitative basis for manage-
 ment decisions.
   It is the objective of this paper to illustrate how pro-
 babilistic density functions have been applied to inter-
pret data on the University Lakes system and how they
can assist management of similar lake systems. Rela-
tionships between certain parameters and probability
distributions were developed for evaluating restora-
tion  efforts, in-lake  changes, monitoring  programs,
and  summer fishkill frequencies for this  southern
hypereutrophic lake  system.
SITE DESCRIPTION AND WATER
QUALITY MONITORING

The University Lakes System consists of six hyper-
eutrophic lakes, located approximately 1.5 miles east
of the Mississippi  River and 2.5  miles southeast of
downtown Baton Rouge, La., occupying a total area of
118.7  ha (293.1 acres) (Fig. 1, Table 1). Historically,
these  lakes have  been used  for  recreational  and
educational purposes by the surrounding community
of Baton Rouge and Louisiana State University. High
biomass concentrations, warm climatological condi-
tions, and sewage contamination have caused hyper-
eutrophic conditions in these lakes. In 1977, a joint ef-
fort by the U.S. Environmental Protection Agency and
the city/parish of East Baton Rouge was initiated to
restore the University Lakes system. Restoration of
               Table 1.—Morphological characteristics of the University Lakes prior to dredging activities.
Lake
University
Campus
College
Crest
Erie
City Park
Mean Depth
(Meters)
0.61
0.46
1.10
1.52
0.61
0.73
Detention Time
(days)
39
16
39
561
14
25
Surface Area
(hectares)
89.20
2.95
2.14
3.43
1.21
21.00
Drainage Basin
(hectares)
332.70
41.20
28.26
2.22
24.63
209.00
                                                 335

-------
 LAKE AND RESERVOIR MANAGEMENT
 the lakes included deepening by hydraulic dredging,
 diverting runoff, eliminating sewage contamination,
 and  altering interlake  connections.  Presently,  all
 dredging activities on the lakes have been completec.
 Sealing of sewage leaks and bank stabilization is ex-
 pected to be completed by the fall of 1984.
   Water quality monitoring on the University Lakes
 has been conducted on  a  monthly basis since Juno
 1979 and supplemented by twice-a-month sampling
 during  summer  months. An additional weekly sampl-
 ing program was initiated on Crest Lake in June 1981
 and continued through December 1982. Data collec-
 tion for this routine monitoring work consisted mainlv
 of water samples taken at a nearshore station in each
 lake .3  m (1 ft) below the surface and .3 m (1 ft) above
 the bottom.
   Thirteen water quality parameters were measured
 during  most of  the  sampling  events, including total
 phosphorus (TP) and dissolved oxygen (DO). Chemical
 determinations were performed in triplicate and in ac-
 cordance  with   procedures outlined  in  Standard
 Methods  (1980).  Vertical  dissolved  oxygen  and
 temperature profiles were taken at each sampling. A
 detailed description of the sampling regime is describ-
 ed elsewhere (Malone, et al., 1980).
   From June 18 through Aug.  10, 1982, an intensive
 daily sampling  program, independent of the routine
 monitoring  work,  was conducted on Crest Lake to
 evaluate the validity of various assumptions used in
 the application of common lake models. Data collec-
 tion consisted primarily of samples being taken from
 three mid-lake stations at  uniform depths of .15, .9,
 and 1.5 m below the surface. Determinations of total
 phosphorus were performed on the same day of  col-
 lection  and were subject to replicate analysis as in the
 routine sampling program. A similar intensive sampl-
 ing program was conducted from June 7 through Aug.
 14, 1983, on Campus Lake to evaluate post-dredging
  College Lak,
Figure 1.—The University Lakes System located in Baton
Rouge, Louisiana.
 effects in this system. An outlined methodology for
 the intensive monitoring programs is presented else-
 where (Mericas and Malone, 1983a).
 PARAMETER DISTRIBUTION

 Ahmed and Schiller (1980) defined trophic status of
 lakes  probabilitically  by  a  distribution of  total
 phosphorus  data.  An  analogous  approach  was
 developed for quantifying  data  and evaluating lake
 system responses in the University Lakes by using a
 probability  density  function.   Variability in  total
 phosphorus concentrations resulting from a variety of
 sources  is well represented  by  this technique. The
 resulting parameter distribution can be used to quan-
 titatively evaluate the impact of the variability  upon
 the  system.  Natural  variability  in total  phosphorus
 levels as well as uncertainties stemming from the ap-
 proximate nature of sampling programs and analytical
 imprecision all contribute to the variance of the  para-
 meter  distribution. Management decisions can thus
 be based  upon a single quantitative measure that ac-
 curately represents both the condition of the system
 and reliability of the monitoring data.
   Several types  of continuous probability density
 functions have been employed in the analyses of data;
 for example, uniform, normal, and gamma distribu-
 tions are a few. Of those mentioned, normal and  gam-
 ma  distributions  have  demonstrated  the  best
 capabilities for  evaluating the University Lakes on
 seasonal and annual basis. To determine what type of
 continuous distribution should be  used, a histogram
 (or discrete probability distribution) is formulated us-
 ing the data. The  histogram is produced by plotting
 the frequency of occurrences within each interval of
 data  (or  parameter concentration)  as  a  bar  that
 represents the number of occurrences, or frequency,
 of the data for thai interval. Figure 2 illustrates the use
 of a histogram for quantifying total phosphorus data
 from all lakes in the University Lake system from July
 1979, to July 1982 (Mericas, 1982). Examination of the
 graph suggests that the data is not evenly distributed
 for the range of phosphorus concentrations. To char-
 acterize the data  in  a more useful format,  a  con-
 tinuous probability distribution must be developed in
 a manner that accurately represents the histogram. A
 gamma probability density function (Benjamin and
 Cornell, 1970) was chosen in the  form of Equation 1:
                                                         fx(c) =
                        forc>0
                                                                                                   (1)
                                                     where
                                                           c = water quality parameter concentration
                                                          k, A = empirical constants
                                                       The empirical  constants  are readily derived from
                                                     simultaneous solution of  Equations 2 and 3 relating
                                                     them to the mean and variance of the data set.
                                                                 k
                                                           mc  = T                                 (2)
                                                                                                   (3)
                                                     where
                                                         m
          =  mean concentration of the water quality
            parameter
    °c   =  variance associated with the water qual-
            ity parameter
  The  gamma  function  illustrates  the  data  in a
statistical format by exclusive use of the mean and
standard deviation from the TP data during that time
                                                336

-------
                                                                                       URBAN LAKE QUALITY
 period (Fig. 3). An important feature of a distribution
 function such as this is that the probability of an event
 occurring  between  any two points on  the curve is
 represented by the total area under the curve between
 those two points. Additionally, the gamma distribution
 takes the left tail of the curve through the origin, which
 is consistent with the observation that the concentra-
 tion of a water quality parameter is always greater
 than or equal to zero.
   Conversely, if the histogram had exhibited a  bell-
 shaped distribution for the data, a normal or Gaussian
 probability density  function (Benjamin  and  Cornell,
 1970)  would  have  been  used.  This  distribution  is
 described  by Equation 4:

          1   ex  r_ i (c-mc\or_^^^   (4)
where
     oc = standard  deviation  associated with the
           water quality parameter

  The normal distribution is frequently used to repre-
sent  random  errors,  such as those resulting  from
analytical error. It is useful for parameter distributions
whose mean values are large enough in comparison to
the variance  to prevent projection  of significant fre-
quency of negative values. Thus, care should be taken
in selecting the type of function used to represent the
system. Additionally, comparison of parameter distri-
butions should be conducted in a  manner represen-
tative of similar sampling conditions.

Interpretation

The  use  of a parameter distribution permits an ac-
curate representation  of  a  hypereutrophic  system.
               TOTAL PHOSPHOROUS MIDPOINT ( MO/L)
Figure 2.—Histogram of all total phosphorus data collected
on the University Lakes from July 1979 through July 1982
(Mericas, 1982).
                  TOTAL PHOSPHORUS (MG/L)
Figure 3.—Gamma probability density function representing
all total phosphorus data from the University Lakes from July
1979 through July 1982 (Mericas, 1982).
                                                      This reflects the ability of the distribution to represent
                                                      the natural variability characteristic of these systems.
                                                      This concept is hypothetically illustrated in Figure 4
                                                      by comparing three normal probability distributions
                                                      with the same mean TP value but different variances.
                                                      Although the mean value hasn't changed, the range of
                                                      TP values increases with the variance as illustrated by
                                                      the lateral spread in the lower portion of the curves.
                                                      The range of values may be as important as the mean
                                                      since adverse conditions, such as fishkills and odor
                                                      problems, are most often associated with intermittent
                                                      peaks (or conversely, minimums) in water quality para-
                                                      meters.
                                                        A parameter distribution also can reflect variation
                                                      in  the  mean  lake condition. Figure 5 illustrates this
                                                      capability in a theoretical manner. Such  changes in
                                                      the mean reflect changing trophic level such as might
                                                      be expected from morphological or loading alterations
                                                      resulting from, for example, a restoration effort. Thus,
                                                      using the parameter distribution enhances one's abili-
                                                      ty  to conceptually and quantitatively evaluate a lake
                                                      system's condition  by providing  a more complete
                                                      description of the lake.

                                                      Applications

                                                      Probability density functions have numerous applica-
                                                      tions in water quality analyses. In many instances a
                                                      simple visual analysis of the data over a given time
                                                      period  may prove to be inadequate. Figure 6 illustrates
                                                        250


                                                        225


                                                        200
                                                      u- I25
                                                      t-
                                                      £ 75
                                                      0.

                                                        50


                                                        25
                                                                                 m -0350
                                                                                 CTj • 0 035
                                                                                     m,= 0350
                                                                                     Oi-0052
                                                          000 006 012 018  024 OL30 036 042 048 054 060066 072 078
                                                                      TOTAL PHOSPHORUS ( MG/L)


                                                      Figure 4.—Comparison of three normal probability distribu-
                                                      tions with the same mean total phosphorus values and dif-
                                                      ferent variances.
                                                      z  4
                                                         000 005 OIO 015 020 025 030 035 040 045 0.50 055 060 065 070
                                                                      TOTAL PHOSPHORUS (MG/L)


                                                      Figure 5.—Comparison of two normal probability distribu-
                                                      tions with different mean total phosphorus values and the
                                                      same variances.
                                                   337

-------
 LAKE AND RESERVOIR MANAGEMENT
 TP data collected in the routine monitoring program
 over a 2-year period for Crest Lake during which an in-
 terim  restoration technique  of  isolation  was  im-
 plemented.  Response of the system to the  initial
 restoration  effort tends to reduce peak TP  levels
 somewhat. However, the overall data during the period
 of  isolation  still  appears  highly erratic, clearly
 demonstrating the instability present within  the lake.
 By using a parameter distribution, the data can be put
 into a form that enables one  to readily quantify the
 state of the system.
   Figure 7 presents a comparison of parameter  distri-
 butions representing  two  phases of  the interim
 restoration effort undertaken  on Crest Lake. Thess
 functions  illustrate two  accomplishments of  ths
 restoration:  (1) mean TP levels  in  the system wers
 significantly reduced as a result of  the isolation, and
 (2) the system  tends to fluctuate less, as noted by ths
 variability, indicating an increased stability within.
   Figure 8 illustrates the  application of parameter
 distributions developed to examine the effect of  tlna
 dredging activities  upon  the DO levels in University
 Lake.  It was feared that suspension  of  the benthic
 muds  would deplete dissolved oxygen levels in  the
 water column.  The expected decrease in  average  DO
 was compensated for by an  increased stability of  tho
 system. Peak  oxygen  values were  controlled  by
 physical transfer phenomena during the dredging ac-
 tivity. Historically, the system had been dominated by
 oxygen  release  and  consumption  associated with
 algal growth and decay cycles; these were virtually
 eliminated by the high turbidity present during active
 dredging. The  parameter functions  clearly illustrate
                         JUL   OCT   JAN8

                      ! OF OBSERVATION
Figure 6.—Mean total phosphorus data collected on Crest
Lake over a 2 year period.
  2 5

  20
  I 0 -

  0 5 -
                       SOLATION PERIOD
                         mc = 0 264
                         crt , 0 066
                              — PRE- ISOLATION PERIOD
                                   mc = 0 354
                                   tre = 0 13 3
   000 005 010 015 020 025 03O 035 040 0.45 050 055 060 065 070 075 080
                  TOTAL PHOSPHORUS (MG/LI

Figure 7.—Quality distributions representing two phases of
the interim restoration technique undertaken on Crest Lake.
 the dramatic change in the nature of DO variations in
 the system during the dredging period.
   Additionally, parameter distributions can be used in
 verifying monitoring programs. Figure 9 illustrates a
 comparison of TP data collected under the routine
 monitoring and the intensive monitoring programs
 undertaken on Crest Lake during the summer of 1982.
 Neglecting the temporal variability present in the data,
 it can be observed that the two sampling programs are
 quite similar. These observations can be confirmed by
 applying the  t-test and F-test. Utilization of the para-
 meter distribution in this manner enabled the authors
 to conclude that the routine monitoring data were in-
 deed representative of the inlake conditions.


 LAKE MANAGEMENT AND  PARAMETER
 DISTRIBUTIONS

 Prior to restoration of the University Lakes system, 12
 summer fishkills were observed over a 3-year period in
 the lakes. These kills have been associated with domi-
 nant algal populations which, under certain meteoro-
 logical conditions, collapse,  and in  turn, deplete ox-
              v-*	DREDGING PERI03
              \     me = 5 0
              \     CT-, - 2 03
                                                            0 I 234 56 7 8 9 10 II 12 II 11 15 16 17 18 19 2021 22232425 2627282930
                                                                        DISSOLVED OXYGEN (MG/LI


                                                      Figure 8.—Comparison of quality distributions representing
                                                      dissolved oxygen data on University Lake during periods of
                                                      pre-dredging and dredging.
                                                          10
 > 8
 o
 •z.
 UJ 7
 o
 UJ .
 o 4
 cc
 t 3

    a

    i
                              INTENSIVE STUDY
                                                                                      ROUTINE MONITORING
           01     02    03     04    05

             TOTAL  PHOSPHORUS , MG/L
                                                                                                 06
Figure 9.—Comparison of quality distributions  for total
phosphorus data  resulting from  routine  and  intensive
monitoring programs undertaken on Crest Lake.
                                                  338

-------
                                                                                        URBAN LAKE QUALITY
ygen levels present in the system (Swingle, 1968). Con-
sequently, massive fishkills and odor problems result.
Management of water quality to avoid such conditions
requires a methodology permitting an evaluation in a
minimal period of time and at a  reasonable cost. Us-
ing parameter distributions permits such an  evalua-
tion.
   Mericas and Malone (1983b) presented an objective
function  or  a  "system  response  function"  for
evaluating the occurrence of summer  fishkills in the
University Lakes in conjunction with parameter distri-
butions. The  summer fishkills, which  predominantly
took place during periods of high productivity,  usually
summer months, resulted from complex meteorolog-
ical factors. Additionally, water quality analyses in-
dicated  all  kills  occurred at TP levels in  excess of
0.400 mg/l. The objective function (describing the pro-
bability  of a  fishkill on any given  day) was approx-
imated  by  a  uniform  distribution  based  on  total
phosphorous  concentration:
      P (FK/TP) = 0.030
      P (FK/TP) = 0
                            for TP>. 0.400
                            elsewhere
(5)
where
     P(FKTTP)  =  probability  of
                  given TP
                                fishkill  occurrence
  The probability of the occurrence of a fishkill can be
calculated for any  period  by considering the  inter-
section of the parameter distribution and the objective
function:
     (P(FK) =  P(FK/TP)flP(TP> 0.400)

where
                                                (6)
     P(FK) =  daily probability of fishkill
      P(TP>OAOO) =  probability of daily observation
                      exceeding 0.400 mg/l TP

   This technique was used for a preliminary evalua-
tion of post-restoration efforts on Campus Lake (Fig.
10). Development of the parameter distribution for the
period of pre-restoration consisted of summer TP data
collected in the routine monitoring  program  on Cam-
pus Lake over  a 4-year  period. Data used  with the
distribution for the post-restoration period were col-
lected during the first  28 days of an  intensive monitor-
ing program. Calculated fishkill estimations  for each
period are listed in Table 2. This preliminary  compari-
           A
              	POST RESTORATION
                  mc -02I5
                           3 0% PROBABILITY OF
                               FISH KILL
  60

  55

  50

z 45
Id

s"°
0-35

H 30

o 25

S 20
        Ol  02  03  04  05  06  07  06  09  10  II  12
                  TOTAL PHOSPHORUS ( MG/L)


 Figure 10.—Quality distributions for total phosphorus data
 collected on Campus Lake during periods of pre-dredging
 and post-dredging overlayed with the fish kill response func-
 tion as designated by the shaded area.
                                                      son  indicates  the  probability  for a  fishkill  was
                                                      significantly reduced by the restoration. Verification
                                                      of this finding will require more post-restoration obser-
                                                      vations. This  type of approach has been  used with
                                                      other lakes in the system and proven to be quite useful
                                                      for decisionmaking  processes (Mericas and  Malone,
                                                      1983b).

                                                         Table 2.—Results  of integration for the areas under
                                                         each distribution above the fishkill response function
                                                            and projected daily fishkill frequencies for each
                                                                             period.
Time Period
Pre-dredging
Post-dredging
P(TP>0.4)
0.7164
0.0117
P(FKm>>0.4)
0.030
0.030
P(FK)
0.0215
0.0004
CONCLUSIONS
The use of parameter distributions in conjunction with
objective  functions  representing  adverse  system
responses have proven  to  be a useful management
tool  for the University Lakes system. The parameter
distributions can be formulated from routine monitor-
ing data according to well-accepted statistical theory.
These provide the user  with both a quantitative  and
pictorial tool for interpreting results. The gamma func-
tion  has proven particularly applicable to the  hyper-
eutrophic University  Lakes  because of the  skew
characteristic of  data  collected from this  system.
Functions based on total phosphorus have been most
often used to influence management policy.

ACKNOWLEDGEMENTS:  Funding for  this research  was
derived in part from  the U.S. Environmental Protection Agen-
cy, City/Parish Government of East Baton Rouge, and the
State of Louisiana through a cooperative lake restoration ef-
fort under the Clean Lakes  Program. This paper was not sub-
ject to review by the funding agencies. Findings of this paper
reflect the opinions  of the authors only. The data base used
in this paper includes contributions by Glenn McKenna, Con-
stantine Mericas, Andrew Eversull, and  Paul Gremillion.

REFERENCES
Ahmed, R., and R.  Schiller.  1980. Quantification of phos-
   phorus input to lakes and its impact on trophic conditions.
   In  Restoration  of  Lakes  and Inland Waters.  EPA
   440/5-81-010. U.S.  Environ. Prot. Agency, Washington, D.C.
Barica,  J. 1978.  Collapse of Aphanizomenon flos-aquae
   blooms  resulting  in massive fish kills in eutrophic lakes:
   Effect of weather. Verh.  Int. Verin. Limnol., 20: 208-13.
	1980. Why hypereutrophic systems? In Develop-
ments in Hydrobiology. Vol. 2. W. Junk Publisher, The Hague,
Netherlands.
Benjamin, J.R., and C.A. Cornell. 1970. Probability, Statistics,
   and Decisions for Civil Engineers. McGraw Hill, Inc. New
   York.
Malone, R.F., H.  Saidi, and  G.  McKenna. 1980. University
   Lakes Restoration Project. Annu. Water Qual. Monitor.
   Rep. Inst. Environ. Stud, Louisiana State Univ.
Mericas, C.E. 1982. Phosphorus dynamics and the control of
   eutrpphication  in a southern urban lake. Master's thesis,
   Louisiana State Univ.
Mericas, C.E., and R.F. Malone. 1983a.  Mathematical repre-
   sentation of short term  phosphorus  variations in a non-
   stratified southern lake. Environ. Monitor. Assess. (In
   press.)
	1983b. A phosphorus based fish kill response func-
   tion for  use with  stochastic lake models. N. Am. J. Fish.
   Manage. (In press.)
Standard Methods for the Examination of Water and Waste-
   water. 1980. 15th ed. Am.  Pub. Health Ass. Am. Water
   Works Ass., Water Pollut. Control Fed., Washington, D.C.

Swingle,   H.  1968.  Fish  kills caused by phytoplankton
   blooms and   their  prevention.  Food Agric.  Organ.
   44(5):407-11.
Uhlmann,  D. 1978.  Stability and multiple  steady states of
   hypereutrophic ecosystems. Develop. Hydrobiol. 2:235-47.
                                                    339

-------
 OCCURRENCE AND CONTROL  OF  TASTE AND
 ODOR IN SYMPSON  LAKE
 G. C. HOLDREN
 R. MAJOR WALTMAN
 Water Resources Lab
 University of Louisville
 Louisville, Kentucky


             ABSTRACT
             For the past few years Bardstown, Ky., has experienced taste and odor problems with its drinking
             water. The problems originated in Sympson  Lake, the raw water source for the city, but the exact
             cause was unknown; a chemical and biological investigation was undertaken in 1982 to determine
             the source of the problems. Chemical analysis indicated that Sympson Lake experiences hypolimnetic
             oxygen depletion during summer stratification. Although manganese concentrations reached high levels,
             no chemical cause for the taste and odor was found. Algal counts indicated that maximum taste and
             odor complaints coincided with increases in diatom population, especially Stephanodiscus. The max-
             imum diatom count was only 126 cells/ml, much lower than counts reported to cause problems in
             previous studies, but no complaints were noted when other algal species predominated  Several dif-
             ferent control methods were investigated. Copper sulfate was applied in April 1982 to control diatom
             population. Athough a decline in diatom counts was noted after the CuSO4 application, a concurrent
             reduction in soluble reactive phosphorus to < 1 ^g/l and rising water temperature may have also con-
             tributed to the decline. Aeration, alum treatment, nutrient control, and potassium permanganate plus
             activated carbon were also investigated as control techniques. Alum and activated carbon/potassium
             permanganate were judged to be both ineffective and expensive. Aeration is relatively inexpensive
             but would be expected to have little or no effect on the spring diatom bloom. Control of influent nutrients
             should prove effective, but further study wou d be required to determine the costs involved. In the
             meantime, CuSO4 appears to be the cheapest and most effective treatment method.
INTRODUCTION

Sympson Lake was  formed in 1962 by the impound-
ment of Buffalo Creek, Nelson County, Ky. The laks
serves as the sole source of raw water for the Bards-
town Public Water Supply System and is also used fcr
recreation. Some characteristics of Sympson Lake ars
listed in Table 1.
  Water quality in Sympson Lake has  generally been
quite good but the water utility has experienced occa-
sional taste and odor problems for the past few years.,
primarily during the  spring and fall months. The pro-
blems  originated  in Sympson  Lake,  but the exact
cause was unknown. Water samples from Sympson
Lake were collected  from January through November
1982, to examine the  water chemistry of Sympson
Lake, determine the cause of the taste and odor prob-
lems, and evaluate various treatment alternatives.
EXPERIMENTAL PROCEDURES
Sampling and Chemical Analysis

Water samples from Sympson Lake were collected ap-
proximately monthly from January to November 1982,
for chemical analysis and algal counts. A Van Dorn
sampler was used to obtain water from depths of 0,2,
5, 10, 15, and 17 m near the intake  structure for the
water treatment plant on each sampling date.
  Monthly  measurements  included  temperature,
dissolved oxygen (DO), conductivity,  pH,  alkalinity,
total organic  carbon  (TOG),  calcium, magnesium,
sodium, potassium, iron, manganese, soluble reactive
phosphorus  (SRP),  total phosphorus,  nitrate  and
nitrite (NOa  +  N02" -  N), ammonia-N, and Secchi
depth. Chlorophyll a was also measured, but less fre-
quently.  All analyses  were completed  using pro-
cedures found in Standard Methods (1980) or U.S. EPA
(1979).
Algal Enumeration and Identification

An  adaptation of  the  filter  technique  (Standard
Methods, 1980) was used for plankton enumeration in
this study. Water samples were preserved in a final
concentration of 6 percent gluteraldehyde. The count-
ing procedure began with filtering 100 ml of preserved
sample  through  a 0.45 j^m membrane filter. Filters
were immediately washed with 40 to 60 ml of 50 per-
cent ethanol followed by three rinses with 40 to 50 ml
of 95  percent ethanol. Wet filters were  placed  on a
glass  slide with a drop of clove oil between the filter
and slide. Another drop of clove oil was placed on top
of the filter, followed by the cover slip. The preparation
was allowed  to sit for 24 to 48 hours to allow the clove
oil to render the filter transparent. Algae were counted
in seven or more fields and results were combined to
calculate the number of algal cells/ml of sample.

RESULTS AND DISCUSSION

Examination  of  the  water  quality  parameters
measured during this study indicated Sympson  Lake
could  be classified as a hardwater,  monomictic lake
showing  some signs of eutrophication.  Complete
results of the  chemical and physical analyses  were
presented by Waltman (1982) and are summarized in
Table 2.
                                                340

-------
                                                                                   URBAN LAKE QUALITY
  Sympson Lake was thermally stratified from May to
November,  with  oxygen  depletion  occurring   im-
mediately after stratification  in the bottom waters.
Maximum oxygen depletion was noted in September,
when oxygen was absent from all water below 3 m and
iron, manganese, and ammonia reached maximum
levels.  The presence of these compounds  is often
associated with taste and odor problems in reservoirs,
but Sympson Lake is equipped with a variable-depth
intake structure and no hypolimnetic water was with-
drawn during the summer months. The lake developed
an ice cover during January 1982, but the cover did not
last long enough for winter stratification to occur and
iron  and manganese  concentrations were  below
detection limits during the winter and spring months.
These  results indicated that chemical causes were
not  responsible for the taste and  odor  problems
detected.
  Taste and  odor problems in Sympson Lake were
found to coincide with large increases in the popula-
tions of the diatom, Stephanodiscus. This species has
been shown to cause taste and odor problems in many
lakes  and produces a musty,  geranium-like  odor
(Palmer, 1977) that has occurred in Sympson Lake dur-
ing the spring and fall.
  The occurrence of Stephanodiscus is tied closely to
the annual temperature cycle  in lakes and explains
the  seasonal  pattern of taste and  odor problems.
Studies summarized by Hutchinson (1967) indicate the
optimum temperature  range  for  the  growth  of
Stephanodiscus is 4 to  12°C,  with  maximum popula-
tions occurring at 5 to 7°C. Diatom  blooms can there-
fore be expected in early  spring after ice out, or as
soon as the  water  begins to  warm in ice-free lakes,

  Table 1.—Morphologic and hydrologic characteristics
                 of Sympson Lake.
Draingage basin area
Lake surface area
Lake volume
Mean depth
Maximum depth
Hydraulic residence time
2.28 x 107 m2
6.8 x 105 m2
5.44 x 106 m3
8.0 m
18.3 m
160 days
                and again in late fall. In Sympson Lake this tempera-
                ture range was found in surface waters in March and
                April and late  November, coinciding with taste and
                odor complaints.
                  The total diatom concentration reached a maximum
                of  126 cells/ml in March and total algae concentra-
                tions peaked at 143 cells/ml in May (Fig. 1). These con-
                centrations  are well below the densities of 1,000 or
                more cells/ml often associated with algal blooms in
                eutrophic lakes (Hutchinson, 1967). Because diatom
                populations were not extremely large,  it is possible
                that the aquatic actinomycetes often associated with
                algal populations may have contributed to the taste
                and odor problems (Symons, 1956a); however, taste
                and odor  problems  occurred  only  when  diatom
                populations  were  at maximum  levels,  indicating
                diatoms were the primary cause of the observed prob-
                lem.
                  Nutrient dynamics  in Sympson Lake further favor
                the growth  of diatom populations during the spring
                and fall. Maximum whole-lake nutrient concentrations
                of 0.040 to 0.057 mg/l total phosphorus and 1.2 to 2.8
                mg/l total inorganic nitrogen, here defined as the sum
                of NOJ + NC>2  -  N and ammonia  - N, were observed
                                                           -40
                                         o  Total Algae

                                         •  Total Diatoms
                                         a  Secchi Depth
      JFMAMJJA  S   OND
                     Month
Figure 1.—Algal populations and Secchi depth in Sympson
Lake, 1982.
                         Table 2.—Water quality parameters for Sympson Lake, 1982.
Parameter (units)
Temperature (°C)
Dissolved oxygen (mg/l)
Conductivity (mS/m)
pH
Alkalinity (mg/l as CaCOs)
Total organic carbon (mg/l)
Calcium (mg/l)
Magnesium (mg/l)
Sodium (mg/l)
Potassium (mg/l)
Iron (mg/l)
Manganese (mg/l)
Soluble reactive phosphorus (mg/l)
Total phosphorus (mg/l)
Nitrate + nitrite - N (mg/l)
Ammonia - N (mg/l)
Secchi depth (m)
Chlorophyll a (^g/l)
Annual Mean
12.8
7.4
35.9
8.0
150
2.3
35.3
19.8
2.0
3.4
<0.1
0.3
0.004
0.035
0.96
0.18
1.6
—
Range
1.0-30.0
<0.5- 14.3
29.8 - 43.0
6.8 - 8.7
138-200
1.5-3.1
29.3 - 43.9
16.0-27.1
1.8-6.7
0.8 - 7.6
<0.1 -0.6
< 0.1 • 3.2
< 0.001 -0.102
0.008 - 0.510
0.01 - 3.22
<0.1 - 1.9
0.5 - 3.1
1.2-8.1
                                                 341

-------
 LAKE AND RESERVOIR MANAGEMENT
 during late winter and early spring. Average rainfall in
 this area is at a maximum during the spring months
 (Natl. Oceanic Atmos. Admin. 1982) and runoff from
 the  relatively large drainage basin (Table  1) is ex-
 pected to be the primary nutrient source for Sympson
 Lake.
   Concentrations of soluble nutrients in the surface
 water (Fig. 2) indicate that nutrient limitation may con-
 trol both algal growth and taste and odor in Sympson
 Lake. Surface concentrations of soluble reactive P
 and total inorganic nitrogen dropped quickly during
 spring diatom  bloom. Soluble reactive P dropped ':o
 less  than  0.001   mg/l  in  early  April,  indicating
 phosphorus  limitation may have contributed to  the
 collapse of  the diatom bloom; however, treatment of
 the lake with copper sulfate (CuS04) in mid-April com-
 plicated  evaluation of these results. Total inorganic
 nitrogen concentrations  in the epilimnion also droD-
 ped during the summer, reaching levels of 0.07 mg/l n
 late August and 0.06 mg/l in early November as algal
 uptake and the subsequent deposition of particulars
 to the hypolimnion removed nutrients from  surface
 waters.
   Soluble reactive P increased in the hypolimnion dur-
 ing  the summer  months (Fig.  2). Hypolimnet c
 phosphorus  is expected to be relatively unavailable
 for algal growth during the summer, but  would mix
 with surface waters during the  fall  overturn period,
 when a combination  of  low temperatures and high
 nutrient  concentrations  would  again  favor  diatom
 growth.
Treatment Alternatives

Treatments to  remove  taste  and  odor  problems
associated with algal growth can be divided into three
general categories: treatment to remove the taste and
odor-causing compounds by produced algae, treat-
ment to remove the algae before a problem develops,
and treatment  to  reduce nutrient inputs  to prevent
algal growth.  The cost  of treatment and reliability
must be evaluated before a method is chosen. Several
different treatment methods for taste and odor control
were evaluated for Sympson Lake and are summariz-
ed in Table 3.
  The Bardstown water treatment plant had used ac-
tivated carbon to deal with minor taste and odor prob-
lems in the past; however, these were so severe in the
                           fall of 1981 that activated carbon alone could not con-
                           trol the problem  and potassium  permanganate was
                           added to the treatment process. A combination of 3.0
                           mg/l  activated carbon and  0.5 mg/l potassium per-
                           manganate had little effect on the taste and odor
                           problems. Based on a previous study (Dougherty and
                           Morris, 1967), a combination of 4.5 mg/l activated car-
                           bon and 1.5 mg/l  potassium permanganate was sug-
                           gested for future treatments with  this method.
                             Treatment costs were  estimated using chemical
                           costs (1982 prices) provided by the water utility and
                           local  chemical suppliers. Total treatment cost for both
                           potassium  permanganate $2.16/kg ($118.50/100 lb)and
                           activated carbon $2.85/kg ($64.30/50  Ib) was estimated
                           at $126/day for the 7600  m3/day  (2  MGD) treatment
                           plant. Based on past experience, it was estimated that
                           approximately 60 days treatment would be required
                           each  year,  resulting  in  an annual  cost of  $7,560.
                           Because of limited  effectiveness in the past, and
                           because treatment costs for this method were higher
                           than some more effective methods, treatment with ac-
                           tivated carbon and potassium permanganate was not
                           recommended.
                             Aeration  is another method that  has been  widely
                           used  in lakes with taste and odor  problems;  it was
                           considered  as  a treatment alternative for Sympson
                           Lake.  Aeration  prevents hypolimnetic oxygen  deple-
                                    j   F'MA'M'J'J  ' A'  s ' o ' N ' o
                           Figure 2.—Nutrient concentrations in Sympson Lake, 1982.
                             Table 3.—Treatment alternatives for Sympson Lake.
        Method
                               Approximate Cost
                                    Benefits
                                                          Disadvantages
Alum treatment
Nutrient input
reduction
Aeration
Copper sulfate
Potassium permanganate
and activated carbon
$28,000-$84,000, depending
on volume of lake treated.
Cannot be determined without
additional information.
$16,000 initial investment plus
$3000-$4000 annual operating
cost.

$2700 per treatment with 1-3
treatments required/year.
$7560 per year, based on 60
days treatment/year.
                                                        Reduction of internal
                                                        nutrient loading.
Could reduce algal
blooms and associated
problems. No periodic
treatments required.

Elimination of high
hypolimnetic  Fe and
Mn concentrations.

Temporary elimination
of algal growth and
associated problems.

Removal of taste and
odor at treatment plant.
High initial cost.
Other nutrient sources
must also be removed.

Possible success can-
not be determined at
this time.
May not eliminate
taste and odor
problems.

Periodic treatments
required.
Has not been completely
successful in the past.
                                                342

-------
                                                                                   URBAN LAKE QUALITY
tion and, therefore, reduces hypolimnetic concentra-
tions of  iron and  manganese and internal loading.
Aeration  may also reduce the number of blue-green
algae, but does not always significantly change algal
populations; the changes cannot be predicted in ad-
vance (Pastorok et al. 1981).
  Based  on the experience of the Wisconsin Depart-
ment of Natural  Resources (Wedepohl, pers. comm.),
aeration  systems can  be  expected to  cost  approx-
imately $16,000  for initial  installation and $3,000 to
$4,000 in  annual operating expenses. This would make
aeration  competitive with  other treatment methods,
but the effectiveness of aeration in controlling taste
and odor in Sympson Lake may be limited because
Sympson Lake does not have a large population of
blue-green algae (Waltman, 1982) and the taste and
odor was not readily removed by air stripping.
  Lake treatment with CuSO4 has been widely used to
control nuisance algal growth. Copper sulfate has the
advantages  of being toxic  to most  nuisance
organisms at  low  concentrations, nontoxic  to  fish,
and not harmful to the general aquatic environment in
concentrations  normally  used  (Tuwiner, 1975).  A
CuSO4 concentration of 0.1 to 0.5 mg/l has been found
sufficient for  control of diatoms (Symons,  1956a,
1956b).
  Because Sympson Lake has relatively high alkalini-
ty (annual mean  =  150 mg/l as CaCO3) and total hard-
ness (annual  mean  =  170  mg/l  as  CaC03), no
deleterious  effects were  expected  at  the  copper
sulfate concentration  of 0.5  mg/l (as CuSO4«5H2O)
selected  for treatment. Based on this concentration
and a cost of $1.10/kg ($50/100 Ibs)  for commercial
CuSO4»5H2O, an application  of CuSO4 costs $2,700
for  Sympson Lake.  It is estimated that one to three
treatments per year would  be required, depending on
weather conditions.
  Sympson Lake was treated with CuS04 on April 12
and 15, 1982, and  again in November 1982.  In  both
cases taste  and  odor  problems  were eliminated;
however, water analysis on April 16 indicated soluble
reactive P was below detection limits and the surface
temperature was 12°C.  It is therefore possible  that
nutrient limitation  and  rising temperature  may have
contributed to the decline  in algal populations and
that CuSO4 was not solely responsible.
  Determination of the optimum application time is
one of the problems associated  with  CuSO4 treat-
ment. A comparison of algal counts and Secchi depth
(Fig.  1)  with the  optimum temperature  range for
Stephanodiscus (Hutchinson, 1967) indicates the  best
time for CuSO4 application in Sympson  Lake occurs
when Secchi depth is reduced to  less than 1 m and
temperature is between  4 and 12°C. These two para-
meters  were  chosen   because  they   are  readily
measured by treatment  plant personnel. If it can be
determined that nutrient limitation is likely or that the
temperature is outside  of the 4 to 12°C  range, treat-
ment may  not be necessary. These guidelines were
used in spring 1983, to  determine that  a CuSO4 ap-
plication  was unnecessary.
  Nutrient removal may be the most desirable treat-
ment method because  it is  intended to prevent the
development of  algal blooms and does not  involve
continuing chemical treatment of either the lake or the
raw water at the treatment  plant. The  reduction of
nutrient inputs from runoff could prove successful in
eliminating algal bloom  problems in  Sympson Lake.
The high  spring nutrient concentrations mentioned
previously indicate that spring runoff  may be the ma-
jor  nutrient source at  present.  Because  the  algal
blooms are not exceptionally large and because solu-
ble reactive phosphorus concentrations  in surface
waters dropped to undetectable levels following the
spring diatom   bloom,  reduction  of  phosphorus
loading may eliminate future algal blooms or reduce
them  to the point that additional problems are not en-
countered. As an  added benefit, oxygen depletion  in
the hypolimnion would  diminish if algal blooms were
not present to provide a source of decomposable
organic matter.
  A survey of land use and farming practices in the
Sympson Lake drainage basin would help determine
the type of nutrient controls required and the chances
for success. It is  possible that voluntary changes  in
fertilizer use and  the pasturing of livestock in areas
surrounding the  lake could reduce  nutrient  inputs
below the critical levels required for algal  growth. If
not, controlling particulate matter in  runoff by using
small settling  basins on influent streams could be
successful. Unfortunately, it is not possible to deter-
mine  either the cost or effectiveness of  such a pro-
gram  at present,  but nutrient input reductions could
prove beneficial to Sympson Lake and should be in-
vestigated further.
  Alum treatment to reduce internal nutrient loading
was another method suggested for treating Sympson
Lake. Alum treatment is primarily intended to prevent
internal loading after other nutrient sources have been
removed, but examination of nutrient data for Symp-
son Lake indicated runoff was the major nutrient
source and internal  loading was relatively unimpor-
tant. Alum treatment is  also relatively expensive. Bas-
ed on previous studies  by Garrison (1980) and Cooke
and  Kennedy  (1981),  it  was estimated  that  an
aluminum dose of 5 mg/l would be required for treat-
ment. At an alum cost of $147/tonne ($133/ton) with
4.85 kg Al/tonne, the total  treatment cost would be
$28,000 to $84,000 depending on whether the hypolim-
nion alone or the whole lake were treated. This high
cost and limited chances for success eliminated alum
as an alternative treatment.
CONCLUSIONS
Taste and odor problems in Sympson Lake were found
to be associated with the presence of relatively large
populations of Stephanodiscus, although the extreme-
ly large  populations  often associated with  algal
blooms were not observed. Investigation of  various
treatment alternatives indicated that the  best results
could be  expected from methods designed to reduce
diatom populations in the  lake, and that  treatment
with copper sulfate was the most cost-effective of the
methods  evaluated.  Copper sulfate application when
Secchi depth is less than 1 m and water temperatures
are between 4 and 12°C appears to be the most effec-
tive treatment method for the short term. Control of
nutrient  inputs  from  runoff  may be an effective
method for algal control  in the future because of the
limited concentration of nutrients in the lake,  but fur-
ther study would  be required  to determine the cost-
effectiveness of this alternative.
ACKNOWLEDGEMENTS: This  research was  supported  in
part by a grant from the city of Bardstown, Kentucky. Addi-
tional assistance was provided by the Graduate School and
the College of Arts and Sciences, University of Louisville.
                                                343

-------
LAKE AND RESERVOIR MANAGEMENT
REFERENCES

Cooke, G.D., and R.H.  Kennedy. 1981.  Precipitation and
  inactivation  of  phosphorus as a  lake restoration tech-
  nique.  EPA-600/3-81-012. U.S. Environ. Prot. Agency, Cor-
  vallis, Ore.

Dougherty, J.D., and R.L Morris. 1967. Studies on  the re-
  moval  of actinomycete musty tastes and odors in water
  supplies. J. Am. Water Works Ass. 59:1320-26.
Garrison, P.J. 1980. Mirror and Shadow lakes demonstraticm
  project. Final rep.  Prepared for City of Waupaca Public h-
  land Lakes Protection and Rehabilitation District.
Hutchinson, G.E. 1967. A Treatise on Limnology. Vol. 2. Intro-
  duction to lake biology and the limnoplankton. John Wiley
  and Sons, Inc., New York.

National  Oceanographic  and  Atmospheric Administration.
  1982. Climatological data for Kentucky. Vol. 77. Washing-
  ton, D.C.
Palmer, C.M. 1977. Algae and Water Pollution. EPA-600/9-77-
  036. U.S. Environ.  Prot. Agency, Cincinnati, Ohio.

Pastorok, R.A., T.C.,  Ginn, and M.W. Lorenzen. 1981. Evalua-
  tion  of  aeration/circulation  as a lake restoration tech-
  nique.  EPA-600/3-81-014. U.S Environ.  Prot. Agency, Cor-
  vallis, Ore.
Symons, G.E. 1956a. Taste and odors. Part I. Water Sewage
  Works 1903:307-10.

	1956b. Tastes and odor control. Part 2.  Water
  Sewage Works 1903:348-54.

Standard Methods for the Examination of Water and Waste-
  water. 1980. 15th ed. Am.  Pub. Health Ass. Am.  Water
  Works Ass. Water Pollut. Control Fed., Washington, D.C.

Tuwiner, S.B. 1976. Control of microorganisms  in reservoirs.
  Water Sewage Works 123:69-70.

U.S.  Environmental Protection Agency. 1979.  Methods for
  chemical analysis of  water and wastes. EPA-600-4-79-020.
  Cincinnati, Ohio.

Waltman, R.M. 1982. Evaluation of taste and odor problems
  associated with Sympson Lake, Bardstown (Nelson Coun-
  ty), Ky. M.S. Thesis, Univ. Louisville.

Wedepohl, R. Personal  cornm. 1983. Wis. Dep.  Nat. Resour.
  Madison.
                                                    344

-------
                                          Acidic   Precipitation
CALCITE  DISSOLUTION AND ACIDIFICATION
MITIGATION STRATEGIES
HARALD U.  SVERDRUP
Department of Chemical Engineering
Lund Institute of Technology
Lund,  Sweden
            ABSTRACT

            The dissolution kinetics of a calcite powder sinking in acidified water, and the dissolution kinetics
            for the long-term dissolution of calcite from the bottom, taking the deactivation into account are studied.
            The solutions of the differential equations are given as diagrams that can be used to predict the out-
            come of a treatment, and the calcite utilization. These solutions are shown to be in agreement with
            observation from lake liming projects. As a consequence of the verified theory, a mitigation strategy
            can be outlined in terms of meeting neutralization requirements and project economy. This depends
            on such variables as initial pH, particle size, location of treatment, water depth, and such. The acidifica-
            tion mitigation strategy is defined for (1) lakes and near stagnant waters, and (2) running waters. For
            lakes and near stagnant waters a strategy is outlined, and criteria for the  neutralizing agent given.
            The Kalkad Calcite Distributor developed as a result is presented. For running waters a strategy is
            defined, and criteria given for the neutralizing agent and the location of the effort, to optimize the
            result and its cost A short presentation of equipment for treating running water is given—The Fluidiz-
            ed Divertion Well, The Slurry Dosing Equipments and The Dry Powder Dosing Apparatus—along with
            data on their performance
INTRODUCTION

Experience from the Swedish lake and soil liming pro-
gram  has clearly  shown  that  knowledge of the
mechanisms of calcite dissolution is essential in plan-
ning an acidification mitigation strategy.
  Neutralizing  agents other than calcite have been
tried,  but calcite seems to be the most economical
and easiest available agent. Therefore this study will
only consider calcite and the knowledge necessary to
achieve desired neutralization.

CALCITE DISSOLUTION AND LIMING

The dissolution of solid  calcite in dilute acid can be
described  as a heterogenous solid-liquid  reaction.
The chemical reactions involved are:
(1)  CaCO3 + H+
                            + HCOJ
    (2)  CaC03 + H2O  + C02-Ca2+  + 2HCC>3

    (3)  CaCO3 + H20 ^ Ca2+ + HCOJ + OhT

In natural waters with a  pH-value less than pH 6.5,
reaction (1) will dominate, and for values above pH 6.5
reaction (3) will  be  important.  In  most cases the
natural waters considered for liming will contain too
little dissolved carbon dioxide for reaction (2) to be of
any importance.
  The dissolution rate for calcite under different con-
ditions has been shown to be controlled by diffusion
of the reacting species (Sverdrup, 1982,  1983a; Sver-
drup and Bjerle, 1983).
  It is also necessary to distinguish between two
cases which lead to different kinetic expressions: Dur-
ing the inital stage of a  liming operation, the  calcite
                                               345

-------
 LAKE AND RESERVOIR MANAGEMENT
 powder sinks through the water column. The dissolu-
 tion rate will be rapid compared to calcite resting on
 the bottom. There, the diffusion boundary layer will De
 several orders of magnitude larger and the dissolution
 rate accordingly slower. In addition, the stagnant con-
 ditions will permit precipitates to  form on the calcite
 surfaces and thus further reduce the dissolution rale.
 INITIAL DISSOLUTION IN LAKE LIMINGS


 When a calcite powder is spread in a lake, the par-
 ticles will rapidly sort out according to size and sink
 through the water column and dissolve. This process
 is in detail described earlier (Sverdrup, 1983a). The pro-
 cess is controlled by diffusion of the hydrogen ion arid
 the governing differential equation  may, for the
 dissolution of one particle as in Figure 1, be written:
        aM        aC
 which means that the mass dissolved is proportional
 to the surface of the particle, and the flux of hydrogen
 ions  into this surface. When it is considered that a
 powder consisting of many different particle  sizes is
 sinking in acidic water as illustrated in Figure 2, the
 equation becomes:

    d    n            n  A|
                   n-1
                    I
                                                                               M0-
                                                                                                  (2)
                            PARTICLE
                            RADIUS  R

        BOUNDARY  LAYER
        THICKNESS   AR

 Figure 1.—Dissolution model for a single spheric particle.
Figure 2.—Dissolution model for a calcite powder sinking in
a lake. The largest particles will sink the fastest and partly
dissolve and elevate the pH-value before the next particle
size.
                                                    This equation can be solved analytically and express-
                                                    ed as the dissolved fraction X:

                                                                       n  Ah         n-1
       (1 +
            n-2    n-1
  1    n-1    n-1
.!•••._Z    ._!   j)»»»B
-------
                                                                                     ACIDIC PRECIPITATION
  A close study of the solution equation shows that a
high initial pH-value in the water column can, in order
to maintain a good calcite utilization, be compensated
for by choosing an application site with greater depth.
Waters with high initial pH-value or lesser depth, as
for instance running waters, can be compensated for
by choosing a more finely ground powder.
  A  high dosage intensity  will tend to lower the
dissolved fraction by  local  saturation  or complete
neutralization. This may be countered by spreading
the powder  over  a larger part  of the water surface
(Sverdrup et al. 1983).
  It is also of importance that the powder is well mix-
ed and suspended in water when spread. In  a normal
dry calcite powder the smaller particles will adhere to
the surfaces of the larger ones. If not well suspended
in water  prior to liming, the larger particles will carry
the smaller  and, accordingly, the powder is not effi-
ciently utilized. This can be seen  in Figures 6 and 7,
which show a dry calcite particle and one of the  same
size picked from a suspension in water.  It can clearly
be seen how the surface of the dry particle is covered
with smaller ones.
LONG TERM DISSOLUTION OF CALCITE

Calcite resting on the bottom is under hydrologically
stagnant  conditions  compared  with  their  sinking
phase.  The  diffusional  boundary  layer  is  several
orders of magnitude larger and the dissolution rate ac-
cordingly slower even in river beds and littoral zones
of lakes.
  Such stagnant conditions will also allow  precipi-
tates of aluminum and iron and biological growth on
the calcite  surface  to  form, and thus  impede  the
dissolution further. If the calcite is viewed as a sheet
resting on the bottom, the part that  dissolves down-
ward will be bound into the sediment  and released on-
ly when all overlying calcite is dissolved. This implies
that only a fraction of what is resting  on the bottom is
available to  the  water column  and  also only for a
limited period of time. This can be very clearly seen in
Figure 8 which shows an integrated mass balance for
the lake Ovre Bolsjon which was limed in  1976 with
500 tons of calcite ground to 0-0.5 mm (no. 6). The first
stage, A, represents the initial dissolution during sink-
ing, that is 38 percent of  the total amount. The long
term dissolution, B, constitutes approximately 5 per-
cent of the total amount or 8.5 percent of the  amount
left on the bottom.
  It can here be seen, as in many other cases,  that the
calcite deposited on the bottom is poorly utilized and
does  not give a pronounced long-term effect in this
case. Approximately 50 percent is "lost" on the bot-
tom.
  The kinetics for the long-term dissolution can be
described with an expression of the following type:
 am

' at '

 aF
 aF

(at)

 aSY
                  aSx
                 (-
                  at
                      aC
(4)
                                               (5)
where vXt) is a deactivation term taking care of the in-
hibition of the calcite surfaces and the sedimentation
of detritus, and  (aSx/at) a term  for the adsorption-
desorption rate of calcium to the sediments. (aF/at) is
the net flux of calcium from the bottom covered with
      calcite. Based on the dissolution kinetics and on the
      mass balance, a reacidification model was made for
      limed lakes. A test run for the lake Ovre Bolsjon is
      shown in Figure 9 with predictions and observations
      on the pH, the alkalinity, and the calcium.
          100
                                                     in
                                                     O
                                                     (0
                                                     oc
                                                     HI
                                                     <
                                                     a
                                                     HI

                                                     o
                                                     CO
                                                     to
                                                     Q
                 4.0   4.5   5.0   5.5   6.0
                     INITIAL  pH-VALUE IN THE LAKE

      Figure 3.—The dissolved fraction X for different calcite
      powders as  shown in Figure 5 versus initial lake pH. The
      diagram is valid for a sinking depth of 5 meters, but this may
      be  compensated  for  other depths with  the  equation:
      PHdiagram  =  PH,m,ia| - Iog10 (S/5.0 m).
                              +        T-CHALK
                              O 0-1 OMM DOLOMITE
                              D 0-O.2MM CALCITE
                              • 0-0.5 MM CALCITE
                              A 0-1 OMM CALCITE

                              T 0-20MM CALCITE
                                                           4.0   4.5   5.0   5.5   6.0    6.5
               INITIAL pH-VALUE IN THE LAKE


      Figure 4.— The  dissolved fraction X observed  in several
      Swedish lake liming projects as compared to the theoretical
      calculations.
                                                  347

-------
 LAKE AND RESERVOIR MANAGEMENT
   Figure 8 shows the dissolution of calcite together
 with the integrated mass balance. In this case, as v/ell
 as  in  several  others  tested  at  the Institute  of
 Technology in Lund, the model seems to predict the
 reacidification of the lake very well.  The model is
 available at the Institute of Technology in  Lund and
 runs on a small desk computer that makes calcula-
 tions for whole lake systems.
   In Sweden  it has  been  considered that  the water
 movements over hard lake bottoms or in the littoral
 zone could enhance the dissolution. In terms of diffu-
 sion resistance, however, this has negligible effects.
 In these areas lack of sediment makes the downward
    100
  £   90
  £   80
  o   70
  5   60
  S   50
  "-   40
  5 2  30

  u tij  20
  £5  10
  Q. iyi  Q
PARTICLE DIAMETER IN MICRONS
02   0.5  1  24   10   30 60100  300   1000
     Pj~l	1—j--+-rL^—i  i -M-r i^-gt^t-^i^ i 1100
       0.0002   0001   0004 001  003   01   03   10
         PARTICLE DIAMETER IN MILLIMETERS
Figure 5.—Particle size distribution curves for several calcite
powders used in Swedish liming projects. Some of their com-
mercial names are: 1: Agricultural limestone, 6:0-0.5 mm
calcite powder, 7:0-0.2 mm calcite powder, 9:Malmo Chalk-
powder.
Figure 6.—A sample of dry screened calcite powder. It can
be seen that the smaller particles adhere to the surface of
the larger ones. If the powder is not well suspended in water
prior to liming, the smaller particles will follow the larger
ones as they sink.
 diffusion into the sediments as well as the adsorp-
 tion/desorption of calcium very small; hence, a larger
 fraction of calcite is available to the lake water.
 SPECIFICATION OF CALCITE POWDERS
 FOR LIMING

 Technically all powders can be well defined by the par-
 ticle size distribution, and accordingly, the specifica-
 tions should be tied to the particle size distribution
 curves. Commercial names may be misleading. It is of
 economic importance that the calcite is well utilized
 in  relation  to  its  application price. An  economic
 analysis  of the  experiences from  Swedish liming
 operations show that calcite powders coarser than no.
 5 in Figure 5 will give low cost efficiency. In general,
 the calcite powders marked nos. 6, 7,8, or 9 will be the
 most cost efficient in lake limings, and nos. 7, 8, 9, or
 10  best for running water. The exact optimum can'be
 calculated as the lowest neutralization cost which is
 the application cost divided by the  fraction calcite
 utilized.
 LIMING ON LAND

 Liming on land has been tried as a method to treat sur-
 face waters. These efforts have not always been suc-
 cessful and the dissolution efficiency of  calcite in
 relation to the target water was rather  low.
   The dissolution rate differential for calcite in soil is
 of a type similar to equation 4, but the dynamics of the
 system are quite different as  soil limings are  most
 often  performed  in  the  unsaturated  zone  with a
 periodical infiltration flow of an intermittent nature. In
 lake bottoms the conditions  for mass transfer are
 those of a stagnant saturated system.
   The dynamics of the dissolution and the leaching of
 the soil do not follow the dynamics of the watercourse
 flow, hence massive efforts are needed to produce any
 effect.
   Acid soils are greatly depleted of calcium ions, and
 before any substantial amount can be leached to the
 runoff, the calcium will be adsorbed by the soil to fill
 this deficiency—in fact neutralizing the soil. This ex-
 plains why several land limings, intended to affect sur-
 face water,  do not work.
'II U
/loo-
10 -
z eo •
P 7o •
ID
d <°°~
CO i»-
n
m *"
P J°"
D 2°J
S W-

o OH














-


y — \— -^-^^~^~\_
^/^\^-~-


OVRE BOLSJON


,' 	 INTEGRATED MASS BALANCE
' 	 HS-1 MODEL




5
CO
<
o
co
O
l-









                                                               1976
                  1977    1378
                                                                              1979    1980
Figure 7.—A sample of calcite particles from  a powder
suspension enlarged 200 times. It can be seen that no more
small particles adhere to the larger.
Figure 8.—An integrated mass balance for Lake Ovre Bols-
jon on calcium. The lake was limed in 1976 with 500 tons of
powder no. 6 of calcite. Simulation of the integrated mass
balance for calcium is marked with the dotted line.
                                                  348

-------
                                                                                   ACIDIC PRECIPITATION
  American studies (Meyer and Volk, 1952) show that
particles larger than 0.3 mm in diameter are of very lit-
tle  value  in  soil limings.  Accordingly, the coarse
powders, often used in soil limings, demand heavy
doses.
A MITIGATION STRATEGY OUTLINED

It is of importance when a liming project is planned
that the decisions taken are adequate and proceed in
the right order, as they depend upon each other. The
flow sheet in Figure 10 tries to identify the most impor-
tant steps and their best order.
  The target water for the liming project should first
be  characterized in terms of neutralization need as
dissolved calcite. This may either be done by titration
or  calculated  theoretically, taking precipitation  of
aluminium and iron  complexes and such factors into
account. Wright (1983) gives an example of the pro-
cedure.
  The next step will  be to calculate the total dissolved
calcite needed in the project.  It will depend on  the
volume of the  lakes to be limed, the flow rate of the
river to be limed, or the amount of soil to be limed and
neutralized.
  The actual amount to be used for the neutralization
is  the  amount calculated  from the physical and
chemical conditions of the location. For a lake this
can be expressed as:
M =
     dM
VL» (	) • ApH
    dpH         1
    	 •
      X          Y
                                          (6)



_I
u
2
S
0
CO
_J
<
0
_l
>
^
2
>
)—
z
_l
<
^
_J
<
7

o
5

4- •
3 •

2

J2O
115 -
no -
)05 -

) 00







m










r~— -•**-•
' — ~^*-^ *
— -^^^^
	 5-^^
g

OVRE BOLSJON


CALSIUM

i 	 66—-,
°° b~~--o^__^
ALKALINITY °~S~C


] | | 	 1 	 1 	
70 -

65 -

&o -

55 -
                    PH
        1975   1976   1977  1978   1979   1980




         • 0+  OBSERVATIONS

         ~;  HS-1  MODEL

Figure 9.—Predictions made with the reacidification model
as compared to the observations.
                                               The lake volume  VL times the neutralization  need
                                               (dM/dpH) expressed  as  gram calcium carbonate per
                                               pH-unit and m3, times the needed pH-value elevation
                                               will give the theoretical  amount needed as dissolved
                                               calcium  carbonate.  When  this is  divided by the
                                               dissolution efficiency X and the content of calcium
                                               carbonate, Y, in the calcite used, the actual amount to
                                               be added is calculated. For a river the calculation
                                               becomes:
                                                              dM
                                                         QI"(    )
                                                     1        dpH
                                                M = — E(	
                                                     Y  i        X;
                                                                                              (7)
where Q| is the volume of runoff for a period of time in
which the pH-value elevation needed is ApH,, and the
dissolution efficiency, X|, under the circumstances. As
the conditions in a river may vary considerably, the
calculation will have to be made  for shorter periods
and added together as in equation 7.
  Several possibilities for methods and localization of
the neutralization may be suggested. For these alter-
natives, the result must be estimated  in terms of
chemical result and duration of a satisfactory condi-
tion in the system. For lakes, a diagram as in Figure 11
may be used to estimate the time needed for the lake
to be reacidified  to pH 6.
  The expected result in terms of chemistry, biology,
result stability, and duration are then compared to the
intentions  of the  liming  project. All methods and
possible localizations of the neutralizing  operation
which cannot fulfill the intentions of the project must
be refused or redesigned.
  The choice will often  be between methods for run-
ing water and methods for lake liming. Lakes or lake
systems with very short retention times may often be
successfully treated as  running water. In Figure 12 a
decision tree is  suggested to assist  in choosing  a
method. Usually  several alternatives to carry out the
neutralization and to reach a satisfactory result exist.
  At this stage the total project cost or the neutraliza-
tion cost efficiency is calculated for the different alter-
natives and the most  cost-efficient  alternative can
thus be used.
                                                LIMING OF RUNNING WATER:
                                                SOME EXAMPLES

                                                In Sweden today, there are  many technique options
                                                and methods for liming running water.
                                                  The diversion well. The  fluidized diversion well has
                                                been tried in laboratory experiments and since 1980 at
                                                a pilot plant at Piggaboda in Smaland,  Sweden.
                                                  The working principle of the well is  that of a ball-
                                                mill. Water flowing through the calcite gravel will keep
                                                the particles in rapid motion, and the mechanical wear
                                                on the particles will grind them to powder. The powder
                                                then dissolves rapidly. The mechanical  wear also
                                                keeps the calcite  surfaces free from  precipitates
                                                which, if the particles were  at rest would reduce the
                                                dissolution  drastically. The apparatus takes all  its
                                                energy from the water. The diversion well is shown in
                                                Figure 13. The widening at the top will lower the fluid
                                                velocity  below the  fluidization velocity and the par-
                                                ticles will hence be  returned  to the well. In this way an
                                                 349

-------
 LAKE AND RESERVOIR MANAGEMENT
                                      NEUTRALIZATION
  THE SPECIFIC NEUTRALIZATION DEMAND
  CAN BE DETERMINED AS AN EXPERIMENTAL
  TITRATION CURVE
       THE SPECIFIC NEUTRALIZATION DEMAND
       CAN BE CALCULATED THEORETICALLY;
       TAKING PRECIPITATION OF ALUMINNUM,
       RESIDUAL ALKALINITY AND SUCH FACTORS
       INTO ACCOUNT.
                             SPECIFIC NEUTRALIZATION DEMAND

                TO DETERMINE THE NET SYSTEM NEUTRALIZATION AMOUNT, THE WATER
                VOLUMES AND THE RETENTION TIMES INVOLVED HAVE TO BE CONSIDERED
                                            I
                             SYSTEM NEUTRALIZATION DEMAND

                DEPENDENT ON WHERE AND HOW THE EFFECTS ARE WANTED, AND TYPE OF
                SYSTEM, WILL THE EFFORT BE LOCALIZED IN THE SYSTEM
                                            I
                                   EFFORT LOCALIZATION
        CALCULATION OF DISSOLUTION EFFICIENCY
        AND DOSE OF NEUTRALIZING AGENT.
        ESTIMATION FROM DISSOLUTION MODELS OR
        DIAGRAMMS OR FROM EXPERIENCES

        CONSIDER IF CHEMICAL WATER QUALITY
        DEMANDS ARE MET, IF NOT THE EFFORT
        WILL BE RELOCALIZED
CALCULATE THE EXPECTED RESULT IN TERMS
OF RESULT DURATION AND STABILITY.
ESTIMATION WITH REACIDIFICATION MODELS
DIAGRAMMS OR EXPERIENCES.

CONSIDER IF BIOLOGICAL DEMANDS ARE
MET, IF NOT THE EFFORT WILL BE RE-
LOCALIZED.
                            CONSIDER IF DEMANDS ON RESULT DURATION
                            AND STABILITY ARE MET, IF NOT THE EFFORT
                            WILL BE RELOCALIZED
                                            I
                             POSSIBLE EFFORT ALTERNATIVES
                  TECHNICAL DATA ON EFFICIENCY, NEUTRALIZATION AGENT COST PRICES OF
                  EQUIPMENT, TRANSPORTS AND DELIVERED SERVICES ARE USED TO CALCULATE THE
                  COST EFFICIENCY OF NEUTRALIZATION, TO OPTIMIZE THE EFFORT COMBINED WITH
                  CONSIDERATIONS ON THE RELIABILITY OF THE INVOLVED METHODS AND EFFORT
                                            I
                         COST EFFICIENT NEUTRALIZATION EFFORT

                WHICH SHOULD WHEN CARRIED OUT, GIVE THE WANTED RESULT IN TERMS
                OF CHEMISTRY, BIOLOGY AND DURATION IN THE MOST COST EFFICIENT
                WAY
                                            I
                                          RESULT
Figure 10.—Flow sheet for the planning of liming efforts.
                                           350

-------
                                                                                    ACIDIC PRECIPITATION
80-90 percent utilization of the calcite mass can be
achieved.
  The dosage given by the well is more or less cons-
tant  per m3. Experiences from Piggaboda show that
approximately 5 percent of the mass in the diversion
well  dissolved per day, raising the pH-value by  one
unit.
  The diversion well will function well, without freez-
ing, down to at least -15° Celsius (see Fig. 14).
  More information on the fluidized diversion well  can
be found in Sverdrup et al. 1983 or Sverdrup et al. 1981.
  The dry-powder doser. The dry-powder doser  has
been successfully tried in full scale in the salmon river
Fyllean, Halland, Sweden. A plan of the plant is shown
in Figure 15. Calcite  powder, usually of type nos. 6, 7
or 8,  is stored in a large container. The automatic regu-
lating unit will register the level of the river and relate
it to the flow rate. The dry powder is mixed with water
and then pumped to the river. This  will ensure good
suspension of the dry powder and accordingly a fair
calcite utilization.
  The pH-value in the river ranges from pH 4.9 to 5.9
upstream from the  plant.  The dosing  takes  place
where the river is  approximately 80 cm deep.  During a
test period, from January to May 1983, approximately
45 percent of the  calcite was utilized (Fig. 16).
  The slurry doser. The slurry doser has been located
in the Fyllean river further downstream from the dry
powder doser at Marback. The plant also contains a
large tank for the wet suspension of very finely ground
calcite,  and a regulation unit which calculates the
river flow rate from the water level. The calculations
are here performed by a microprocessor containing
the nonlinear pH flow rate relationship and a titration
curve (Fig. 17).
  LESS THAN 0.3 YEARS
0.3-
1.5 YEARS
                                     I
  The calcite suspension used contains 70 percent
calcite suspended in water. The powder is extremely
finely ground to a mean diameter of 0.5-2.0 microns.
Such  a  fine calcite powder will dissolve completely
even  as pH-values near 7.  At such  pH-values the
          003   Q05   01  015 02 03 Ot 05 07 t   15 2

              MEAN RETENTION TIME OF THE LAKE IN YEARS
                                         5456 78910
Figure 11.—The liming result duration until pH 6 plotted ver-
sus the lake mean retention time. The drawn line is the
theoretical calculation made with the "HS-1" model and the
circled observations from: 1.  Hornasjon, Goteborg; 2. Ovre
Bolsjon, Bohuslan; 3. Lysevatten, Goteborg; 4. Smedvatten,
Goteborg; 5. Ski tjarn, Varmland; 6. Blomman, Goteborg; 7.
Sodra Blotevatten,  Bohuslan; 8. Bredvatten, Goteborg;  9.
Stensjon,  Varmland; 10. Grytsjon, Orebro; 11. Ekelidvattnet,
Bohuslan; 12.  Nedre Sarnamannasjon, Jamtland;  13. Tar-
malangen; 14.  Kolabodasjon,  Smaland. For mean retention
times shorter than 0.5  years the resulting duration  is
unstable and sensitive to sudden  pH depressions in the
tributaries.
   1.5-4 YEARS
MORE THAN 4 YEARS
                        FULLFILLMENT OF CHEMICAL
                        AND BIOLOGICAL REQUIREMENTS

                        RESULT DURATION AND
                        STABILITY

                        NEUTRALIZATION COST
                        EFFICIENCY
                                               METHODS FOR LAKE LIMING
                                              RESULT DURATION AND STABILITY

                                              NEUTRALIZATION COST EFFICIENCY
  METHODS FOR LIMING OF RUNNING WATERS
                 METHODS FOR LAKE LIMING
Figure 12 —Flow sheet for deciding between methods for running waters and methods for lake liming.


                                                 351

-------
 LAKE AND RESERVOIR MANAGEMENT
 calcite  utilization  for coarser  powders  falls  off
 drastically.
   Other methods. There are many other, and in terns
 of regulation, simpler solutions available in Sweden,
 such as the Hallefors  apparatus or the Borlange ap-
 paratus which are smaller and simpler versions of dry-
 powder, dry application techniques. They usually give
 a rather low calcite utilization efficiency due to the dry
 application.
   Limestone barriers have been tried on several occa-
 sions in Sweden. They very soon clog and fail to work.
   Long beds of several kilometers of limestone gra/el
 do not work particularly well either in rivers.
   Brook liming with calcite powder usually works for
 only a limited period and not during spring flood con-
 ditions. Projects  as Vannean, Anrasan or Hogvadsan
 in Sweden show  this very clearly.


 THE PERFORMANCE OF THE DOSING
 EQUIPMENT

 All the mentioned equipment can  achieve a correct
 neutralization of  running water as long as it  is well
 designed and adjusted. Most of the regulated equip-
 ment works on the flow rate as it is our experience
 that pH electrodes are very unreliable without tedious
 maintenance during the  winter season in rivers or
 natural waters.
   The calcite utilization differs for the different equip-
 ment, and will affect the neutralization cost efficiency.
Figure 13.—Schematic view of the diversion well.
 The  calcite utilization  efficiency  is summarized in
 Figure 19 together with the calcite utilization for other
 methods.

 LAKE LIMING DEVELOPMENT OF NEW
 TECHNOLOGY

 Conventional lake liming is carried out by boat, and
 preferably the calcite powder should be well mixed
 with  water prior to spreading. This usually ensures
 precision and good results. The capacity is approx-
 imately 30 to 40 tons of calcite per day.
   To improve daily capacity the "Calcade" method
 was developed.  Conventionally a boat must return to
 base to reload  every time its 7 ton supply is ex-
 hausted.  Instead of using a tank on board, in  the
 Calcade method the boat is outfitted with a long rub-
 ber hose and the calcite powder continuously pumped
 as a slurry to the boat. This raised possible delivery to
 150-200 tons per day. When the slurry is pumped at
                    REGULATION

                        UNIT


   CALCTE SLURRY SILO
                                                                  DOSING  UNIT
                                                    Figure 15.—Schematic view of the dry powder doser at
                                                    Ryaberg, Fyllean.
Figure 14.—Photograph of the diversion well at Piggaboda.
Figure 16.—View of the site with the dry powder doser at
Ryaberg.
                                                352

-------
                                                                                ACIDIC PRECIPITATION
high pressure, the hose can be several km long and all
heavy equipment stay ashore by the roadside. The
flexibility of the long hose means that the equipment
does not even  have to be  placed on the shore.  A
distance from the shore of 3 kilometers (or 2 miles)
has been tested successfully so far.
  A  picture of the equipment in  use can be seen  in
Figure  20 where the boat is 1.8  kilometers from the
land station. In this project, 9,000 tons of calcite were
spread in Stora Lee, a large lake in Varmland, Sweden.


DISCUSSION

Despite the fact that the ultimate goal of a liming pro-
ject is to preserve or restore ecological or biological
communities, the problems encountered  in carrying
out the project are often of a strictly technical nature.
  In  complex hydrology or chemical reaction engi-
neering, problems are encountered in almost any lim-
ing project.  When  dealing  with dosage apparatus,
                        regulating problems often appear. Such problems are
                        often more efficiently dealt with by engineers than
                        biologists. It is important that the person planning a
                        liming  project has some  understanding  of  the
                        chemical  reaction  engineering  principles  and the
                        hydrological principles involved in the dissolution of
                        calcite in natural water systems.
                       THE FINAL SOLUTION TO THE
                       ACIDIFICATION PROBLEM


                       What is discussed in this paper does not represent a
                       final solution to the acidification problem. It only gives
                       a perspective of how some of the damage of the en-
                        TOTAL CALCITE   UTILIZATION
            REGULATION

                 UNIT
 CALCITE
 SLURRY  SILO
WATER1
    DOSING  UNIT
                                                  07.     207,
                                          CflLMTE Si-URR.V OOSERS
                        DRY POWDER
                          POScKS  <
                                                  wrr SUSPENSION
                                                                                      LAKE
                                                             • IAN6
                                                             EAM.IERS, 4
-------
 LAKE AND RESERVOIR MANAGEMENT
 vironment  can be temporarily and partially repaired.
 Very severe problems face us in the near future as the
 acidification of the  soil  and the ground  water pro-
 ceeds, threatening our water resources and forests.
 The final solution to  the  problem must be to reduce
 the emissions of acidifying  agents.
Figure  21.—Conventional lake  liming: The Kornsjo-Boksjo
lake  system was limed with 9,000 tons of calcite powder
classified as 0-0.2 mm. The calcite was spread from a boat
with a capacity of 14 tons. Then the powder was applied dry.
Later it has been discovered that the wet technique will bet-
ter use the calcite. The "dust bowl" is also avoided.
                                                          Figure 22.—The calcite powder should be evenly spread over
                                                          the lake surface. A very concentrated dose may locally in-
                                                          crease the density of the water, and plumes of water-calcite
                                                          slurry will sink through the water column. The good spread is
                                                          ensured with the Calcade method, here seen 1.8 kilometers
                                                          out from the supply station.

                                                          ACKNOWLEDGEMENTS: The North American Lake Manage-
                                                          ment Society with  their cosponsors, the Tennessee Valley
                                                          Authority, the U.S.  Environmental Protection  Agency,  the
                                                          U.S.  Department of Agriculture, and the Electric Power
                                                          Research Institute  made my participation at  the Knoxville
                                                          Symposium possible by financing my transatlantic travel.


                                                          REFERENCES

                                                          Meyer, T.H.,  and C5.W. Volk. 1952. Effect  of particle size of
                                                            limestones on soil reaction, exchangeable  cations and
                                                            plant growth. Soil Sci.  1952:31-52.
                                                          Sverdrup, H. 1982.  Dissolution of calcite and  other related
                                                            minerals in acidic aqueous solution in a pH-stat. Vatten
                                                            38:59-73.
                                                             	1983a. Lake liming Chem. Scripta 22:1:12-18.
                                                                  1983b. "HS-1" A simple reacidification  model for
                                                            limed lakes. Pap. pres. George D. Aiken lectures: Acid Rain
                                                            Transportation  and Transformation  Phenomena.  Burl-
                                                            ington, Vt. Sept. 18-20.

                                                          Sverdrup, H., and I. Bjerle. 1983.  The calcite utilization effi-
                                                            ciency  and the  long-term effect on alkalinity  in several
                                                            Swedish lake liming projects. Vatten 39:41-54.
                                                          Sverdrup, H., H.R. Eklund, and I. Bjerle. 1981.  Kalkning av
                                                            rinnande  vatten,  erfarenheter   fran  en  fluidiserad
                                                            Kalkbrunn-Mover.  Vatten  4:388-393. (Liming of running
                                                            waters,  experiences  from  a  fluidized  limestone bed-
                                                            Mover® )

                                                          Sverdrup, H., et al. 1983. Liming of  running waters.  Proc.
                                                            Symp. in Alvesta, Sweden 18/5 -83. To be translated by Bat-
                                                            telle, Columbus, Ohio.

                                                          Wright, R.F. 1983. Liming of Hovvatn.  Rep. No. 3, Norwegian
                                                            Liming  Project. Oslo.
List of Symbols

A

A'

AL

AP

B

C

aC
 aR
Constants

Constants

Area of the lake

Area of a calcite particle
Surface load factor for lake liming

Hydrogen ion concentration


Concentration gradient around particle
                                                     354

-------
                                                                                               ACIDIC PRECIPITATION
aC
aX
D
M
M0
M,
ME
dM
dpH
aM
 at
 aF
Concentration gradient on the boundary
layer

Diffusion coefficient
Mass
Initial mass
Mass of fraction i
Conversion factor

Neutralization need
               Dissolution rate
               Flux of calcium from the sediments
  n   A,
 at
v(t)
X
Y
M*
Fraction of particles with radium R,
Flow rate

Particle size distribution function

Fraction of the lake area covered in the
liming operation
Sinking depth

Sorption rate of calcium to the sediments

Time dependent inhibition function
Dissolved  fraction
Percent carbonate content in calcite
Surface load as amount calcite per area
treated
                                                       355

-------
ONTARIO'S EXPERIMENTAL LAKIE NEUTRALIZATION  PROJECT:
CALCITE ADDITIONS AND SHORT-TERM CHANGES
IN LAKE CHEMISTRY
L A. MOLOT
J. G. HAMILTON
G.  M. BOOTH
Booth Aquatic  Research Group Inc.
Toronto, Ontario, Canada
            ABSTRACT

            This paper presents a preliminary analysis of chemical changes associated with the addition of
            84 tonnes of fine calcite (mean diameter 9 ^m) in dry form during Aug. 11-16, 1983, to Bowland
            Lake, a remote and highly acidic lake near SJudbury, Ontario. A Canso water bomber was used to
            apply the calcite to the lake surface. Immediately following liming, the volume-weighted pH had
            increased from 5.0 to 6.8 and the volume-weighted alkalinity had increased from -0.7 to 3.6
            mg/l. The calcite dissolution efficiency was 40 percent by Aug. 18 and 52 percent by Aug. 31.
            Aerial application with a Canso is a logistically simple way of neutralizing a remote lake when a
            suitable runway is available nearby. However, the calcite dissolution efficiency may be lower
            than the expected value of 73 percent because bombing from low altitudes produces locally high
            concentrations of calcite particles, which may inhibit dissolution.
INTRODUCTION

The Experimental Lake  Neutralization  Project  is a
joint investigation by the Ontario Ministries of Natural
Resources and the Environment designed to test the
feasibility of neutralizing acidic lakes (liming) as an in-
terim mitigative strategy. The objectives are to protect
a lake endangered by acidification and to rehabilitale
an acidified lake. The project is coordinated by Boolh
Aquatic Research Group Inc., of Toronto.
  This paper presents  a preliminary analysis of
chemical changes associated with the addition of 614
metric tonnes (92 Imp. tons) of finely ground calcile
during Aug. 11-16,  1983, to Bowland Lake, a highly
acidic lake. Biological and chemical monitoring of
Bowland Lake began in May 1982 and is scheduled 1o
continue until the fall of 1985. Monitoring components
include  monthly surveys  of routine  limnological
parameters (such as metals, nutrients, plankton, ard
major ions), a spring runoff sampling program, benthic
surveys, fish stock assessment, and bioassays of ear-
ly life stages of lake trout (Salvelinus namaycush).
SITE DESCRIPTION AND PRETREATMENT

CONDITIONS

Bowland Lake is a remote lake without road access 70
km north of Sudbury, Ontario (47°05'N SO'SOW). The
average volume-weighted  pH  was  5.0, the Grun
alkalinity ranged from -1.00 to - 0.20 mg/l as CaCO3,
the mean depth is 7.6 m, the maximum depth is 28 rn,
and the surface area is 109 ha. A bathymetric map is
presented in Figure 1. Aluminum and nickel concen-
trations ranged from 120-160 and 3-5 /ig/l, respective-
ly, while copper concentrations were less than 11 /^g/l.
Bowland Lake is a headwater lake in a granitic basin
forested mainly by black spruce (Picea mariana) arid
Jack pine (Pinus  banksiana). A reproducing popula-
tion of stunted yellow perch (Perca flavescens) is the
only fish population present.  A  native lake trojt
Figure 1.—Bathymetric map of Bowland Lake with station
locations (*).
                                              356

-------
                                                                                  ACIDIC PRECIPITATION
population  has apparently been extinct for over 10
years and subsequent stockings of hatchery lake trout
have failed on  several occasions  (J. Gunn,  pers.
comm.).
RATIONALE FOR CHOICE AND AMOUNT
OF AGENT AND APPLICATION METHOD

Finely  ground calcite (CaCO3) was chosen as the
neutralizing  agent  because it is  readily available in
Ontario, it is not dangerous to handle, and it dissolves
faster  than  coarse grades of  calcite.  Furthermore,
calcite does not result in a temporarily high pH follow-
ing dissolution, except in the immediate vicinity of a
concentrated  dosage.  In  previous neutralization
studies in the Sudbury area, increases  in whole-lake
pH beyond  the final equilibrium  value may  have
resulted  in decreases in plankton  species  diversity
and  abundance in  lakes  neutralized  with  either
Ca(OH)2 or a mixture of Ca(OH)2 and calcite (Dillon et
al. 1979; Scheider et al. 1975). This phenomenon, refer-
red to as 'pH shock,' is undesirable when attempting
to reintroduce  a  fishery  soon after  neutralization
because of potential damage to the forage base. Addi-
tion of calcite to experimental enclosures in  Bowland
Lake during the summer  of 1982 did not increase mor-
tality in the yellow perch and planktonic crustacean
and rotifer communities  (Molot et al. 1983).
   Unfortunately, the slow  dissolution of calcite
results  in  some settling  of undissolved  particles.
Evidence suggests  that once settled, particles are lost
to the system because further dissolution is prevented
by a coating of metal carbonates and organic material
(Sverdrup and   Bjerle,   1983).  Hence,   it  becomes
necessary to incorporate a correction factor for set-
tling losses. This is summarized as:
                    A  = D/X

where A is the amount to be added, D is the amount re-
quired  in solution,  and  X is  the fraction of A  which
dissolves.
  The  amount of calcite required  in solution,  D, in
Bowland Lake  was  determined  by calculating the
amount of calcite required to (1) decrease the proton
concentration to a  target level, (2) raise  the alkalinity
to a value corresponding  to the target pH, and (3)
reduce total aluminum levels by 80 percent. Aliquots
of lake water were titrated with a saturated solution of
calcite to determine the alkalinity at several values of
pH. The target pH of 6.8 had a corresponding alkalinity
of 4.4 mg/l. Assuming an 80 percent reduction in total
aluminum from 140 ^g/l, the total amount of calcite in
solution  theoretically required  to raise  the pH and
alkalinity and reduce the  aluminum concentration was
5.8 mg/l.
  A model developed by Sverdrup (1983) can be used
to calculate the  correction  factor,  X,  for  non-
dissolution of calcite. The model predicts that,  in an
acidic  lake such as  Bowland Lake, calcite  particles
less than 10^m in diameter dissolve completely, those
larger than 70 /^m undergo little dissolution, and par-
ticles between  10  and  70 ^m undergo  intermediate
dissolution.  Hence, fine grades of calcite  will  yield
greater dissolved fractions than coarser grades. The
model  assumes that dissolution is greatest   when
calcite is evenly dispersed over the lake surface to
avoid locally high pH values which inhibit dissolution.
However, in practice, shallows  should be avoided to
prevent large settling losses.
  The grade of calcite chosen was Snowhite 20-2 from
Steep Rock Calcite of Perth,  Ontario. This grade con-
tains few impurities and the median and mean particle
diameters are 7 and 9 urn, respectively, with 73 percent
of the particles less than 10 /^m in diameter. Use of a
finer grade of calcite,  which is available from Steep
Rock Calcite, would require a smaller dose to raise the
pH. However,  it was deemed unsuitable for dry air-
borne application because the  potential  for wind-
blown loss was high.
  Application of Sverdrup's model (1983) predicted a
27 percent loss of Snowhite 20-2  to the sediments in
Bowland Lake because of settling (which is equivalent
to 73 percent dissolution). Assuming this loss,  an ap-
plication of 7.9 mg/l was required. A further 2.3 mg/l
was arbitrarily added to take into  account experimen-
tal uncertainties such as the base neutralizing capaci-
ty of sediments and possible inaccuracies in the ap-
plication of  Sverdrup's  model. Hence,  the  total
amount  applied was 10.2 mg/l or  84 tonnes (0.77 ton-
nes/ha).
  A Canso water bomber  was chosen to  apply the
calcite  in dry form.  Competitive tendering of the
calcite delivery contract showed the Canso to be a
cost-effective option for application when  road access
is not available and a suitable airport is available near-
by (the  distance between  Sudbury  airport  and
Bowland Lake is 52 km). Calcite was blown from a
pneumatic tanker truck into the Canso. The Canso has
two holds which can be discharged separately.
METHODS

Water samples were collected once daily at three sta-
tions (A-C, Fig. 1) at 2 m intervals from surface to bot-
tom during Aug. 8-18. Sampling occurred only once at
Station D at 2 m intervals on Aug.  18. Samples were
analyzed  immediately for  conductivity  and
temperature and within 6 hours for pH, Gran alkalinity,
and calcium in unfiltered water (Aug. 8-10) or filtered
water (Aug. 11-18). A pore size of  1.2 ^m was used
from Aug. 11-16 and 0.45 ^m from Aug. 17-18. Alkalini-
ty and pH were measured using a Radiometer pH
meter and combination  pH  electrode.  Conductivity
was measured with  a Lisle-Metrix  C-45  conductivity
meter. Calcium concentrations were measured using
an  Orion 701A ion meter and ion-specific electrode.
Temperature was  measured  with  a YSI  Model 43
telethermometer. Samples of unfiltered water were
stored  for later laboratory analysis of  dissolved in-
organic carbon and total aluminum  (Ministry Environ.,
in prep.)
  Aliquots of unfiltered water from Aug. 11-14 were
stored for 4 days before measuring pH, alkalinity, and
calcium.  Because of logistical constraints, aliquots of
unfiltered water from Aug. 15-18 were stored for 10 to
16 days before analysis. In the latter case (unfiltered
aliquots  from Aug.  15-18),  calcium  concentrations
were measured using atomic absorption.
RESULTS AND DISCUSSION

Logistics: The round trip flight time for the Canso bet-
ween Bowland Lake and Sudbury Airport averaged 40
minutes. The loading operation required an average of
25 minutes to load  approximately 2.4 tonnes  (or 1.2
tonnes per hold) which was only 65 percent of max-
imum capacity. The reduced payload resulted from the
low density of air-blown calcite. The aerial application
occurred during Aug. 11-16, 1983. Minor delays were
encountered  because  of problems with  weather,
mechanical failure,  and calcite delivery. Calcite was
                                                 357

-------
 LAKE AND RESERVOIR MANAGEMENT
 dropped from heights of 15 to 25 meters at speeds of
 145 km/hr. The two holding tanks were discharged
 simultaneously during  17 drops and  then separately
 seconds  apart during the  remaining 22  drops to
 achieve greater dispersal on the lake surface. Thirty-
 six of 39 drops were discharged over waters deeper
 than 6 m, with deeper waters receiving proportionately
 more drops than  shallower waters. The drops  were
 confined  to the middle regions of the lake to avoid
 shoreline hazards.
   The total cost of calcite purchase, delivery by truck,
 and subsequent aerial application was $30,000 Cana-
 dian  or $275  per hectare. Fraser and Britt (1983)
 estimated costs of $265-$500 (1981 U.S.; approximate-
 ly $380-$720 1983 Canadian) per hectare for applica-
 tion of  lime slurry by various fixed-wing aircraft.
   Chemistry: Thermal stratification was stable during
 the sampling period with the metalimnion consistert-
 ly occurring between 8 and 12 meters. The epilimnetic
 and hypolimnetic  temperatures were  21  and  7°C,
 respectively. The percentage of the total lake volume
 was 71,12, and 17 percent for the epilimnion, metalirn-
 nion, and hypolimnion,  respectively.
   Preliminary data for Station A are shown in Figuress
 2 to 4. The volume-weighted pH increased from 5.0 on
 Aug. 10 to 6.8 on Aug. 18,2 days after the liming opera-
 tion ended. The epilimnetic pH was  7.1 on Aug. 13,
 decreasing to 6.0  at 20 m (Fig. 2). The relatively  high
 epilimnetic pH noted during the liming operation was
 probably caused by rapid dissolution  of smaller (0-10
 ^m) calcite particles which would, in  turn, inhibit the
 dissolution of larger particles which dissolve slowl/.
 The smaller increase in hypolimnetic  pH suggesls
 that only a small fraction of particles  less than 10 ^m
 in diameter settled below the thermocline and dissolv-
 ed, although larger particles settling  below  the  ther-
 mocline may have dissolved slightly.
   The decrease in  epilimnetic  pH and increase in
 hypolimnetic pH between Aug.  16 and 18 suggesls
 that some mixing occurred between the epilimnion
 and the hypolimnion although not enough to affect the
 temperature profile. Settling of particles into more
 acidic hypolimnetic waters after liming ceased on
 Aug. 16 and subsequent dissolution  may have  con-
 tributed to the increased hypolimnetic pH evident on
                        PH

                        60
 Aug.  18.  In  general,  however,  comparison of pH,
 alkalinity, and calcium of filtered water analyzed on
 the day of collection with unfiltered water analyzed 4
 days  after  collection suggests  that  settling of
 Snowhite 20-2  is  rapid.  Large differences  between
 filtered and unfiltered samples were noted only when
 water samples were obtained shortly after a drop. The
 pH profiles from Stations B and D on Aug. 18 showed
 a pH increase just above the bottom.
   The  volume-weighted  alkalinity  increased  from
 -0.7 mg/l on Aug. 10 to  3.6 mg/l on Aug. 18 and 4.2
 mg/l on Aug. 31. A metalimnetic maximum of 4.48 mg/l
 occurred at 10 m on Aug. 18 (Fig. 3). A metalimnetic
 calcium maximum also occurred (Fig. 4) but this was
 not supported by the presence of a pH maximum. Den-
 sity currents (swirling motions of water) were observ-
 ed by divers in the metalimnion on Aug. 23. Divers also
 observed a dark layer about 4-5 cm thick at the bot-
 tom of the epilimnion which was not present 1  week
 prior to  neutralization. Discoloration may have been
 caused  by solutes  or a suspension  of  very  fine
 material. A fine layer of calcite could be seen on rocks
 but not in soft sediment.
   Calcium data  are shown in Figure 4. The volume-
 weighted calcium  concentration increased  from 2.4
 mg/l on  Aug. 10 to 4.1 mg/l  on Aug. 18.  Hence, an
 average  of 4.3  mg/l  of calcite dissolved yielding  a
 dissolution efficiency of  40 percent which  is  lower
 than the expected value of 73 percent. The dissolution
 efficiency increased to 52  percent  by  Aug. 31. The
 ratio of dissolved calcite to total increase in alkalinity
 on Aug.  18 was 1.05:1. The low dissolution efficiency
 may have been  caused  by high concentrations of
 calcite particles descending  through the water col-
 umn. This is a function, in turn, of the very short time
 required  for the calcite to  clear the Canso holding
 tanks  and the fact that while an airborne cloud ap-
 peared visually dramatic, the bulk of the calcite could
 be seen shortly after a drop covering a relatively small
                  ALKALINITY , mg-l"1

            -1    0   1    23   45    6    7
     6



     3


     10

 E

 I   12 -
 I-
 Q.

 Q   1<*


     16 -



     18 -


     20 -
                                                                 10 •  T
                                                                                   AUG ' • AUG 18
                                                                                    16  ',
Figure 2.—pH versus depth in Bowland Lake (Station A). Ap-
plication of calcite to the lake surface occurred from Auci.
11-16. Water samples were filtered prior to analysis except
on Aug. 10.
Figure 3.—Gran alkalinity versus depth in Bowland Lake
(Station A). Application of calcite to the lake surface occur-
red from Aug. 11-16. Water samples were filtered prior to
analysis except on Aug. 10.
                                                  358

-------
                                                                                       ACIDIC PRECIPITATION
                 CALCIUM,  rng.T1

                      2       3
    8


    10
e
i
    18


    20
Figure 4.—Calcium versus depth in Bowland Lake (Station
A). Application of calcite to the lake surface occurred from
Aug. 11-16. Water samples were filtered prior to analysis ex-
cept on Aug. 10.

area of 0.25 to 0.50 ha. Third, the maneuvering of large
aircraft  over  lakes the size  of  Bowland  Lake is
restricted,  thereby  limiting the  lake surface  area
receiving calcite. Furthermore, complete mixing was
retarded by  thermal stratification which  would  in-
crease settling losses to epilimnetic sediments due to
lateral transport to shallow waters. Chemical disper-
sants such as sodium hexametaphosphate or sodium
pyrophosphate may increase the dissolution  efficien-
cy.
  Impact  of the liming on the biota is still  to  be
assessed  but no acute effects on the yellow  perch
population were noted  by divers swimming transects
or searching for dead fish  in nearshore areas.
  Future studies: Water  chemistry and perch and
plankton communities will  be monitored monthly until
the fall of  1985, thus allowing us to document  any
changes in community  composition and measure the
rate of reacidification. The  nearshore impact of spring
runoff will  be given special  attention  because of
potential detrimental effects on resident fish popula-
tions. Lake trout will  be reintroduced in the fall of
1983. Introduced trout  will be monitored for growth
rate, trace metal concentrations, and recruitment. It is
hoped that reclaimed lakes such as Bowland Lake will
support  healthy,  self-sustaining  sport fish popula-
tions.
CONCLUSIONS

Aerial application of calcite in dry form with a Canso
water bomber is an effective and logistically simple
method  for neutralizing  remote  lakes.  However,
preliminary analysis of the data indicates that calcite
dissolution efficiency was low, possibly because of
high concentrations of calcite particles descending
through  the water column.  The  liming of  Bowland
Lake did not induce acute mortality among the resi-
dent yellow perch population.

ACKNOWLEDGEMENTS: The Experimental Lake Neutraliza-
tion Project  is cosponsored  by the  Ontario Ministries of
Natural Resources and the Environment. The authors wish to
thank the members of the project Steering Committee and
Task Force for their efforts and input to the study design and
review of the manuscript.


REFERENCES

Dillion, P.J., N.D.  Van, W.A. Scheider, and N. Conroy. 1979.
  Acidic lakes in Ontario, Canada: characterization, extent
  and  response  to base and  nutrient additions.  Arch.
  Hydrobiol. Beih. 13: 317-36.
Fraser, J.E., and D.L Britt. 1983. Liming to mitigate surface
  water acidification: international  programs, strategies,
  and economic  considerations. In Lake Restoration, Pro-
  tection and Management. Proc. 2nd  annu. conf. N. Am.
  Lake Manage. Soc., Vancouver, B.C. Oct. 26-29,1982. EPA
  440/5-83-001. U.S. Environ. Prot. Agency, Washington, D.C.
Gunn, J.  1982. Pers. comm. Ontario  Ministry Nat. Resour.,
  Sudbury,  Canada.
Molot, LA.,  J.G. Hamilton, and G.M. Booth. 1983. Biological
  and chemical changes in enclosures of acidic lakewater
  treated with calcite. Unpubl. rep. Exper. Lake Neutral. Proj.
Scheider, W., B. Cave, and J. Jones. 1975. Reclamation of
  acidified  lakes near Sudbury, Ontario, by neutralization
  and  fertilization. Tech.  rep.  Ontario Ministry Environ.
  Toronto, Canada.
Sverdrup, H. 1983. Lake liming. Chem. Scripta 22:8-14.
Sverdrup, H., and  I. Bjerle. 1983. The calcite utilization effici-
  ency and  the  long term  effect on alkalinity  in several
  Swedish lake liming projects. Vatten 1:41-54.
                                                    359

-------
 ADIRONDACK EXPERIMENTAL LAKE  LIMING PROGRAM
 DOUGLAS L. BRITT
 JAMES  E. FRASER
 General Research Corporation
 McLean, Virginia
             ABSTRACT

             An experimental lake liming program in the Adirondack Mountains of New York State—the largest
             and most comprehensive in the United States—will initially entail neutralizing two acidified Adiron-
             dack lakes now devoid of fish. One lake will be limed with CaCO3 and maintained in a circumneutral
             condition. Brook trout will be stocked in the lake, and their survival and growth monitored  The se-
             cond lake will be treated in a similar mannei, but allowed to reacidify during the study period. The
             project will address (1) renovation and protection of fishery resources threatsned or affected by acidifica-
             tion, and (2) the most critical deficiencies h our understanding of the fundamental mechanisms
             associated with the neutralization of acidified surface waters. Seven activities constitute the core of
             this experimental program: (1) evaluation  ard selection of CaC03 treatment strategies; (2) predic-
             tions and monitoring of lake reacidification; (Ji) evaluation of effects of liming on acid lake sediments;
             (4) evaluation of effects of liming on phosphorus in the water column; (5) evaluation of effects of lim-
             ing on dissolved organic carbon; (6) determination of temporal and spatial variations in water column
             chemistry and metal speciation; and (7) determination of biological responses of fisheries, macroben-
             thos, and zooplankton to liming and reacidification. Data from Canadian and Scandinavian acid lake
             renovation projects will be analyzed and integrated into the Adirondack study.
 BACKGROUND

 Acidic deposition has been implicated in direct and in-
 direct  adverse effects on aquatic resources in some
 regions of North America and Scandinavia. Currently
 the major method of protecting and renovating acid
 sensitive lake ecosystems involves the application of
 base materials (usually  some  form  of limestone or
 hydrated lime), by boat,  aircraft, or  trailer (Scheider
 and  Dillon, 1976; Bengtsson et al. 1980; Blake, 1981;
 Fraser et al. 1982; Fraser and Britt, 1982).
   In Scandinavia, liming  programs began as early es
 1926, when operators of  salmon hatcheries began 1o
 experiment with chemical treatments to offset losses
 caused by acidic influent waters (Muniz, 1981). Todc.y
 operational liming programs exist in Sweden, Norwa/,
 Canada, and the United States; however, only the first
 three countries have established large-scale experi-
 mental programs designed to evaluate the efficiency
 of various treatment strategies (Fraser  and Britt,
 1983). In the United States, over 100 individual liming
 projects have been initiated during the last 25 years to
 reduce  the  acidity  of   surface  waters  allegedly
 acidified by  acidic deposition or by  unknown and
 presumably undeterminable sources (Fraser and Britt.
 1982). Most of these projects have been conducted in
 the following States: Massachusetts,  New York, North
 Carolina,  Pennsylvania,   Rhode  Island, and West
 Virginia. Most  have been implemented as  a "last
 resort" to restore fisheries in affected bodies of water.
 Few of the projects have entailed either the collection
 of baseline data or post-treatment monitoring.
  With  the increasing concern  of both private and
 governmental organizations regarding the potential
 acidification of surface waters in the  United States, it
 is anticipated that liming  will receive  increased atten-
tion  as a  management technique to ameliorate sur-
face water acidification. As a consequence, the poten-
tial benefits, impacts, and uncertainties of liming sur-
face waters to  protect or restore aquatic  resources
 need to be identified. The effectiveness of alternative
 liming techniques and strategies also needs to be
 assessed. This information will be useful in develop-
 ing optimum acidic  deposition control strategies. It
 also will be particularly relevant to fishery biologists
 and aquatic resource managers who must address the
 problem of managing acidified or acidifying lakes and
 streams.
   Based upon recent reviews of international  and
 domestic  liming  programs  (Britt  and Fraser,  1983;
 Fraser and  Britt, 1983) and the results  of an  Inter-
 national Liming Workshop (1983), it is reasonable to
 conclude that a research program is needed. Critical
 data deficiencies must be addressed, as must con-
 tradictory results of recent chemical neutralization
 projects. This research will  enable aquatic resource
 managers to assess the efficacy and risks of future
 U.S. liming projects.
   The  Ecological  Studies Program of the  Electric
 Power Research Institute (EPRI) has initiated a Lake
 Acidification Mitigation Project (LAMP) designed to
 address such concerns. This project is the largest and
 most comprehensive experimental liming  aquatic pro-
ject in  the United States.
OVERVIEW OF LAMP

The project is initially scheduled as a 3-year program
consisting of three major tasks: (1) historic and inter-
national data analysis, (2) field and laboratory experi-
ments, and  (3) program integration. Each of these
tasks is  further divided into several subtasks  (see
Table 1). The program was begun in August 1983. Ma-
jor  milestones  and  the anticipated period  of  per-
formance for major project activities are illustrated in
Figure 1.
  The research project is being coordinated by the
General Research Corp. (GRC) of McLean, Va., and in-
                                                 360

-------
              YEAR AND MONTH
 MAJOR ACTIVITIES
  AND SUBTASKS
   INTERNATIONAL AND HISTORIC DATA
     ANALYSIS
   SELECTION OF EXPERIMENTAL LAKES
     AND FIELD BASE EQUIPMENT
     ACQUISITION

   EVALUATION AND SELECTION OF
     TREATMENT STRATEGIES
   PREDICTION AND MONITORING OF LAKE
     ACIDIFICATION
   EVALUATION OF EFFECTS OF IIMING ON
     ACID LAKE SEDIMENTS
   EFFECTS OF LIMING ON PHOSPHORUS
   EFFECTS OF LIMING ON DISSOLVED
     ORGANIC CARBON

   TEMPORAL AND SPATIAL VARIATIONS II
     WATER COLUMN CHEMISTRY AND
     METAL SPECIATION

   DETERMINATION OF BIOLOGICAL
     RESPONSES
   DATA MANAGEMENT

   PREPARATION OF REPORTS
                                        BASE
                              1Qfld    ADDITION
                              1984    LAKES 1,2
                                        SEPT
 REPEAT
  BASE
ADDITION
 LAKE 2
  SEPT
                                     1   2   3   4   5   6   7   8   9  10   11   12   13   14   15  16  17  18  19  20  21  22  23  24  25   26   27  28  29  30   31   32   33 34  35  36  37  38  39  40  41
                                                BASE LINE OATAiLAKtS 1 ?
BASE LINE DAIAtt
              PLANKTON BASE II


+
o
B
•Jf
A

o

INITIAL FORECAST OF COSTS
MONTHLY COST REPORTS
QUARTERLY PROGRESS REPORTS
DRAFT FINAL REPORT
FINAL REPORT
TASK 1 TRIP REPORT
S
PLANNING WORKSHOP REPORT
Figure 1.—Schedule of  major activities  and subtasks.

-------
LAKE AND RESERVOIR MANAGEMENT
volves major contributions from Clarkson College of
Technology, Cornell  University, Syracuse University,
New York State Department of Environmental Conser-
vation, and the U.S. Geological  Survey. The organiza-
tion of the project is illustrated  in Figure 2, and func-
tional responsibilities for  major tasks and subtasks
are displayed in Figure 3. Principal investigators for
General  Research  and the universities are  listed n
Table 2. The following sections describe in more detail
the LAMP activities associated  with Tasks 1 and 2.
Task 1: Historic and international Data
Analysis

This task continues work previously begun by General
Research to Identify and assess the relevance of com-
pleted/ongoing international liming projects (Fraser et
al. 1982). Information  obtained during the analysis of
existing  data  and historic  experience will be  in-
tegrated with  information developed from  Task 2,
Field and Laboratory Investigations.  In this  manner,
information obtained from  foreign researchers and

   Table 1.—Tasks and subtasks of Core LAMP Program.

 Task 1: Historic and International Data Analysis
     •  Evaluation of stream liming practices.
     •  Evaluation of biological response data.
     •  Evaluation of water quality responses
     •  Evaluation of costs of operational liming
       programs.
     •  Updating of liming directory
     •  Task review meetings.

 Task 2: Field and Laboratory Experiments
     •  Selection of experimental lakes.
     •  Evaluation and selection of treatment
       strategies.
     •  Prediction and monitoring of lake
       acidification.
     •  Evaluation of effects of liming on acid lake
       sediments.
     •  Evaluation of effects of liming on phosphorus
       in the water column.
     •  Evaluation of effects of liming on dissolved
      organic carbon.
     •  Determination of temporal and spatial
      variations in  water column chemistry and
      metal speciation.
     • Determination of biological responses of
      fisheries, zooplankton, and macrobenthos to
      liming and reacidification.

Task 3: Program Integration

     • Quality assurance.
     • Data management
     • Review of work.
     • Report  preparation.
     • Field coordination.
 from analyses of existing empirical data bases can be
 made available  to the field researchers if it is deter-
 mined to be relevant to the  solution of a problem,
 evaluation of an hypothesis, or verification  of pre-
 liminary field research results.
   Similarly, field research activities may provide new
 hypotheses that can  be  tested on the empirical data
 bases compiled in  Task 1. Some critical topics that
 cannot  be adequately addressed at this time by  the
 field research program will also be evaluated  in Task
 1. Examples of such topics include the effectiveness
 of base addition to  lotic systems and the comparison
 of  base  addition   responses   in  clearwater  and
 dystrophic ecosystems.
   Much  of the historic information on  acid lake
 renovation is of  Scandinavian  origin. Several Swedish
 and Norwegian  research organizations (see Table 3)
 are  cooperating with the LAMP study team  by pro-
 viding data, assisting  in analyses, and/or participating
 as project advisors. Such international cooperation is
 anticipated to continue in future years. In addition to
 the Scandinavian data, the LAMP researchers will at-
 tempt to include data from studies conducted by the
 Ontario  Ministry of  the Environment, Massachusetts
 Department of Environmental  Quality, West Virginia
 Department of  Natural  Resources,  and New  York
 State Department of Environmental Conservation.
   Major subtasks related to Task 1 are as follows:
   Subtask 1: Evaluation  of Stream Liming. The suc-
 cessful neutralization of  streams has been identified
 as one of the most difficult objectives to achieve in
 acidification  mitigation  programs  (Swedish  Minist.
 Agric. Environ.  Comm.,  1982).  However, to  protect
 stream spawning fish and downstream lakes  from
 periodic acid surges,  some evaluation of base addi-
 tion strategies will have to be made relative to the lim-
 ing of streams.
   The project team  investigators will  obtain and
 automate several stream  liming  data sets in order to
 address  some of the concerns  of  neutralizing  lotic
 systems. Preliminary discussions regarding the avail-
 ability of the data from liming projects in Otter Creek,
 W.Va.  (West  Virginia  Department  of  Natural
 Resources); Unnamed Creek, Ontario (Ontario Ministry
DC
"
SAGE/SEOIME
CHEMISTRV
U..ONCOU
"'
OE
BIOL
CORNELL
ONSEB
Figure 2.—Organization of the Lake Acidification Mitigation
Project.
                                   Table 2.—LAMP Principal Investigators.
                      General Research Corporation
                      Clarkson College of Technology
                      Cornell University
                      Syracuse University
   Douglas Britt, James Fraser
   Joseph DePinto, Thomas Young
   Steven Gloss, Carl Schofield
   Charles Driscoll
                                                  362

-------
                                                                                      ACIDIC PRECIPITATION
 of  the   Environment  and   Ministry  of  Natural
 Resources);  Hogvadsan and  Anrasan, Sweden  (Na-
 tional Fisheries  Board of Sweden);  and an influent
 stream of Lake Hovvatn, Norway (Norwegian Institute
 for Water Research),  have already been  conducted
 with the principal researchers associated with each of
 these projects. Analyses and evaluation of  these
 stream liming data will address some of the uncertain-
 ties associated with  the  liming  of  running waters.
 Analytical activities will focus on the reactivity/deacti-
 vation of limestone, the speciation and precipitation
 of metals, and the ability of the treatment systems to
 neutralize acid waters under fluctuating flow regimes.
   The Lund  Institute of Technology is developing a
 calcite dissolution model  that may be applicable to
 running  water (Sverdrup, 1983; Sverdrup and  Bjerle,
 1983). If this model proves successful, it may  be ap-
 plied to existing stream liming data bases containing
 sufficient water quality and hydrologic information.
   Since the field research components to Task 2 are
 not focused  on  lotic systems, this subtask will com-
 plement the field program in a critical area. The data
 obtained on  the  response of  aluminum  and other
 metals to base addition in streams also may directly
 apply to the understanding of the fate and speciation
 of metals at the lake/stream interfaces in Task 2.
   Subtask 2: Evaluation of Biological Response Data.
 This  subtask involves  assessing   biological  data
 recently compiled for several Swedish and  Norwegian
 limed lakes.  It is anticipated that  biological response
 data  acquired  from  Scandinavia  will supplement
 biological investigations being conducted in Task 2
 (fish, zooplankton, and macrobenthos) and facilitate
 evaluation of the research. Furthermore, international
 data  on  phytoplankton,  macrophyte, and benthic
 microbial community structure and biological produc-
 tivity in treated lakes will complement work being per-
 formed on other trophic levels in Task 2.
  Subtask 3: Evaluation of Water Quality Responses.
 Existing empirical data bases also will be used to  test
 hypotheses relevant to the water quality components
 of Task 2.  General  Research has   been provided
 physical and  chemical data from limed lakes in the
 United   States,  Sweden,  Canada,  and   Norway.
 Preliminary observations of  these data suggest a
 more detailed evaluation of the relationship between
 base addition and alkalinity maintenance is warranted
 for these and  other lake data sets. Additionally, water
 quality responses to base addition will be analyzed by
 using  cluster analyses,  regression  analyses,  and
 where  data  permit,  other  multivariate  statistical
 techniques.
   Specific hypotheses to be tested include:
   • Overdosing of CaCO3 produces a long-term buf-
 fering effect (either through long-term dissolution pro-
 cesses  or by altering cation exchange capacity of
 sediments).
   • CaCO3 addition changes the water column con-
 centrations of potentially toxic metals in a predictable
 manner.
   • Dystrophic  systems respond differently to base
 addition than  clearwater oligotrophic systems.
   Initial testing  of these specific hypotheses will rely
 significantly  upon available  Scandinavian data.
 Although data have been collected from many Scan-
 dinavian liming  projects,  the  frequency of  chemical
 analyses and the total number of in-lake sampling sta-
 tions are often inadequate for meaningful statistical
 analyses. Therefore, prior to hypothesis testing, the
 foreign  data  will  be  carefully  screened (with  the
 assistance of  Swedish and Norwegian researchers) to
 select the most  appropriate and comprehensive data
 sets for hypotheses testing. If data permit, additional
 hypotheses also may  be tested (see  Table 4). New
 hypotheses developed  from  observations of prelimin-
 ary research results from Task 2 also  may be tested
 with the foreign  data.
   In addition, information on the applicability of ex-
 isting,   computerized,  reacidification  models and
 models   to  predict the  dissolution  rate  of base
 materials will  be assessed for use in the Task 2 field
experiments in Adirondack lakes.
  Subtask 4: Evaluation of Costs of Operational Lim-
ing Programs. The  economic  costs associated with
         Table 3.—Preliminary list of Scandinavian research organizations scheduled to participate/cooperate
	with LAMP study.

• Swedish National Board of Fisheries—Goteborg, Sweden
• Swedish Water and Air Pollution Research Laboratory—Goteborg, Sweden
• Swedish National Environmental Protection Board—Solna, Sweden
• Institute of Freshwater Research—Stockholm, Sweden
• Lund Institute of Technology—Lund, Sweden
• Swedish University of Agricultural Sciences—Uppsala, Sweden
• Norwegian Institute for Water Research—Oslo, Norway
• Norwegian Liming Project  Office—Arendal, Norway


                       Table 4.—Additional hypotheses that may be tested if data permit.

• Aluminum-phosphorus relationships are significantly altered by base addition
• Seepage  lakes  respond  to  liming differently  than drainage  lakes m terms of water chemistry (alkalimty/phos-
  phorus/metals).

• Lakes in watersheds impacted by  melting snow pack respond  differently (pH/alkalinity/metals)  than  lakes  in
  geographical areas without snow pack

• CaCO3  reactivity  and resulting  alkalinity  values  are affected differently by the timing  of  base addition [e g,
  summer versus winter (ice)].

• Shoreline application of  CaCo3 is  just as effective as whole lake liming  (on an  equivalent  weight  basis)  in
  surface water neutralization and reduction of aluminum concentrations.

• Watershed  liming  combined with  lake liming  is more effective  than  only  surface water  liming  in  reducing
  surface water aluminum concentrations.
                                                  363

-------
 LAKE AND RESERVOIR MANAGEMENT
 operational and large-scale experimental liming pro-
 grams will be documented from recent Scandinavian
 projects. Similar information will  be requested from
 Canadian  provincial  and  Federal agencies.  During
 previous assessments of historical liming data, cost
 information for materials, transportation, equipmenl,
 and to a limited extent labor, have been summarized
 (Blake, 1981; Fraser and Britt, 1982, 1983; Natl. Fish.
 Board Sweden,  1982). This information will  be sup-
 plemented  with  more detailed manpower  re-
 quirements (converted to U.S.  dollars), data on pre-
 liming planning costs, and post-treatment water quali-
 ty monitoring costs  associated with  various liming
 strategies. This will provide for better estimates of the
 actual costs of liming.
   Subtask 5: Updating of Liming Directory. Fraser et
 al. (1983) recently prepared an "International Directory
 of Data Bases of Limed Aquatic Ecosystems." Addi-
 tional physical, chemical, and biological data on limed
 surface waters obtained from foreign research institu-
 tions and agencies will be summarized and included
 in an updated version of the Directory.
   Subtask 6: Task Review  Meeting. At  least one
 researcher associated with the primary data bases of
 each country (Sweden,  Norway, and Canada) will be
 invited to participate in a preliminary review of Task 1
 activities  at the  end of the first  project year. This
 review process will assure that the results of Task  1
 are interpreted with the benefit of a more complete
 understanding  of  the  caveats  and  background
 associated with the foreign liming projects. Additional
 guidance  for continuing  analytical   activities and
 future hypothesis testing will be  solicited from the
 review panel.

 Task 2:  Summary: Field and Laboratory
 Investigations

The experimental design for the field and laboratory
 research  components of this project address both (1)
 fishery management concerns related to the protec-
tion, enhancement, and/or rehabilitation of fisheries
 affected or threatened by acidification; and (2) funda-
 mental  scientific  issues  associated  with  under-
standing  the responses and interactions of sediment,
water chemistry, and biota after base addition.
  The preferred field research program would involve
whole lake liming under the following three sets o'
conditions:
   • Liming  of an  acidified lake to permit establish-
 ment of a stocked fish population, followed by natural
 reacidification to examine population responses
   • Liming  and  maintenance of an acidified lake to
 permit  reestablishment  of  a  self-reproducing  fish
 population.
   • Maintenance liming of an acidified lake where ex-
 tant fish populations exhibit acid stress symptoms.
   Because of budget  limitations, the  maintenance
 liming project involving a lake having the last set of
 conditions has been deferred.
   Field  experiments will initially be conducted in two
 clearwater, oligotrophic,  drainage lakes  within  the
 Adirondack Mountains of New York State. The Adiron-
 dacks were selected because of (1) the susceptibility
 of the  region to acidification, (2) availability of fishless
 lakes,  (3) availability of at least some useful baseline
 data, and (4) an established State  liming program in
 New York.
   Some of the selection criteria for candidate experi-
 mental lakes (referred to as  Lakes 1 and 2,  corres-
 ponding to the aforementioned general conditions)
 are described  in Table 5. Other criteria include  ac-
 cessibility, ownership, and ability  to gauge influent
 and outfluent streams.
FUNCTIONAL RESPONSIBILITIES
DATA MANAGEMENT
PROGRAM INTEGRATION
ANALYSIS OF HISTORICAL AND INTERNATIONAL DATA


PREDICTION AND MONITORING OF LAKE ACIDIFICATION
SEDIMENTS
EVALUATION OF EFFECTS OF LIMING ON PHOSPHOROUS

DETERMINATION OF TEMPORAL AND SPATIAL VARIATIONS
DETERMINATION OF BIOLOGICAL RESPONSESOh FISHERIES
ZOOPLANKTON ANO MACHO8ENTHOS TO LIMING AND
REACIDIFtCATION
PREPARATION OF INTERIM AND FINAL REPORTS
QUALITY ASSURANCE
S
•
•
•
o
0
o






•
•
1
o


•
o
o
A
A
A


•
o
•
M
D
M
O


o
o
o
•



•

o
•
K
O
0


o






o

o
•
NY DEC 1



0
0






o

n
i










•
O

o
              • PRIMARY RESPONSIBILITY
        LEGEND   O SECONDARY RESPONSIBILITY
              A RESPONSIBLE FOFI SAMPLE COLLECTION ONLY


Figure 3.—Functional responsibilities of major project par-
ticipants.
                          Table 5.—Some selection criteria for LAMP candidate lakes.

Lake 1 (Liming
followed by
reacidification)
Lake 2 (Liming
and maintenance


Lake 3* (Liming and
maintenance to
protect native
fish population
Flushing
Rate '
High


Med-High



Low



PH
<50


<50



<5.5



Fishery
Size Status
< 25 ha Absent (but
historically
present)
< 25 ha Absent (but
historically
present and
reproducing)
< 50 ha Present but
stressed
(historically
reproducing)
Fish Spawning
Habitat Al * + *
Not necessary High


Necessary Med-High



Necessary Med-High



DOC
Low


Low



Low



Some candidate lakes

Lake 1: Cranberry Pond
Lake 2  Woods Lake
Lake 3*  Little Simon Pond
'The third lake is deferred from the Core Program but is recommended for addition to program m 1984
                                                 364

-------
  Using the specific selection criteria, two acidified
drainage lakes, initially devoid of fish, will be selected
for the field research program. At this  writing, only
Woods Lake (see Fig. 4) in the Big Moose watershed
has been established as a LAMP experimental lake.
Negotiations are currently underway with the owners
of other lakes for inclusion in the program. One of the
lakes will be chemically neutralized, stocked with fish,
and  maintained  in  a  circumneutral  condition
throughout the  duration of the  project.  The other
acidified lake  will  be neutralized, stocked  with fish,
and then allowed to reacidify.
  Examining chemical changes during the reacidifica-
tion process will  facilitate the assessment of the long-
term implications  of  maintenance liming practices.
The  initial  two study lakes will  have similar water
quality and biological characteristics and both will be
limed in the fall  of 1984 or possibly  in the spring of
1985. It  is  anticipated that reapplication of  base
material will be  performed in the autumn of 1985  in
Lake 2  to  maintain  it in  a circumneutral  condition
through the end  of the research project.
                               ACIDIC PRECIPITATION

  Baseline data to be collected prior to the treatment
of the experimental lakes will include a characteriza-
tion of sediments, soils, hydrology, and lake morpho-
metry. These baseline data will be integrated with the
existing background data for the two lakes. An inten-
sive  sediment and  surface water quality monitoring
program  will  be  implemented in  the experimental
lakes. Some of the major physical and chemical para-
meters to be measured are listed in Table 6. Biological
monitoring  of  fisheries, zooplankton,  and  macro-
benthos also will be conducted in the lakes. Table 7
summarizes the major biological monitoring activities
before, during, and after treatment of the two experi-
mental lakes.
  Specific subtasks to be performed under Task 2 are
described in more detail in the following sections:
  Subtask  1: Evaluation and Selection of Treatment
Strategies. Jar tests, a pH-stat apparatus, and contin-
uous-flow microcosms will be used to evaluate a varie-
ty of  candidate  neutralization  materials  and  to
develop a model for calculating appropriate dosages
prior to lake treatment.
                         Table 6.—LAMP sediment and water quality monitoring program.
Measured Parameters
Temperature
DO
pH
ANC
DOC
DIG
SO4
NO,
NH4
Cl
F
Ca
Mq
Na
K
Al (Monomeric)
Al (Acid soluble)
Al (Organic rnonomenc)
Fe
MM
Pb
Zn
P (Dissolved orthophosphate)
P iTotal)
Chi a
Turbidity
Suspended solids
Moisture content
Loss on ignition
Sediment size
Sediment density
Total exchangeable cations
Carbonates
Sediment alkalinity demand
Trace metal fractionation
Phosphorus fractionation
Water
Sediment Column
Collection Collection
Sites Sites
X
X
X X
X
X
X
X
X
X
X
X
v X
Y X
X
X
X
x x
X
X X
'« K
'

X v
X \
K
X
X
X
X

X
X
X
X
X
X
                                                  365

-------
LAKE AND RESERVOIR MANAGEMENT
   Subtask  2:  Prediction  and  Monitoring  of  Lake
Reacidification. Appropriate dose/response model(s)
will be applied to the experimental systems to predict
immediate dissolution of base material and the dura-
tion  of the effect. The reacidification process in the
treated   lakes   will   be  monitored,  and  the
dose/response model(s) will be refined based on com-
parison of simulations  with field observations. The
model(s) also will be used to evaluate the significance
of each acid-contributing  component  of the lake
systems.
   The main chemical parameters to be monitored in-
clude  pH,  ANC,  dissolved  inorganic carbon,  and
calcium. Intensive chemical monitoring will  be per-
formed on  each treated lake for 1 to 2 weeks after
treatment;  this short-term  monitoring will determine
the immediate efficiency of the treatment in terms 01
                                               lake buffer system response and residual undissolved
                                               neutralizing   material.  Long-term   monitoring
                                               associated with  this  subtask will  focus on docu-
                                               menting the in-lake time profile of the major chemical
                                               parameters during the reacidification process.
                                                 Subtask 3: Effects of  Liming on  Acid Lake Sedi-
                                               ments. The extent to which lake sediments can affect
                                               water column acidity, before and after liming, and the
                                               effect  of liming on major  physical and chemical char-
                                               acteristics of  sediments will be investigated.
                                                 Subtask 4: Effects of Liming on  Phosphorus in the
                                               Water  Column.   Phosphorus will  be  carefully
                                               monitored to evaluate the  direction and extent to
                                               which  liming  acid lakes  affects  concentrations of
                                               phosphorus, the nutrient  that  most commonly limits
                                               productivity  in  high  altitude, oligotrophic  aquatic
                                               systems.
   (0
   o>
   <0
   £

   I
   Q.
             Table 7.—Biological monitoring activities scheduled for the two experimental lakes systems.

                                                                                Location
                           Biota
 Fisheries
 1. Cage survival experiments
 2. Pre-liming survival and growth rates
 3  Egg and fry survival
 4. Emigration during episodic acidic events

 Zooplankton
 1. Species richness and individual abundances
 2. Size frequency distribution
 3. Developmental stages
 4. Biomass
 5. Reproductive rates

 Macrobenthos
 1. Taxa to lowest practical level
 2. Relative abundances
                                                            Lake 1
                                                        (Reacidification
                                                            lake)
                          Lake 2
                       (Maintenance
                        liming lake)
                                                                      X
                                                                      X
 X

 X

 X

 X

 X


 X

 X
                            X


                            X
                                                                                               X

                                                                                               X

                                                                                               X

                                                                                               X

                                                                                               X


                                                                                               X

                                                                                               X
   
-------
                                                                                        ACIDIC PRECIPITATION
Figure  4.—One of the Adirondack lakes (Woods  Lake)
selected for experimental neutralization.
   Subtask 5: Effects of Liming on Dissolved Organic
Carbon. A previous study of an acidified Adirondack
lake suggested that dissolved organic carbon (DOC)
did not  return to pre-liming levels after  other water
quality parameters indicated  that the lake  had re-
acidified (Driscoll et al. 1982).  Since DOC may be im-
portant in determining the availability and toxicity of
aluminum, the  extent  to  which  liming affects DOC
concentration and the relative abundance of organic
metal-binding ligands and humic substances will be
evaluated in the study lakes.
   Subtask  6:  Temporal  and   Spatial  Variations  in
Water Column Chemistry and Metal  Speciation. The
temporal and spatial variations  in the water quality
parameters identified in Table  6 will be monitored. The
extent of changes to trace metal (Al, Fe, Mn, Zn,  Pb)
cycling and pools within acidic lake systems also will
be determined after base  addition.
   Subtask 7: Biological Response to  Liming. Optimal
restocking strategies for brook trout in acidified lakes
will be developed based upon  measured responses of
various life history stages and selected strains to lim-
ing.  Both acute  and chronic effects  associated with
the chemical treatment will be determined for fish,
zooplankton, and macrobenthos in the experimental
lakes. Special emphasis also will be  placed upon  the
relative  importance of episodic intrusions of waters
with low pH and/or elevated metal  concentrations to
the biota of  lakes neutralized  by  liming.
Future Modifications

It is anticipated that the LAMP study will be expanded
in 1984 to include a third lake (Little Simon Pond). This
lake contains  existing but seriously stressed brook
trout and lake  trout fisheries. Adding this lake to the
project will permit us to investigate the effects of lim-
ing on existing fish stocks. For example, changes in
body  burdens  of some metals after  liming  can be
studied in the  resident fish populations.
   In addition,   it  is  anticipated that  phytoplankton
abundance and species distributions,  as well as pro-
ductivity, will be  measured in all  of the study lakes
during 1984.
                                                       CONCLUSIONS

                                                       We  believe  the EPRI-sponsored  Lake Acidification
                                                       Mitigation Project will help answer some of the major
                                                       questions regarding the long-term effects of liming on
                                                       aquatic  ecosystems.  In addition, it should  provide
                                                       useful information to fishery  biologists and  aquatic
                                                       resource managers who must now devise renovative
                                                       strategies for lakes that already are acidic and others
                                                       that may be undergoing acidification.
REFERENCES

Bengtsson, B., W. Dickson, and P. Nyberg. 1980. Liming acid
  lakes in Sweden. Ambio 9:34-26.
Blake, L. 1981. Liming  acid ponds in  New York. N.Y. Fish
  Game J. 28:208-14.
Britt, D.L, and J.E.  Fraser. 1983. Effectiveness and uncer-
  tainties associated with  the chemical neutralization of
  acidifed surface waters. Pages 96-103 in Lake Restoration,
  Protection and Management. Proc. 2nd Annu. Conf. N. Am.
  Lake Manage. Soc. Oct. 26-29, 1982, Vancouver, B.C. EPA
  440/5-83-001. U.S. Environ. Prot. Agency, Washington, D.C.
Driscoll, C.T.,  J.R. White, G.C. Schafran, and  J.D.  Rendall.
  1982. Calcium carbonate neutralization of acidified sur-
  face waters. J.  Environ. Eng. Div. Am. Soc. Civil Eng. 108:
  128-45.
Fraser, J.E., and D.L. Britt. 1982. Liming of acidified waters:
  a review of  methods and effects on aquatic ecosystems.
  U.S. Fish Wildl. Serv. Off. Biolog. Serv. Eastern Energy and
  Land Use Team. FWS/OBS-80/40.13. Kearneysville, W.Va.
	1983. Liming  to mitigate surface water acidifica-
  tion: international programs,  strategies,  and economic
  considerations. Pages 141-147 in Lake Restoration, Pro-
  tection and Management. Proc. 2nd Annu.  Conf. N. Am.
  Lake Management Soc. Oct. 26-29, 1982, Vancouver, B.C.
  EPA 440/5-83-001,  U.S. Environ. Prot. Agency, Washington,
  D.C.
Fraser, J.E., D. Hinckley, R. Burt, and R. Severn. 1982. Feasi-
  bility study to utilize liming as a technique to mitigate sur-
  face water  acidification.  Electric  Power Res.  Inst.
  EPRI-2362. Palo Alto, Calif.
Fraser, J.E., R. Fares, and R. Chadduck. 1983. International
  Directory of Data Bases of Limed Aquatic Ecosystems.
  Palo Alto, Calif.: Electric Power Res. Inst. Interim rep. for
  RP1109-14.
International Liming Workshop. 1983. The liming of acidified
  waters: issues and research - a rep. Water Air Soil Pollut.
  (in press).
Muniz, I.P. 1981.  Acidification and the Norwegian salmon.
  Pages 65-72 in L. Sochasky,  ed. Acid Rain and the Atlantic
  Salmon. Proc. Conf. on Acid  Rain and the Atlantic Salmon.
  Nov. 22-23, 1980, Portland, Maine.  Int. Atlantic Salmon
  Found. IASF No. 10.  New York.
National Fisheries Board of Sweden. 1982. Rad och riktlinjer
  for kalkning  av sjoar och vattendrag. Rep. No. 1.
Scheider, W., and P.J. Dillon.  1976. Neutralization and ferti-
  lization of acidified  lakes near Sudbury, Ontario. Water
  Pollut. Res. Can. 11:93-100.
Sverdrup, H.U. 1982. Lake liming. Chem.  Scripta 22(1): 12-18.
Sverdrup, H.U., and I. Bjerle. 1983. The calcite utilization effi-
  ciency  and  long-term  effect  on  alkalinity  in several
  Swedish lake liming projects. Vatten 39(1): 41-54.
Swedish Ministry of Agriculture Environment Committee.
  1982. Acidification Today and Tomorrow. Risbergs tryckeri
  AB, Uddevalla, Sweden.
                                                    367

-------
 CONSIDERATIONS OF PRUDENCE AND  EQUITY FOR PROTECTING
 LAKES  FROM ACID PRECIPITATION
 ALFRED M. DUDA
 Environmental Quality Staff
 Tennessee Valley Authority
 Knoxville, Tennessee

            ABSTRACT
             Deposition of atmospheric pollutants is occjrring over much of eastern North America. While
             evidence exists of long-term reductions in pH and alkalinity of sensitive surface waters and
             adverse impacts on aquatic life, much of it is circumstantial. Cause and effect relationships
             have not been substantiated and many scientific uncertainties  must be resolved before deci-
             sions concerning emissions controls can be made. This paper addresses the issues of informa-
             tion gaps, scientific uncertainties, and risks in making policy decisions to protect sensitive lake
             resources from acid precipitation. Circumstantial evidence may be all that decisionmakers can
             realistically expect from science in the short term, and information gaps inevitably will remain.
             On such an issue of public significance as ac d precipitation, a cogent understanding of existing
             scientific facts and the use of prudence and equity are needed on the part of decisionmakers to
             ensure that necessary steps are taken in  he face of uncertainty to protect sensitive lake
             resources from acidification.
 INTRODUCTION

 Wet  and  dry atmospheric  pollutants  are  being
 deposited  over much of  eastern  North America.
 Evidence exists of long-term reductions  in pH  and
 alkalinity of sensitive surface waters. More than  half
 of 214 high elevation lakes sampled in the Adiron-
 dacks in 1975 had pH levels less than 5.0 (Schofield,
 1976). Of 95 New England lakes with pH data available!
 from the 1930's, more than 60  percent have shown
 trends of  decreasing  pH  (Haines and Akielaszek,
 1983).
   Adverse changes  in aquatic  life have been  cor-
 related with the reductions of  pH. Atlantic Salmon
 populations have disappeared from nine rivers in Nova.
 Scotia, but they remain in adjacent rivers with higher
 buffering capacity (Farmer et al.  1980). While these ef
 fects have been attributed  to acid precipitation, the
 evidence is largely circumstantial in nature, cause and
 effect relationships have not been substantiated,  and
 consequently some groups say  decisions to control
 emissions should not be made.
  This paper examines the role that information gaps,
 scientific uncertainties, and risks can play in making
 policy decisions to protect sensitive  lake resources
 from  acid precipitation. It also examines the role of
 government in protecting environmental  resources
 held in trust for the public minimizing external  costs
 and  risks imposed on society.  In complex natural
 resource management problems, there will always be
 uncertainty and risk that science  is not able to resolve,
 and  consequently, decisionmakers cannot wait for
 answers  to every  possible scientific  uncertainty.
 Policy decisions must be  based  on the best existing
 information, and considerations of prudence and equi-
 ty must become more dominant in choosing a course
 of action.


UNCERTAINTY AND ACID PRECIPITATION

If a decision is to be made with certainty, the decision-
maker must have complete  and  accurate knowledge
of the consequences of alternative actions. With to-
 day's complex technological problems, such certainty
 is not possible  and risks  are associated with any
 chosen course of action. Uncertainty may result from
 errors in  measurement, the  variability  of complex
 natural processes, man-made conditions, budgetary
 or  technical  limitations  for  data collection  and
 analysis,  personal  biases,  and  random  events.
 Science  cannot  eliminate  such  uncertainty  and
 associated risks, but it can reduce them.
   Many issues of uncertainty  have been raised in the
 acid precipitation debate. The magnitude and extent
 of the problem, significance of economic and environ-
 mental damage, and the effectiveness of mitigative
 measures are subjects of wide dispute among dif-
 ferent interest groups and scientists. This dispute
 centers around  the  extent of  information  needed
 before decisions about control  measures can  be
 made, the practicality of demonstrating cause and ef-
 fect relationships when there are financial and scien-
 tific limits to such understanding, and  the confusion
 that is caused when special interest groups make in-
 accurate or misleading statements.
  Acid precipitation  is not a  new issue. In the mid
 1600's, long-range transport of air pollutants between
 England and France was identified, and the use of tall
 chimneys was suggested to disperse these pollutants;
 and by the 1850's an English chemist noted damage to
 plants and materiafs caused by "sulfur acid" in the air,
 described  long-range transport of pollutants, and
 coined the term acid  rain (Cowling,  1982). By the late
 1800's, vegetation damage was evident on over 30,000
 acres in three Southeastern States from a smelter
 located at Copperhill, Term. Evidence from paleolog-
 ical studies of lake sediments in the northeastern por-
 tion of the United States found acidification of lake
 waters and deposition of lead and zinc dating back to
 the mid 1800's (Hanson et al. 1982). Similar evidence
 has been found in Scandanavia (Davis et al. 1980).
  At issue is the extent of information needed before
decisions can be made and whether cause and effect
proof  can  be  demonstrated with certainty in a short
                                               368

-------
                                                                                    ACIDIC PRECIPITATION
time frame. Groups opposed to acid precipitation con-
trols point out legitimate uncertainties such as detail-
ed chemical  reactions  occurring  in clouds and in-
tricate processes occurring  in terrestrial and aquatic
ecosystems upon which to base dose-response rela-
tionships. There is  no  consensus as to how much
basic research  into natural processes, how much
testing  of  hypotheses, or  how detailed  an  under-
standing of cloud physics is needed before a "sub-
stantial" or a "reasonable" basis may be established.
The result is a call  for more research on the part of
some advocates before  decisions can be made.
  A  more fundamental question involves  the suffi-
ciency of  field studies that are circumstantial  in
nature and do not vigorously demonstrate cause and
effect relationships. At best, field studies provide data
for  statistical  associations  between variables and
cannot prove causality.  Proving causality in the field
for every investigation  is an impossible task given
variability in  natural processes, lack of appropriate
control sites for comparison, lack of historical data,
and the  length of time  required to document  subtle
changes such as those that may occur in soils and
forests. There  are practical  financial and scientific
limits to man's ability to  understand  natural pro-
cesses. Decisionmakers cannot wait for all scientists
to agree or all questions to  be answered before they
act if damage is suspected.
  Misinformation can also create  uncertainty and
confusion.  In  particular,  there is  a  tendency  for
organized groups to represent their particular views
and biases in the political decisionmaking arena by
using scientific evidence in a selective manner  and by
making misleading  statements  to support their posi-
tion. In a report to Congress, the Comptroller General
of the United States (1981) examined the acid precipi-
tation debate and  noted  that misleading and inac-
curate statements were commonly being  made and
that  they were causing  confusion. Such misinforma-
tion reaching decisionmakers has inhibited the search
for  consensus,  obscured  the findings of important
scientific research,  and  confused the public.
  To better appreciate the nature of this uncertainty,
an example of current allegations of natural soil acidi-
ty being responsible for recent acidification of surface
waters is examined in the following section.
  Natural Soil Acidification. Natural  processes in ter-
restrial  ecosystems result  in the release and con-
sumption of hydrogen ions. Acids are formed naturally
in soils through the decomposition of organic matter,
uptake of cations,  and production of carbonic and
organic  acid  compounds.   In   addition,  land  use
changes (conversion of agricultural  land to forests)
can  result  in soils becoming more acidic over the
years as more organic  matter  accumulates. Allega-
tions have  recently been made that  this natural pro-
duction  of  hydrogen  ions  far   exceeds  those
associated with acid precipitation, consequently re-
cent acidification of streams, rivers, and lakes is caus-
ed by natural processes such as the aging of forests
(Krug and Frink, 1983; Marcus et al. 1983). This  allega-
tion provides an excellent example of how uncertainty
can be raised in science.
  Rosenquist (1977) first popularized the notion that
recent acidification  of surface waters in Scandanavia
was  being caused by natural acidity rather than acid
precipitation.  He used laboratory leaching  experi-
ments on small portions of soil profiles to simulate
the equivalent  of  150  years of  precipitation. The
results  of  this  laboratory work over a several day
period as well as other laboratory studies involving the
stirring of small amounts of soil with solutions were
taken by the author as proof that natural exchange-
able protons were present in the soil and were respon-
sible for surface water acidification.
  Drablos et al. (1980) did not rely on  questionable
laboratory work but rather utilized field  studies to in-
vestigate the allegations about natural acidity follow-
ing changes in land use in Norway. They correlated
historic land uses in  watersheds with reductions of
fish populations in sensitive lakes and found no effect
of land  use on fish population reductions. Acidifica-
tion seemed to proceed despite land use history.  !n
fact, the final report of the Norwegian SNSF project
acknowledged that while some natural soil acidity cer-
tainly is present, there was no doubt that the change
in composition of precipitation has played an impor-
tant role in acidification of freshwater (Overrein et al.
1980). Likewise in Sweden, research indicated that
while soils did become more acid following aging over
thousands of years, acidification of waters  does not
necessarily occur in the short term because naturally
produced organic acids remain largely in the soil due
to their complex molecular structure and  insolubility
(Swedish Ministry of Agriculture, 1982). Consequently,
natural acidification of soils in Scandanavia was con-
cluded to be less important in recent acidification of
surface waters than inputs of acid precipitation.
  In North America, the fierce debate concerning acid
precipitation has resulted in the resurrection of  these
allegations  and has created  much uncertainty. Van
Miegroet and Cole (1982) computed  the  amount of
hydrogen ions that theoretically would be generated in
forests to show that  proton production was greater
than proton  input from precipitation. Marcus  et al.
(1983) and Krug and  Frink (1983) also  refer to such
computations of proton production to make their point
that natural acidity is more important than  acid pre-
cipitation in  causing  recent acidification  of surface
waters.
  This natural soil acidity is said to have affected sur-
face waters in the Northeast only recently  because of
the  recovery  of  forests—especially coniferous
species—during the 20th century and natural acidity
from the aging of forest so'ils. Of course, objections
may be  raised regarding the  field applicability of
laboratory  tests on disturbed  soils,  the  validity of
assumptions concerning theoretical computations of
proton generation,  and  the relative significance of
natural  acidification   in the  Northeast when  acid
Figure  1.—While  much  attention has been  focused on
glaciated areas of the north, there is much uncertainty about
acid precipitation effects in the sensitive ecosystems of the
southern Appalachian mountains.
                                                  369

-------
 LAKE AND RESERVOIR MANAGEMENT
  precipitation has apparently been occurring there for
  the last 100 years (Hanson et al.  1982).
    To deal with such  uncertainties that are raised in
  science, decisionmakers should first examine existing
  data to determine whether facts are consistent with
  the alternative theories that are raised. If the data aie
  not consistent with the allegations, less attention can
  be paid to them. Table 1 contains a compilation of ex-
  isting data from the literature regarding pH of water in
  coniferous  forests.  It presents a  comparison  of
  average pH of water in forest ecosystems that are sub-
  ject to high  amounts  of acid precipitation with those
  subject to low amounts of acid precipitation. Note the
  higher pH values for rainfall, throughfall, and soil solu-
  tion water in the old  coniferous  forests that receive
  "unpolluted" precipitation. The soil water solution is
  naturally somewhat  acid in  the undisturbed, old
  growth forests of Alaska and Oregon, but compare the
  pH with much lower values in soil solutions of con-
  iferous forests that receive high amounts of acid pre-
 cipitation. In fact, the B horizon  soil solution in the
 Adirondacks   is  one hundred  times  more  acid
 (hydrogen ion concentration) than natural  acidity of
 the Oregon old growth forest, and the A horizon (top-
 soil) soil solution  is almost one thousand times mors
 acid  than the naturally  acidic soil  solution  in ths
 Oregon forest. These data are not consistent with ths
 allegations  that  natural  soil  acidity  is  primaril/
 responsible for acidification of surface waters. How-
 ever, it is clear from the Alaska and Oregon data that a
 portion of soil water acidity—albeit small—is natural
 in origin.
Figure 2.—Widespread dieback of Frasier Fir trees has oc-
curred  in  high  elevation  forests  of  the  southern  Ap-
palachians during the last decade.  Has acid precipitation
contributed to the death of these sensitive forests?
                                 Other pertinent data come from Canada. Over the
                               last 40 years, the average pH of surface waters in the
                               southern part of the large Laurentides Park (this por-
                               tion  receives higher  loadings of  acid precipitation
                               than northern parts of the park) has decreased signifi-
                               cantly—up to one pH unit; stream pH has not decreas-
                               ed in the northern  portion  of the  park, which has
                               similar vegetation as the southern portion of the park
                               (Jones et al. 1980).  In Alberta, Baker et al. (1977) re-
                               ported on acidity in spruce forests and forest soils
                               receiving high  amounts of acid  precipitation  and
                               similar spruce forests receiving low amounts of acid
                               precipitation. The investigators found  much greater
                               acidity—up to two pH units in spruce forests receiving
                               higher amounts of acid precipitation. Existing data, al-
                               though circumstantial in nature, are not consistent
                               with the allegation that natural soil acidity is responsi-
                               ble for recent acidification of surface waters.
                                 Liming Strategies. If  a decisionmaker chooses a
                               highly uncertain course  of action substantial  risks
                               may accompany that decision. An instructive example
                               may be the choice of a liming program to address the
                               acid precipitation issue. An excellent compilation of
                               information  on liming  strategies  was   recently
                               prepared  by Fraser and  Britt  (1982).  In addition, a
                               review of liming  programs is included in Marcus et al.
                               (1983), and other papers at this symposium discuss re-
                               cent  findings   from  lake  liming  in   Scandanavia
                               Canada, and the United States.  While  there  are
                               notable success stories such as the Swedish lake lim-
                               ing program, lake liming may not work as  predicted,
                               and it remains an extremely complex undertaking to
                               prescribe appropriate applications.
                                Lake liming may be toxic to fish as  illustrated  by
                               Bengtsson (1980) in Sweden. In the Sudbury lakes of
                              Canada, liming  has  improved pH but toxic levels of
                              metals  persist  and  stocked  fish have not survived
                              (Marcus et al. 1983). The Swedish government does
                              not consider liming to be an effective countermeasure
                              to solve acidification problems but rather a temporary
                              defense to reduce risks to valuable aquatic resources
                              in some lakes until emissions reductions are imple-
                              mented (Swedish Ministry Agric., 1982). The Swedes
                              are concerned about surges of acidity and aluminum
                              following snowmelt and high rainfall. These episodic
                              surges may kill fish following reintroduction into lim-
                              ed lakes and they may mobilize toxic metals that were
                              previously  precipitated  by  liming.  Such   episodic
                              events have also been  reported in North America, in-
                              cluding the Great Smoky Mountains (Jones et al. 1983).
                              It appears that lake liming has risks associated with it.
                              Repeated applications may be necessary, loss of rein-
                              troduced aquatic life  may occur in some lakes, many
                              inaccessible lakes and those susceptible  to  large
                          Table 1.—Average pH of water in coniferous forest ecosystems.
  Parameter
Low Acid Precipitation Areas
 Oregon^        Alaska2
  Source:
  ' Sollins et al. (1980)
  '. Johnson (1981)
  !. Cronan and Schofield (1979)
  '. Monitor and Raynal (1982)
  '. Jones et al (1983)
            High Acid Precipitation Areas
New Hampshire*      New York*
Rainfall
Throughfall
Litter solution
A Horizon solution
B Horizon solution
5.3
5.3
5.7
6.5
6.3
5.5
5.3
4.7
5.4
6.0
4.2
4.0
—
4.0
—
4.2
4.2
3.7
3.6
4.3
4.3
4.6
4.1
4.3
4.9
                                                 370

-------
                                                                                   ACIDIC PRECIPITATION
hydraulic surges of acidity may not be effectively re-
stored, and adverse ecological changes may occur.
  To combat episodic surges of acidity and alum-
inum, liming of streams, rivers, and terrestrial areas
may be needed. Watershed liming would seem to be
prohibitively expensive and stream liming seems to be
less cost effective than lake liming (Fraser and Britt,
1982). Additional risks are raised about the success of
watershed liming. Terrestrial liming has been reported
to reduce the growth of trees,  kill vegetation, and
adversely affect soil organisms (Tamm, 1976; Swedish
Ministry Agric., 1982). It is clear that liming strategies
have uncertainties associated with them and risks
would be present if widespread liming were chosen as
the primary means for combatting acidification. While
some lakes would be restored and others would face
less risk of acidification, many lakes, rivers, streams
and terrestrial areas  would continue to be at risk
because liming strategies are not feasible for them.


UNCERTAINTY IN  THE SOUTHERN
APPALACHIANS

While much of the concern about effects of acid preci-
pitation has  centered on the glaciated regions of
North  America,  little  attention has focused on the
potentially sensitive  areas  of  the  Southern Ap-
palachians.  Extending from  Maryland  to  Georgia,
these mountains are characterized by thin, sensitive
soils and fragile trout and  smallmouth bass streams.
Sulfur emissions have increased more in the South-
east the last decade than elsewhere on the North
American  continent. In fact, wet deposition of sulfate
and hydrogen ions is almost 50 percent  higher than
upstate New York or  New England and precipitation
pH averages  about 4.3 in the southern Appalachians
(U.S./Can.  Work Group 1, 1983).
  The natural blue haze of this  mountainous region
has been replaced by a white  haze dominated by acid
sulfates (Stevens et al. 1980). This man-induced haze
causes substantial  loss of visibility from the Great
Smoky Mountains to the Blue Ridge and Shenandoah
Valley. Research has demonstrated that this acid sul-
fate haze has as its origin anthropogenic emissions of
sulfur dioxide in the Midwest (Ferman et al. 1981).
   A recent TVA  analysis  of  historical water quality
data in these sensitive watersheds of the TVA region
found a trend of decreasing pH and alkalinity at most
stations (Meinert and Miller,  1982). In some of these
mountain streams, pH levels  are depressed almost 2
pH units following heavy inputs of precipitation, and
high concentrations of aluminum accompany these
surges of  acidity (Jones et al.  1983). The National Park
Service reports  that  fishkills  occur  following  heavy
rainfalls in at least seven fish hatcheries in the vicinity
of the Great Smoky Mountains (Mathews and Phillips,
1982).  An unusually high  frequency  of spinal defor-
mities has been noted in smallmouth bass (almost 20
percent of the fish) in one high-elevation TVA reservoir
and similar deformities have been reported in other
sensitive  TVA  reservoirs.  Beamish et  al.  (1975)
reported such deformities to  be symptomatic of re-
cent acidification in  Canada by acid precipitation.
Chemical toxicants  may  also  be  responsible for
skeletal deformities.
   Uncertainity  concerning adverse  impacts is  not
limited to aquatic resources  in the  southern  Ap-
palachians. Tens of thousands of acres of fir trees are
becoming infested with insects  and  are dying in the
higher elevations. While the balsam wooly aphid has
been blamed for killing the trees, recent research by
scientists from Oak Ridge National  Laboratory has
found that the trees have experienced  unexplained
decreases in growth up to two decades before the re-
cent infestations began (Foster, 1983). This finding
raises speculation that acid precipitation (including
dry deposition of pollutants) may act synergistically
with other factors to weaken the unique fir trees so
that they are  susceptible to insect  attack.  If such
uncertainties can be resolved by science, it will only
be after many years of carefully designed research. In
fact, it may be impossible for cause  and effect rela-
tionships to be determined with reasonable certainty.
Meanwhile, decisionmakers  must  weight  the risks
associated with  potential  irreversible  damage to
sensitive ecosystems.
PRUDENCE AND EQUITY

Air resources,  fisheries, water quality, and  public
lands are referred to as public goods because they are
held in  trust and  managed by government for the
public's sustained  use. Because some would exploit
or degrade such resources to keep production costs
of a product low, it is the role of government to protect
these vital public resources and the interests of future
generations in using them.
   Decisionmaking in the public arena is difficult when
the consequences and risks of alternative actions are
not fully  known  and  economic  costs for control
measures seem significant. It is made more difficult
when advocates confuse the public and decision-
makers  with misinformation and inaccurate informa-
tion. Unfortunately, science is not capable of resolving
all such uncertainties in the short  run. Consequently,
ethical and social considerations  involving prudence
and equity become more dominant  in choosing a
course of action to protect the public interest.
   Prudence refers  to the use of judgment—or being
overly cautious—in protecting one's interest. When
there is  risk of irreversible loss of public goods or risk
of damage too costly to  mitigate, prudence dictates
that action be taken in short order to  reduce these
risks to  allow for a margin of error in protecting sen-
sitive resources. Such action is needed because costs
and benefits of technological  advances are not uni-
formly distributed over society. One region may reap
benefits and another  region  may bear the social,
economic, and environmental costs without sharing in
the benefits. It is the role of government to ensure that
the external costs and risks are minimized  by achiev-
ing the  incorporation of  pollution control costs  into
the cost of production of a product. Equity considera-
tions dictate that those who impose risks  and costs
on others should be responsible for eliminating such
externalities borne involuntarily by others.
   When the Tennessee Valley Authority was estab-
lished in 1933, Congress charged  it with promoting
proper use, conservation, and unified development of
the Tennessee Valley's natural resources to improve
the standard of living for Valley residents. As the Na-
tion's largest producer of electrical energy as well as a
Federal  instrumentality  with  responsibility  for the
wise use of the region's natural  resources, TVA  is
vitally concerned with  the  adverse impacts of acid
precipitation on sensitive ecosystems.
   In March 1982,  TVA issued an  acid precipitation
policy statement that has proven to be very contro-
versial on the national level. In preparing the state-
ment, TVA realized that many years will pass before
                                                 371

-------
  LAKE AND RESERVOIR MANAGEMENT
  definitive scientific answers  are  developed to de^al
  with  uncertainties and  that  the  Tennessee Vail ay
  region and the Nation cannot afford the risk of perrr a-
  nent damage to sensitive ecosystems if impacts are ir-
  reversible or too costly to mitigate. Based on a con-
  sideration of facts, prudence,  and  equity, TVA issued
  a policy statement recognizing the  likelihood of a rela-
  tionship  between acid precipitation and the total load
  of SO2 and NOX in the atmosphere. TVA also recogn z-
  ed  that  long-range transport  and  transformation  of
  pollutants from fossil-fuel boilers produce sulfate arid
  nitrate particles that are linked to  acid precipitation,
  and it called  for further reductions in emissions  of
  these precursors. With the acidification occurring  in
  glaciated regions of North  America  and with the
  uncertainty surrounding  adverse impacts  in the
  southern Appalachians, such a position supporting
  equitable emissions reductions  is certainly prudert,
  especially when the public is  faced with possible  r-
  reversible damage to  its sensitive  environmental
  resources.
 CONCLUSIONS

 Decisionmaking in  today's technological society  is
 difficult when the consequences and risks of alter-
 native actions are uncertain. This uncertainty can be
 legitimate or  it can  be artifically  created through
 misinformation and selectively  used science.  The
 misuse of science in policy controversies is not just a
 result of disagreements among competent scientists
 but an  attempt to support special interests. As was
 pointed out by the Comptroller General of the United
 States (1981), such misinformation has confused the
 acid precipitation debate and decisionmakers must be
 cognizant of it. When risks and uncertainty cannot be
 reduced, public policy decisions become difficult to
 make and no-action alternatives or calls  for more
 research often become attractive options to choose.
 While some calls for research  may  be legitimate,
 others may simply be made to justify continued finan-
 cial support for research staffs, to delay decisions on
 action,  or to discredit one position  so that another
 may appear more  favorable.
   The example of allegations of natural soil acidit/
 being responsible for  recent  acidification of surface
 waters  was raised to  illustrate that decisionmakers
 must seek out all facts—not just some selectively pre-
 sented facts—when they evaluate the significance of
 uncertainties that are raised. Existing data on soil
 solution water  in  coniferous  forests receiving high
 amounts of acid  precipitation compared to similar
 data in old growth  forests receiving low acid precipita-
 tion demonstrate that  soil water  can be  much more
 acid—up to  1000 times more acid—in areas receiving
 high amounts of acid precipitation compared to areas
 receiving low acid precipitation. While it is clear that
 there are natural contributions to soil  water acidity,
 existing data are not consistent with the recent allega-
 tions regarding natural soil  acidity  being primarily
 responsible for acidification of lakes  and streams.
   If liming strategies are chosen  to address the acid
 precipitation   issue,  substantial  risks  may  bo
 associated with that decision. There is a risk that lake
 liming effectiveness  may be short-lived in many lakes;
and that adverse impacts may be  caused  in lake eco-
systems. It may not be feasible to  lime many lakes  be-
cause of practical, technical,  and cost constraints;
and stream, river, and terrestrial ecosystems would bei
resources remaining at risk because of  the lack ol
 Figure 3.—There will always be uncertainty present in deal-
 ing with complex technological problems. When there is risk
 of permanent  damage to sensitive ecosystems, prudence
 dictates  that action be  taken to protect these sensitive
 public resources.
 feasible  liming strategies. The initiation of experi-
 mental liming programs such as the Swedish model
 may be prudent to provide a temporary defense for
 some lakes until emissions are reduced, but such a
 program  can  be aimed at only a small percentage of
 resources at risk.
   No matter  how  much information is  available on
 practically any issue, questions of scientific uncer-
 tainty in cause and effect relationships can be raised
 by scientists  representing opposing views. Decision-
 makers must  understand that there will always be in-
 formation gaps and uncertainty present  in  dealing
 with  complex technological  problems  because
 resources for data collection  are limited, because
 man's ability to understand highly variable  natural
 processes is limited, and because of an inherent lack
 of certainty associated with the scientific method in
 the  short run. On issues of great public  significance
 such as acid  precipitation, where there  is risk of ir-
 reversible damage to sensitive  public  resources,
 decisionmakers cannot afford to wait until definitive
 answers to all possible uncertainties  are researched.
 In fact, many questions regarding  cause and effect
 relationships  may take many years to answer—if in-
 deed they can ever be  answered with  certainty  by
 science. Consequently, ethical  and social considera-
 tions involving prudence and equity become more im-
 portant in choosing a course of action to protect sen-
 sitive  public  environmental  resources  and  the  in-
 terests of future  generations  in  using these vital
 resources.
REFERENCES

Baker, J., et al. 1977. Acidity of open and intercepted precipi-
  tation in forests and effects on forest soils. Water Air Soil
  Pollut. 7:449.

Beamish,  R.J., et al. 1975. Long-term acidification of a lake
  and resulting effects on fishes. Ambio. 4:98.
Bengtsson, B. 1980. Liming acid lakes in Sweden  Ambio
  9:34.

Comptroller General of the United States. 1981. The debate
  over  acid  precipitation:  opposing  views,  status  of
  research. EMD-81-131. U.S.Gen. Account.  Off Washing-
  ton, D.C.
                                                 372

-------
                                                                                              ACIDIC PRECIPITATION
Cowling, E.B. 1982. Acid precipitation: historical perspective.
  Environ. Sci. Technol. 16:110A.
Cronan, C.S. and C.L.  Schofield.  1979. Aluminum leaching
  response  to acid precipitation: effects on high elevation
  watersheds in the Northeast. Science. 204:304.
Davis,  R.B.,  et al. 1980. Atmospheric deposition  in Norway
  during the last 300 years as recorded in SNSF lake sedi-
  ments. Pages 274-275 in Proc. Int. Conf. Ecol. Impact Acid
  Precip. SNSF Project, Norway.
Drablos, D.,  et al. 1980. Historical land-use changes related
  to fish status  development in different areas in southern
  Norway.  Pages 367-369 in  Proc.  Int. Conf. Ecol.  Impact
  Acid Precip. SNSF Project.
Farmer, G.J., et al. 1980. Some effects of the acidification of
  Atlantic salmon  rivers in Nova Scotia. Can. Tech. Rep.
  Fish. Aquat. Sci.  No. 972.
Ferman, M.A., et al. 1981. The nature and sources of haze in
  the Shenandoah Valley/Blue Ridge Mountains area. J. Am.
  Pollut. Control Assn. 31: 1074.
Foster, C. 1983. Smokies firs, acid rain. The Oak Ridger. July
  30:1.
Fraser,  J., and  D. Britt. 1982.  A technical report on  liming
  acidified  waters. Report No.  1285-01-82-CR.  Gen. Res.
  Corp. McLean, Va.
Haines, T.A., and J.J. Akielaszek. 1983. Acidification of head-
  water lakes and streams in  New England. Pages 83-87 in
  Lake Restoration,  Protection, and  Management. EPA
  440/5-83-001.  U.S. Environ.  Protection Agency,  Washing-
  ton, D.C.
Hanson, D.W.,  et  al.  1982. Modern and  paleolimnological
  evidence for accelerated leaching and metal accumulation
  in soils in New England, caused  by atmospheric deposi-
  tion. Water Air Soil  Pollut. 18:227.
Johnson, D.W. 1981. The natural acidity of some unpolluted
  waters in southeastern Alaska and potential impacts of
  acid rain. Water Air Soil Polut. 16:243.
Jones, H.C., et al. 1983. Investigations of cause of fishkills in
  fish rearing  facilities  in Raven  Fork  waterhsed. Tenn.
  Valley Auth. Knoxville, Tenn.
Jones, H.G., et al. 1980. The evolution of acidity in surface
  waters of Laurentides Park (Quebec, Canada) over a period
  of 40 years. Pages 226-227 in Proc. Int. Conf. Ecol. Impact
  Acid Precip. SNSF proejct.
Krug, E.C., and C.R. Frink. 1983. Acid rain on acid soil: a new
  perspective. Science. 221:520.
Marcus, M.D., et al. 1983. An assessment of the relationship
  among acidifying depositions, surface water acidification,
  and fish populations in North America. Vol. 1. Final Rep.
  EA-3127. Electric Power Res. Inst., Palo Alto, Calif.
Mathews, R.C., and  C.A. Phillips. 1982.  Survey of fishery
  losses in  hatcheries  near Great Smoky Mountains Na-
  tional  Park and relation  to acid precipitation. Nat. Park
  Serv. U.S.  Dep.  Inter., Gatlinburg, Tenn.
Minert, D.C., and  F.A. Miller. 1982. A review of water quality
  data in acid sensitive watersheds in the Tennessee Valley.
  Tenn. Valley Auth., Chattanooga, Tenn.

Mollitor,  A.V., and D.J. Raynal. 1982. Acid precipitation and
  ionic movements in Adirondack forest soils. Soil Sci. Soc.
  Am. J. 46:137.
Overrein, L.N., et al. 1980. Acid precipitation-effects on forest
  and fish. Final  Rep. SNSF Project, Norway.
Rosenquist,  I. Th. 1977. Alternative sources for acidification
  of river water in Norway.  Sci. Total Environ. 10:39.
Schofield, C.L. 1976. Acid  precipitation: effects  on fish.
  Ambio 5:228.
Sollins, P., et al. 1980. The internal element cycles of an old-
  growth Douglas Fir ecosystem in western Oregon. Ecol.
  Mono. 50:261.
Stevens, R.K., et al. 1980. Characterization of the aerosol in
  the  Great  Smoky Mountains.  Environ. Sci.  Technol.
  14:1491.
Swedish Ministry of Agriculture.  1982. Acidification today
  and tomorrow.
Tamm, C.0.1976.  Acid precipitation: biological effects in soil
  and on forest vegetation. Ambio 5:235.
U.S./Canada  Work Group  1.  1983. Impact assessment-
  final report. Memorandum of Intent on Transboundary Air
  Pollution.
Van Miegroet, H., and D.W. Cole. 1982. Effects of acid rain
  on the soil nutrient status and solution acidity. Pages 1-18
  in Air Quality Protection Aspects of Forestry Management.
  Tech. Bull. 390. NCASI, N.Y.
                                                        373

-------
  STUDIES ON THE USE OF LIMESTONE TO RESTORE
  ATLANTIC SALMON  HABITAT IN ACIDIFIED RIVERS
  W.  D. WATT
  G. J. FARMER
  W.  J. WHITE
  Fisheries Research Branch
  Halifax, Nova Scotia


             ABSTRACT
             Liming experiments were conducted using two approaches: instream placements of limestone gravel
             and headwater lake neutralization with limestone powder. Instream gravel was  ineffective at low
             temperatures and high flows, and required sit least 100 tonnes for each m^.s"1 of flow to achieve
             biologically useful pH increases. The Atlantic salmon showed positive responses, but only in the im-
             mediate vicinity of the limestone. Headwater lake liming produced high neutralization efficiencies
             but, there were rapid fluctuations in river pH's, which resulted from the natural stratification and mix-
             ing cycles within the lakes, and from rapid flushing due to high runoff volumes. Inverse stratification
             under winter ice occasionally produced acidic surface layers and dramatic lowering of downstream
             pH's. Positive responses were obtained from salmon in the lake's outlet stream. The problems of en-
             vironmental chemistry and logistics seriously limit the practicability of using limestone to mitigate the
             acidification of Atlantic salmon rivers.
 INTRODUCTION

 Acidification has reduced Atlantic salmon habitat in
 Nova  Scotia by one  third during the past 30 years
 Watt (1981) and Watt et al. (1983) have reported the ex-
 tirpation of salmon runs to seven Nova Scotian rivers
 where the average pH's are now below 4.7 and reduc-
 tions  in the  salmon runs to  nine other rivers where
 pH's now range 4.7 to 5.0. Salmon  stocks show no
 acidity related  signs of decline in Nova Scotian rivers
 of pH  above 5.1.
   One of the principal characteristics of the Atlantic
 salmon is that each river tends to have its own distinc-
 tive stock or stocks. Thus, loss of a river's native stock
 is irreversible. It may be possible to restock the rivers
 by using donor salmon stocks from  other areas, but
 the  general  experience  in salmon  restoration  pro-
 grams with nonnative stocks  has been that restora-
 tion is very expensive, it takes many years, and the
 success rate is low if the donor stream is not  nearby.
   In an effort to preserve the nuclei of remnant runs in
 those rivers where average pH's are presently between
 4.7 and 5.0, we have then investigated the feasibility of
 establishing de-acidified refugia  in tributaries  so that
 the genetics of the native stocks will not be entirely
 lost. These nuclei stocks, if successfully preserved in
 refugia, will then be used to restock the rivers after the
 acid rain problem has been corrected.
  The  potential for taking the remnant stocks into our
 fish  culture system for preservation has  been  con-
 sidered. The objection to this approach is that environ-
 mental and genetic selection  in fish  culture stations;
 tends to be quite  rigorous. We  are  concerned that
 after several generations in a fish culture environment
 we would no longer have the wild stock. We have also
 considered cryogenics as a possible solution. The dif-
 ficulty  here is  that to establish the  feasibility  o':
 preserving salmon eggs and sperm by freezing for 20
 to 50 years, it would be necessary to  do it first to ac-
cess the actual survival. This may be a feasible solu-
tion, but it is not yet a proven  technology.
   The difficulty involved in controlling the pH of an
 Atlantic salmon river is impressive. The annual flow
 variation is two orders of magnitude, from a minimum
 flow which usually occurs in late summer to maximum
 flow which usually occurs in  late winter. The annual
 range of variation in  pH amounts to a variation in
 hydrogen  ion concentration of one order of magni-
 tude,  which unfortunately is an amplification of the
 flow variation  since pH and  flow  are inversely cor-
 related (Fig. 1), the lowest pH's occurring during max-
 imum flows and vice versa. So, in the Atlantic salmon
 rivers of Nova Scotia we have to deal with an annual
 variation in hydrogen ion discharge of three orders of
 magnitude. A successful liming approach must incor-
 porate this degree of responsiveness.
   Protection against episodic pH depressions is also
 of concern in Nova Scotian rivers. On two or three oc-
 casions each year, warm mid-winter depressions pro-
  pH 5 5  —
             GOLD RIVER DISCHHRGE  (Log  m Vs)
Figure 1.—Relationship between pH and flow for Gold River,
N.S., during 1980-1982. The regression equation for pH on
log flow is pH =  5.83 - 0.63 (Iog10 m3 s~1), r2 = 0.82. The
Gold River is moderately acidic with a mean annual pH of 5.1.
                                                 374

-------
                                                                                    ACIDIC PRECIPITATION
duce heavy rainfalls on frozen soils, causing rapid sur-
face runoff. The runoff pH may actually be lower than
that of  the  precipitation  if  partial  freezing  con-
centrates the acid. In lakes under ice cover, mixing is
minimal and an inverse thermal stratification prevails.
The runoff  water enters the streams and  lakes near
0°C and it spreads in a surface layer 0.5 to 1.5 meters
thick immediately under the ice. This phenomenon is
depicted (Fig. 2) for Eastern Lake, N.S.,  for Jan. 8,
1981, after  a  32  mm rainfall of pH  4.7 on Jan. 7. The
runoff was  partially frozen by contact with snow and
frozen ground and by the low overnight temperatures,
resulting in a pH of the runoff under the ice in the top
meter  of Eastern  Lake  that  is below that of the
precipitation. Unfortunately, it is this low-pH surface
water that  is discharged to the rivers by the natural
surface outlets.
  Three methods of applying limestone were initially
considered, based on personal communications and
review of previously published accounts of experience
in Sweden (Grahn and  Hultberg,  1975),  the United
States (Pearson  and  McDonnell, 1975) and  Canada
(Schieder et al. 1975; Dillon et al. 1977): use of a silo to
lime the river as it flows past the site, use of crushed
limestone gravel in the river bed, and use of powdered
limestone to neutralize headwater lakes. Possible use
of various  other mechanical devices and wells was
tentatively  rejected because of the high probability of
their being jammed or blocked by ice during the four
months of  freezing conditions typical  of  most  Nova
Scotian winters. The maximum limestone requirement
occurs  during the winter months when  discharges
tend to be  high and pH is lowest.
  The  silo  concept was  initially rejected because of
the difficulty of  designing a dosing apparatus with a
three  order-of-magnitude  response  range,   and
because of the potential  breakdowns which would be
unavoidable  in a wilderness situation where winter
power failures of one  to four days duration are not in-
frequent. Possible failure  of  the  liming  apparatus,
especially  during the low pH winter  period, would
almost certainly result in a major fishkill.
 LIMESTONE GRAVEL IN RIVERS

 De-acidifying rivers by adding limestone gravel to the
 stream bed has promise (Pearson  and McDonnell,
 1975). This technique has been tried on a small scale
 (100-300 tonnes) at six sites in Nova Scotia. Average
 pH's at all six sites are within the range 4.8-5.0. The
 sites are monitored for chemical  effectiveness in pH
 control and for biological effectiveness by survival of
 juvenile Atlantic salmon. Thus far, the pH increase in
 the streams has varied from 0.0 when discharge was
 maximum during the  winter to +1.6 pH units during
 summer low flow conditions.
   Under winter temperatures and  high flow rates, very
 large quantities of limestone  gravel are required  to
 significantly increase the pH of acidified rivers. Data
 from the liming sites have been used to derive a multi-
 ple regression relating the pH increase below the lim-
 ing sites to water  discharge, water temperature, and
 the amount of limestone gravel  which has been ap-
 plied.  Figure  3 shows   the results  of  the  multiple
 regression analysis on 2  years of data from the liming
 sites. Data collected from the sites during the second
 year after placement of the gravel did not  differ
 significantly from the first year's data. The regression
 uses information on  an  acidified river's present pH
 level, discharge, and temperature to predict the quan-
tities of limestone gravel required to achieve signifi-
cant pH alteration.
  Farmer et al. (1980) have shown that the most sen-
sitive phase of the Atlantic salmon's life cycle is the
early feeding stage that occurs just after the fish have
absorbed the yolk sac and are swimming up out of the
gravel to commence  the fry  stage. At this swim-up
stage, the LD50 for acid conditions is quite close to pH
5.0. In Nova Scotian rivers, the swim-up stage is usual-
ly reached in May. To protect the most sensitive phase
of the life cycle enough limestone must be used to en-
sure a river pH above 5.2 during the month of May.
Temperatures  in the river during  May would normally
be about 10°C; about 120 tonnes of limestone gravel
per m3s-1 of flow would increase pH by about 0.4 dur-
ing May in a river presently near pH 4.8 (borderline for
salmon survival). To do a small Nova Scotian  salmon
river such as the East River, Lunenburg County, which
    0    1    2    3   <
     TEMPERRTURE  C °C)
 5
PH
Figure 2.—A midwinter low pH episode in the surface water
under the ice of Eastern Lake, N.S., on Jan. 8,1981, after a 32
mm  rainfall  of  pH  4.7 on  Jan.  7. The  maximum air
temperature on Jan. 7 was +6°, the overnight low was -10°
and on Jan. 8 the maximum was  +3°. Prior to the rainfall
event temperature had been below 0°C for at least 7 days,
and  there  were  21  cm of snow on  the ground  from a
snowstorm of pH 5.6 on Jan. 2-4.
Ap-J
            100      200      300

            LIMESTONE DOSE (tonnes.m~
Figure  3.—A  multiple  regression equation  was derived
relating the data on pH increase (ApH) in river water passing
over limestone gravel placements to stream discharge (m3
s~1), water temperature (t, in °C), and the limestone dose in
metric tonnes. The equation is: ApH =  0.237 (loge DOSE) +
0.008 (t) -0.809; and r2 is 0.84.
                                                  375

-------
 LAKE AND RESERVOIR MANAGEMENT
 has  a total Atlantic salmon  potential of  about 200
 adults per year and a 1 in 10-year maximum daily flow
 in May of approximately 21 m3s -1 would require 2,500
 tonnes of gravel.
   Approximately 15 percent of the limestone  gravel
 will be dissolved each year by the acid waters in the
 river, and approximately 10 percent will be deactivated
 by being buried or otherwise physically  displaced.
 Hence, maintaining this river will require approximate-
 ly 25 percent or 600 tonnes of limestone gravel each
 year. Cost per tonne of gravel, including  trucking and
 spreading  in the stream bed,  has varied from $20 lo
 $40.  Trucking costs represent the greatest  expen-
 diture  for  these projects. Theoretical cost calcula-
 tions, using tonnages derived from the regression of
 Figure 3, indicate that this method may be cost effec-
 tive  for those rivers within about  100 kilometers of
 limestone sources  and with present mean annual pH
 levels above 4.7.
   Difficulties  encountered with  using  limestone
 gravel for pH control are:
   1.  Tendency of the  gravel to be coated  and deac-
 tivated by chemical or biological coatings;
   2.  If placed in too low an energy area,  there is a
 tendency for the gravel to become covered by sedi-
 mentary deposits; and
   3.  If placed  in too high  an  energy area,  there is a
 tendency for the gravel to be washed downstream into
 gravel banks, and so effectively deactivated.
   We have  taken considerable care to choose gravel
 of a size comparable to the substrate already presert
 at the liming site. By matching the gravel sizes in this
 manner, it  is possible to  achieve  conditions  where
 enough movement  occurs to  prevent  the  limestone
 gravel from  becoming coated, but not sufficient to dis-
 place it into gravel banks. Since the material is placed
 in the vicinity of natural gravel deposits, there is no
 tendency for it to be covered by sediments. Although
 careful site selection  and gravel  matching is  quite
 possible on  a small scale, it may be impractical to do
 it  for an entire river, as for instance in placing  2,500
 tonnes in East River. In practice,  we shall probabl/
 have  to accept quite high rates of deactivation.
   The efficacy of limestone gravel in mitigating the ef-
 fects of acidification on Atlantic salmon  is indicated
 by the juvenile salmon densities found in waters up-
 stream, on,  and below liming sites in Liscomb  River,
 N.S.  (Table 1). Juvenile densities are significantl/
 higher (p< 0.001) in the immediate vicinity of the lime-
 stone gravel for both fry and parr (juvenile salmon in
 their  second year).
   For Liscomb River, the 1 in 10-year maximum dail/
 flow in the month of May is 123 m3s-i; hence about
 15,000 tonnes of gravel would be required to protect
 the entire  river. The  Atlantic salmon  potential of
 Liscomb River is about 2,000 adults per year.

 LAKE NEUTRALIZATION

 An alternative  approach  to  the  problem  of de-
 acidifying Atlantic salmon habitat  is using powdered
                   limestone  to neutralize headwater lakes which then
                   release the treated water to the rivers. The dissolution
                   efficiency of powdered limestone is much higher than
                   gravel. The treatment method would partially correct
                   the increased acidity of higher flows by concomitant
                   higher releases from the neutralized lakes.
                     The capacity of river and  lake water to release
                   hydrogen ions is measured by the acidity; hence, in
                   neutralization  experiments we  have  expressed  the
                   limestone doses  in terms  of lake acidities. Severely
                   acidified (pH< 5.0) Nova Scotian lakes generally have
                   acidities in the 100 to 300 /^eq/1  range. A 1X acidity
                   limestone dose lor a lake with acidity of  150  ^eq/1
                   would be 7.5 grams of calcium carbonate equivalence
                   for every cubic meter of water in the lake. It is difficult
                   to relate dosages  based on acidities and lake volumes
                   to the doses per  unit area which are common in the
                   literature.
                    The dolomitic  limestone employed in  our initial
                   studies is finely ground with a mean particle diameter
                   of 30  micrometers as measured by a Coulter counter.
                  The limestone contains 17.1 percent calcium, 11.3 per-
                  cent magnesium,  and 89.8 percent calcium carbonate
                  equivalence. Figure 4 indicates the impact on pH of
                  adding a dosage of 2.7X acidity to Eastern Lake,  Nova
                  Scotia.  The theoretical curve is the  calculated  pH
                  decline resulting from dilution  and monitored acid in-
                  puts. The theoretical curve assumes that after the in-
                  itial neutralization and mixing (about 2 weeks) no fur-
                  ther limestone is  dissolved. As Figure 4 shows,  how-
                  ever, the excess limestone remains chemically active
                  in  inhibiting re-acidification,  especially during the
                  seasonal turnover times in spring (March-April) and
                  autumn (September-October). Nova Scotian  lakes are
                  subject to rapid flushing due to high rainfall (1,400 mm
                  per year), high rates of runoff (70 percent), and low
                  ratios  of lake  volume  to  drainage area. The mean
                  residence time of Eastern Lake is approximately 6
                  months. The excess neutralization capacity appears
                  to have been effective for about 1 year, or two mean
                  residence times.
                    An estimate of the proportion of the  limestone that
                  is initially dissolved in the process of raising the  lake-
                  water  pH was obtained by comparing the pH rise with
                  a preliming titration curve of the lake water.  After the
                  whole-lake  pH  maximum  is  reached,  significant
                  dissolution of the  remaining limestone can be follow-
                  ed  by  monitoring acid inputs (water volume, pH, and
                  acidity); and calculating theoretical pH changes from
                  dilution, from acidification by hydrogen ion inputs,
                  and from titration  of the runoff up to  the prevailing
                  lake pH.  Carbonate dissolution  is  then calculated
                  from the difference between theoretical and actual
                  pH's. These results can be expressed, as in  Figure 4,
                  as a curve of cumulative percent dissolution efficien-
                  cy. The initial percent dissolution of 25 percent occur-
                  red 2  weeks after liming,  when  the whole-lake pH
                  reached  its  maximum  value  of  6.5.  The   percent
                  dissolution  efficiency almost  doubled  during   the
                  following year, largely in two spurts during turnovers
                  in march-April and again in September. After the  first
Table 1.—Juvenile Atlantic salmon densities calculated from 1981-82 electrofishing data in Liscomb River, N.S. (mean pH 4 9)
One electrofishing site is 5 km upstream of the limed areas, three sites are in areas spread with limestone gravel (total of 180
                       tonnes), and three sites are > :tO km downstream of the limestone.
Electrofishing Sites
No. of Site Visits
                                                                  Fry/100 rr»2
                                                       Parr/100
Upstream of treatment
Within limed areas
> 30 km downstream
                                 7.1
                                 47.9
                                 6.0
1.1

3.7

0.4
                                                 376

-------
                                                                                     ACIDIC PRECIPITATION
year the rate of dissolution slowed, being only an addi-
tional 4 percent during the second year after liming.
   Using dolomitic limestone on two other lakes  has
yielded similar results (Table 2). At lower doses, the
percent dissolution efficiency is higher; but the pH in-
creases achieved were lower, subsequent pH declines
were faster, and  the  pH  peaks at the spring  and
autumn turnovers were less pronounced.
   We then began experiments with another form  of
limestone having much  lower magnesium  content.
This calcitic limestone is 36.0 percent calcium, 2.7 per-
cent magnesium, and  100 percent calcium carbonate
equivalance.  The grinding was similar  to that  of
dolomitic limestone, and the powder has a mean parti-
cle diameter of 32 micrometers.  Laboratory tests in-
dicated that this material was more effective in raising
the pH of acidified water and would have a higher per-
cent dissolution efficiency.
   Sandy  Lake was limed with  a 3.3X acidity dose  of
the  calcitic limestone. The percent dissolution  effi-
ciency after 1  year was 11 percent higher than that ob-
tained when dolomitic limestone was used in Eastern
Lake (Table 2), and pH control was substantially better
with pH's remaining above 5.2 for 16 months. The max-
imum  post-liming whole-lake pH  was 6.9, but whole-
lake pH was held above 6.0 only for about 3 months.
The percentage of the limestone dissolved at the time
of pH  maximum (Table 2) in Sandy Lake is similar  to
that for Eastern Lake, but the percent dissolution  effi-
ciencies  in Sandy are substantially higher than  in
Eastern after 1/2 and 1 year, in spite of Sandy's having
received a 20  percent higher dose.
   Sandy Lake's 24 percent initial dissolution efficien-
cy is lower than the 30 percent predicted by Sverdrup
and Bjerle (1983) for calcite of this mean particle  size
in a lake initially at pH 4.8. However, the observed 54
percent dissolution efficiency after 1 year is con-
siderably  higher than the predicted total long-term
dissolution efficiency of 44 percent. Sverdrup  and
Bjerle's model  for predicting  short-  and  long-term
dissolution  appears   more   appropriate   for  the
dolomitic  limestone  usage (Table  2)  than for  the
calcitic.
   In the Sandy Lake experiment, in  addition to  the
estimate  of percent  dissolution  efficiency derived
from  the  acid/base relations,  estimates were also
calculated from the calcium budget (inputs, outputs
and concentrations in the lake). Results of both  cal-
culations  are  compared  in  Figure 5. The  calcium
budget includes  both  dissolved  and  particulate
calcium  because the  samples were treated with  a
strongly acidic solution of lanthanum chloride prior to
analysis on an atomic absorption spectrophotometer.
Initially,  the calcium  budget  gives very high com-
parative values because of the high proportion of lime-
                    stone in particulate suspension.  Both methods  of
                    estimating percent dissolution efficiency tend toward
                    similar values after 3 months and continue more or
                    less  together reaching  approximately  60 percent
                    dissolution after 1 year. The persistently higher values
                    from the calcium budget  may indicate that a small
                    fraction of the limestone (about 7 percent) was flushed
                    out of the lake in particulate form.
                      Sandy Lake was chosen for limestone addition to
                    test the effectiveness of lake liming in rehabilitating
                    acidified Atlantic salmon habitat in the outlet stream.
                    Electrofishing in the stream and gill netting in the lake
                    failed to find any juvenile  salmon  prior to liming,
                    though  the  habitat  of  the stream was physically
                    suitable. A flow-through bioassay using water from
                    the lake (pH 5.0) yielded an LT50 for salmon fry of 2.5
                    days. Additional bioassays  conducted  at the same
                    time, using lake  water limed to pH's 6 and 7, showed
                    no significant fry mortalities.
                      The lake was  limed in August and in October the
                    outlet stream was  stocked with 2,000  9-month-old
                    salmon  parr. During the following  summer, electro-
                    fishing revealed that parr had survived in the outlet
                    stream.  In addition, there had been successful natural
                    spawning and a population of salmon fry was also pre-
                    sent. Sandy Lake is on the  Sackville River system, and
                    a remnant run of Atlantic salmon is known to persist
                    in the higher pH portions of this river system. Some of
                    these fish were attracted  into the Sandy Lake outlet
r\

nn-rr'-rn
— — THEORETICRL pH ~~
	 PERCENT DISSOLVED —
- \\
                    Figure 4.—Results of treating Eastern Lake with a powdered
                    dolomitic limestone dose of 2.7X acidity. The  theoretical
                    curve assumes that re-acidification proceeds by dilution and
                    acid inputs, without further  neutralization from the excess
                    limestone dose. The sedimented limestone contributed to
                    the lake's de-acidification during spring and autumn turn-
                    overs, thus giving two  post-liming pH peaks with very little
                    contribution thereafter. The percent dissolved curve is deriv-
                    ed from acid/base monitoring (see text).
 Table 2.—Liming doses and de-acidification responses from four Nova Scotian lakes. All four lakes have mean residence
      times of 4 to 6 months. The percent dissolution of the limestone is calculated from acid/base data (see text).

                                                       	     Lakes
                                  Big
                   Paterson
                   Eastern
                                                              Sandy
Limestone type
Dose (X acidity)
Dose (tonnes/ha)
Initial pH
Post-liming pH max.

% Dissolution
At pH max.
After 1/2 year
After 1 year
After 2 years
Dolomitic
0.7
0.2
4.6
5.3
Dolomitic
2.0
0.6
4.4
5.6
Dolomitic
2.7
0.4
4.6
6.5
Calcitic
3.3
1.7
4.8
6.9
32
41
60
62
19
30
50
50
25
30
43
47
24
38
54
                                                  377

-------
 LAKE AND RESERVOIR MANAGEMENT
 stream after the lake was limed (spawning is in Oc-
 tober-November), and the pH elevation was sufficient
 to ensure fry survival through  to the following sum-
 mer.
   The average total cost of powdered limestone acd-
 ed to four Nova Scotian lakes has been $87 per tonre.
 Routine operation on a large scale should permit bUk
 handling economies sufficient to reduce the overall
 costs to about $70 per tonne. The greater part of the
 cost (60 percent) is in the spreading.
   To provide this form of pH control  for East River
 (dosing  at 3X acidity) would require the adding 500
 tonnes of limestone per year to the lakes at a cost of
 approximately $35,000 per year. To apply the lake lim-
 ing  method to the much  larger Liscomb River system
 would require liming 20 lakes with a total of 2,200 ton-
 nes per year.

 COST COMPARISON

 The relative costs of using limestone gravel in the river
 and neutralizing headwater lakes with powdered lime-
 stone would be difficult  to compare precisely.  A
 realistic comparison of the two liming approaches for
 year-round pH control  would be misleading, because
 the  large quantities of limestone gravel that  wou d
 theoretically be  required  to maintain acceptably high
 pH's during the winter and early spring  periods, when
 pH and temperature are low and flows are high, render
 this  approach  impractical  under these conditions
 (East River would require 23,000 tonnes).
   The lake liming approach could, however, be scaled
 down to a 1X acidity dose applied in May so as to pro-
 vide  effective protection only for the  early feeding
 stage when salmon fry emerge from the gravel (May-
 June); this can be compared directly in terms of ton-
 nages and costs with the gravel approach. Such a
 comparison is presented in Table 3 for East  River.
 East River is a  convenient  site for the comparison
      SEP  OCT NOV DEC JRN TEB MRR  RPR MHY JUN JUL RUG SEP  OCT NOV
                        MONTH
Figure 5.—The percent of the calcium carbonate equivalenca
accounted for in Sandy Lake after liming with a 3.3X acidity
dose of powdered calcitic limestone. The acid/base curve
gives percent dissolved. The curve derived from the calcium
budget gives the  percent dissolved and/or  in particulate
suspension.
 because suitable access roads are already in place for
 either liming method. A requirement to construct new
 access roads could double the 20-year total costs for
 either method, but the relative costs of new road con-
 struction for lake liming or gravel placement may dif-
 fer greatly between river systems. For the comparison
 in Table 3 the lake liming approach has a definite cost
 advantage, about half as much as the gravel method.
   In spite of its higher costs, the limestone gravel ap-
 proach  has  an attractive  advantage  of superior
 reliability because, once in place, a failure to renew
 the gravel in  any  one year would not seriously pre-
 judice the survival potential of  a restored  salmon
 population. A failure  to lime the  lakes in any 1  year
 could  result in the loss of most of one and possibly
 two year classes.


 NEW STUDIES

 Consideration  is  now  being given  to using much
 higher limestone doses in the lakes (10X acidity) in an
 attempt to achieve a more sustained de-acidifying ef-
 fect, and  to liming lakes during the winter  season. A
 1983 attempt to pump a 2 percent suspension/solution
 of calcium hydroxide into Sandy Lake via  a diffuser
 under the ice was  a failure.  Inadequate mixing under
 the ice cover resulted in the accumulation of most of
 the high pH water  near the lake bottom.
  The  concept of liming from a silo is now being re-
 considered. If the silo  were sited above a lake of more
 than  a  1 month  mean  residence  time, then  a
 mechanical or power failure would not pose an  im-
 mediate threat to the  salmon habitat downstream of
 the lake, and so 100 percent reliability would not be re-
 quired.
  Liming  from a silo has been included for cost com-
 parison as a third possibility in Table 3. The  powdered
 limestone would be stored in an agricultural-type silo
 beside the river, and limestone would be fed con-
 tinuously to the river through a slurrying device. Since
 we are only dealing with the months of May and June
 for  this   comparison   Ihe  problem  of  ice  in  the
 mechanism need not be considered. Flows during the
 May-June period typically vary by less than  one order
 of magnitude, so it appears quite feasible  to design
 adequate  responsiveness into the delivery unit.  The
 silo system has the highest initial cost because of the
 capital equipment, but once operating the annual cost
 of the silo method would be the  lowest of the three.
 Hence,  on a  20  year  basis, the silo  method is
 economically competitive with headwater lake liming.


 CONCLUSIONS


 1. Instream limestone gravel can be used to raise the
 pH's of moderately acidified (mean pH 4.8-5) Atlantic
salmon rivers to acceptable levels during the sensitive
early feeding stage in  May-June.
 Table 3.—A comparison of required limestone quantities and total estimated costs for maintaining the pH of East River, N.S.
 above 5.1 during May-June of each year by placement of limestone gravel in the river, neutralization of headwater lakes with
                   limestone powder, or adding limestone powder directly to the river from a silo.


Limestone gravel
Lake liming
Liming from silo

Tonnes
2,500
170
120
First Year
SX103
75
17
100
Subsequent
Tonnes $
600
170
120
Years
X1Q3
20
12
10
Total
Tonnes
13,900
3,400
2,400
for 20 Years
SX1Q3
455
245
290
                                                 378

-------
                                                                                           ACIDIC PRECIPITATION
   2. Instream limestone  gravel cannot provide year-
round  pH control  because under  winter and early
spring conditions, the quantities of gravel that would
theoretically be required to achieve the necessary pH
elevation are impractically large.
   3. Headwater lake liming using a calcitic limestone
dose of 3X acidity can provide effective pH control for
approximately 1  year.
   4. For pH control during the  May-June period,
headwater lake liming would be a  more economical
approach than instream limestone gravel, but it would
also be  more vulnerable to interruptions in annual
treatments.
REFERENCES

Dillon,  P.J., N.D. Van, W.A. Scheider, and N. Conroy. 1977.
  Acidic  lakes  in  Ontario: characterization, extent, and
  responses to  base and nutrient  additions. Ont. Ministry
  Environ. Rep., Rexdale, Ontario.
Farmer, G.J., T.R. Goff, D.O. Ashfield, and H.S. Samant. 1980.
  Some effects of the acidification  of Atlantic salmon rivers
  in Nova Scotia. Can. Tech. Rep. Fish. Aquat. Sci. 972.
Grahn, O., and H. Hultberg. 1975. The neutralization capacity
  of 12  different lime products used for pH-adjustment of
  acid water. Vatten 2: 120-32.
Pearson, F.H., and A.J. McDonnell. 1975. Limestone barriers
  to neutralize acidic streams. Am. Soc. Civil Eng. J. Environ.
  Eng. Div. 101: 425-40.
Scheider, W., J.  Adamski, and M. Paylor. 1975. Reclamation
  of acidified lakes near Sudbury, Ontario. Ont. Ministry En-
  viron.  Rep., Rexdale, Ontario.
Sverdrup, H., and I. Bjerle. 1983. The calcite utilization effi-
  ciency and the long-term effect on alkalinity in several
  Swedish lake  liming projects. Vatten 39: 41-54.
Watt,  W.D.  1981. Present and potential effects of acid pre-
  cipitation  on  the Atlantic salmon in Eastern Canada. In
  Acid Rain and the Atlantic Salmon. Proc. Conf. Nov. 22-23,
  1980.  Int. Atlantic Salmon Found. Spec. Publ. Ser. 10:
  39-45.
Watt,  W.D., C.D. Scott, and W.J. White. 1983. Evidence of
  acidification of some Nova Scotian rivers and its impact
  on Atlantic salmon, Salmo salar. Can. J. Fish. Aquat. Sci.
  40: 462-73.
                                                     379

-------
 LAKE ACIDIFICATION AND THE BIOLOGY OF ADIRONDACK  LAKES-
 CRUSTACEAN ZOOPLANKTON COMMUNITIES
JAMES W. SUTHERLAND
SCOTT O. QUINN
JAY A. BLOOM FIELD
Bureau of Water Research
New York State Department of Environmental Conservation
Albany, New York

CLIFFORD A. SIEGFRIED
Biological Survey
New York State Museum & Science Service
Albany, New York
            ABSTRACT

            Fifty-five lakes in the Adirondack Mountains of New York were surveyed during 1982 for crusta-
            cean zooplankton, rotifers, phytoplankton, and water chemistry. The midsummer surface pH
            values ranged from 3.60 to 7.25. Lakes were selected in order to have a wide range of mor-
            phometry and watershed characteristics. Zooplankton species richness and diversity declined
            with pH. In lakes with pH less than 5.0, three species tended to dominate the growing season
            community. A discussion of the zooplankton communities in acidic lakes that have recently
            been treated with either agricultural limestone or sodium carbonate is included.
INTRODUCTION

The Adirondack Mountain Region of New York State
encompasses 2.4 million hectares including approx-
imately 2,900 ponded waters (114,000 ha surface area)
(Pfeiffer and  Festa, 1980) and 9,370 kilometers (6,700
ha surface area) of  significant fishing streams (Col-
quhoun et al. 1982). Based on surface water alkalinity,
this area is one of the largest regions in the eastern
United  States that  is  susceptible  to acidification
(Omernik and Powers,  1982), and the Adirondacks
receive substantal inputs of acidic deposition annual-
ly (Gibson and Linthurst, 1982).
Figure 1.—Map showing location of study area and waters
surveyed.
   A broad range of pH in Adirondack surface waters
(Pfeiffer and Festa, 1980) presents a  unique experi-
mental situation that can contribute significant infor-
mation  toward  the  understanding of  biotic  com-
munities and their functioning in acidified environ-
ments. A synoptic survey of ponded  waters  in the
Adirondack Mountain Region was initiated during the
summer of 1982 to assess the status of aquatic com-
munities in relation to water chemistry. The aquatic
communities surveyed included the phytoplankton,
planktonic rotifer,  crustacean zooplankton, and ben-
thos. This paper presents the results  of the crusta-
cean zooplankton  surveys in the Adirondack waters.
The results  of  the rotifer  community  surveys are
presented in Seigfried et al.  (1983).
   The  effect of acidification  on the  plankton  of
Adirondack ponded waters has received limited study.
The relationship between phytoplankton and crusta-
cean zooplankton species richness and pH of Adiron-
dack waters has been reported by Hendrey (1980) and
Confer et al.  (1983), respectively.
METHODS

The 55 ponded waters included in the synoptic survey
are within the Adirondack Park (Fig. 1). Selection was
based on a stratified random design, stratified accor-
ding to pH, elevation, size, accessibility, and historical
water quality information. The survey waters ranged in
pH from 3.60 to 7.25 and, in general, were small (<100
ha), unproductive (chlorophyll a < 5 ^g • I' 1) ponds
and lakes between  500 and 700 meters in elevation
(Fig. 2).
                                             380

-------
                                                                                     ACIDIC PRECIPITATION
   Fifty waters were surveyed  between July 26 and
Aug. 5, 1982, while the  remaining five waters were
surveyed  between  Aug. 31  and  Sept. 2, 1982. All
physiochemical and biological collections were made
at the  deepest portion of the survey water basin. The
crustacean zooplankton community was sampled by
triplicate vertical tows from near the bottom of each
water body to the surface using a No. 10 plankton net.
The plankton collected in the net for each tow were
concentrated in 50  ml  screwtop vials, narcotized with
carbonated  water,  and  preserved  in  a  formalin-
sucrose-rose bengal solution.
   pH  was  determined  on temperature-equilibrated
 duplicate  samples within 3  hours  of collection.
 Measurements were recorded to ± 0.02 pH units us-
 ing an Orion Research-lonanalyzer/Model 407A with
 an Orion Model 91-62 combination electrode for low
 ionic strength solutions. The electrode was calibrated
 before, during, and after each daily use with the two
 buffer technique. The  method of analysis for pH was
 CO2 - equilibration  as described by Schofield (1978).
   Crustacean  zooplankton   were  identified and
 counted in  triplicate aliquots,  adjusted to include a
 minimum of ~ 100 organisms of each of the dominant
 species. No net efficiency correction factor was used
 in the  calculation of crustacean zooplankton density.
 Sample counts were converted to densities (# 'm-3),
 standardized using natural logarithms and then com-
 pared   with  those  of every  other sample  using
 Sorensen's  (1948) index to derive a similarity matrix.
 Similarity values were then used in a Q-mode cluster
 analysis to  partition  samples into discrete  groups
 (Park,  1968), prior to Q-mode ordination analysis (Park,
 1968).
                    and  Bosmina  longirostris occurred  in almost  every
                    water body surveyed. Five other species, Mesocyclops
                    edax  (42   lakes),  Holopedium  gibberum  (36),
                    Diaphanosoma leuchtenbergianum (33), Tropocyclops
                    prasinus (30), and  Daphnia catawba  (28), occurred in
                    more than  half the waters surveyed. The number of
                    crustacean  zooplankton species generally  declined
                    with decreasing pH (Fig.  3).  Waters with pH > 5.50
                    generally had  seven or more species, while waters
                    with pH < 5.50 generally contained fewer than seven
                    species.  The greatest  number of species  in survey
                    waters was 14  from circumneutral  Upper  Saranac
                    Lake (Lake No. 55). Only one  species, Diaptomus
                    minutus, occurred  in Avalanche Lake (Lake No. 52).
                      Four crustacean zooplankton species,  Holopedium
                    gibberum,  Diaphanosoma  leuchtenbergianum,
                    Bosmina longirostris, and  Diaptomus minutus, occur-
                    red throughout the entire range of pH (3.60 to 7.25)
                    sampled in this survey (Table 1). Two species, Daphnia
                    retrocurva and Orthocyclops modestus, occurred only
                    at pH values ^ 7.00. Seven additional species, Sida
                    crystalline, Daphnia galeata,  D. dubia,  D. longiremis,
                    Ceriodaphnia  quadrangula,  Cyclops  bicuspidatus,
                    and Cyclops vernalis, were restricted  to survey waters
                    with pH values ^ 6.00. Only one limnetic zooplankter,
                    Diaptomus leptopus, occurred only in acidic waters. A
                    few  limnetic species,  Leptodora  kindti, Diaptomus
                    oregonensis, and D. sanguineus, occurred once each
                    in circumneutral survey waters.
                      Compositional trends of crustacean zooplankton
                    species  in the survey waters are illustrated  in  the
                    results of  the cluster  analysis (Fig. 4).  Three large
                    distinct clusters are present:  the Woods Lake, Razor-
                    back Pond and Heart Lake clusters. The Woods Lake
 RESULTS

 Thirty crustacean zooplankton species were identified
 from the study waters (Table 1).  Diaptomus minutus
20-
15-
10-
5-
n




r
••—

















                            -o 30-
            PH
 0»  15-

 I
<5  5-10 10-20 >20
chlorophyll a
  Cug-l-1)

           Elevation Cm)
£\i
I '5"
S
5 ""
V)
0>
-*
CO
_J 5-







il n
'-. -.^












.






-10 10-50 50-100 00-5(x!
Lake Surface Area 1





>500P
ha)
Figure 2.—Distribution of study waters in relation to  pH,
chlorophyll a, elevation, and surface area.
                     Table 1.—Crustacean zooplankton species identified from
                     Adirondack survey waters with percent occurrence and pH
                                     range of occurrence.
                                                                 Organism
Holopedium gibberum
Sida crystalline
Diaphanosoma leuchtenbergianum
Daphnia catawba
D. parvula
D. galeata
D. pulicaria
D. dubia
D. longiremis
D. ambigua
D. retrocurua
Ceriodaphnia quadrangula
Bosmina longirostris
Ophryoxus gracilis
Acroperus harpae
Chydorus spp.
Polyphemus pediculus
Leptodora kindti
Orthocyclops modestus
Cyclops bicuspidatus
C. scutifer
C, vernalis
Tropocyclops prasinus
Mesocyclops edax
Epischura lacustris
Diaptomus minutus
D. oregonensis
D, sanguineus
D. leptopus
Eubosmina coregoni
'Single occurrence

45674567  pH Occurrence
                                                             Percent
                                                           Occurrence
65.5
 5.5
60.0
50.9
30.9
 5.5
14.5
18.2
14.5
 9.1
 3.6
 7.3
92.7
 3.6
 1.8
 9.1
 3.6
 1.8
 3.6
 7.3
12.7
 7.3
54.5
76.4
25.5
98.2
 1.8
 1.8
 5.5
 1.8
                                                  381

-------
  LAKE AND RESERVOIR MANAGEMENT
 cluster is a group of 11 acidic waters (pH = 4.81, see
 Table 2). The crustacean zooplankton community of
 this cluster is characterized by the dominance of Diap-
 tomus   minutus,  Bosmina   longlrostris,  and
 Mesocyclops edax (Fig. 5). In fact, these three species
 make up more than 90 percent of the crustacean com-
 munity in survey waters with pH values ^ 5.50, but
 less than 60 percent in survey waters with pH > 5 50
 (Fig. 6).
   The Razorback Pond cluster includes  10  survey
 waters ranging in pH from 5.05 to 6.60 (pH = 5.73, see
 Table 2).  The crustacean community of this cluste- is
 characterized by the same three species found in the
 Woods Lake  cluster is characterized  by  the  same
 three  species found in the Woods Lake cluster plus
 two additional  species, Holopedium gibberum amd
 Diaphanosoma leuchtenbergianum (Fig. 5).
   The  third major  cluster, Heart  Lake,  made up
 primarily of circumneutral waters (pH  =  6.46) ex-
 hibited the greatest diversity of species (Fig. 5). The
 crustacean community of this cluster of survey waters
 generally includes the  five species from the Razor-
 back Pond cluster plus Daphnia catawba, D. dubia,
 and Tropocyclops prasinus. The increased number of
 species   with  increased  pH of  survey waters  is
 reflected  in the  mean values of  the diversity indices:
 0.679 for the Woods Lake cluster, 1.231 for the Razor-
 back  Pond cluster, and  1.527  for  the  Heart Lake
 cluster (see Table 2).
   The ordination  model for the surveyed  waters is
 presented in Figure 7. The model illustrated accounts
 for 58  percent of the dissimilarity within the data set.
 The addition of a third axis (not  illustrated) increases
 this value to 68 percent. A plot of pH on the ordination
 model (Fig. 8) illustrates a distinct gradient, indicating
 the  important  effect of pH on the compositional
 trends.
   (I I2
   8
   ti  8
      345678
                        pH


Figure 3.—Relationship between number of crustacean zco-
plankton species identified from surveyed waters and pH.
 DISCUSSION

 Early studies of zooplankton in the Adirondack Moun-
 tain  Region  by  the Biological  Surveys during the
 1920's  and 1930's were limited to generic level iden-
 tification.  Hall and  Waterman (1968)  first described
 crustacean   zooplankton  species  occurrence   in
 Adirondack   waters.  These  authors  reported six
 species,  Epischura  lacustris, Diaptomus  minutus,
 Mesocyclops edax,  Holopedium gibberum, Daphnia
 a)
J£
 O
        (00   90
   % Similarity

80    70    60
                                     50    40
                                                                 90
                                                                                        50
                                                                                              40
                                                                                                    30
                  80    70    60
                     % Similarity

          Q-Mode  Cluster  Analysis
Figure 4.—Q-mode cluster analysis of Adirondack crusta-
cean zooplankton community composition.
 Table 2.-Mean values (± standard deviation) of various parameters for each of the three clusters identified in the survey of
	Adirondack crustacean zooplankion communities (see text for explanation).

                                                         Mean Values ± Standard Deviation

Clusters

Heart Lake
Razorback Pond
Woods Lake
'Shannon and Weaver, 1949
Number of
Survey Waters

12
10
11

Diversity
Index*

1.527 ±0.328
1.231 ±0.339
0.679 ±0.1 81


PH

6.46 ±0.37
5.73 ±0.59
4.81 ± 0.41

Secchi
Depth
(meters)
4.5 ±1.3
2.5+1.3
5.1 ±3.1


Chlorophyll a
04)'l-1)
3.86 + 2.26
8.35 ±6.73
2.87 ±2.17

                                                 382

-------
                                                                                        ACIDIC PRECIPITATION
 catawba, and Bosmina  longirostris, as  being wide-
 spread and abundant in  20 or more of the 36 Adiron-
 dack waters surveyed.  The most  recent  report  of
 crustacean zooplankton species in Adirondack waters
 was presented  by Confer et al. (1983). The study  in-
 cluded 10 Adirondack waters plus 10 waters in New
 Hampshire; the surveyed waters ranged in pH from 4.5
 to  7.2.  These authors report the occurrence  of five
 species,  Diaptomus  minutus,   Mesocyclops  edax,
 Polyphemus  prediculus,  Holopedium gibberum, and
 Dlaphanosoma  leuchtenbergianum, over almost the
 entire pH range. They also report that the occurrence
 of  four  species,  Epischura   lacustris,  Daphnia
 catawba,   Bosmina  coregoni,   and  Bosmina
 longirostris,  was restricted to pH values > 5. Daphnia


           Crustacean  Community  Composition

       Heart Lake               Razorback Lake
       Woods  Lake
                            Bosmina longirostris
                            Diaptomus minutus
                           > Daphnia catawba
                            Mesocyclops edax

                        ^^» Diaphanosoma  leuchtenbergianum

                        GH3» Daphnia dubia

                        ^2B=- Holopedium gibberum

                            Tropocyclops prasmus

                           • Other
Figure 5.—Crustacean zooplankton community composition
in three major clusters of Q-mode cluster analysis of Adiron-
dack crustacean community composition.
                                                    species were virtually absent from survey waters with
                                                    pH values < 5.
                                                      The present study is in general agreement with Con-
                                                    fer et al. (1983) regarding the dominant crustacean
                                                    zooplankton  species in the range  of pH  surveyed.
                                                    However, in contrast to Confer et al. (1983),  the  pre-
                                                    sent study demonstrated the occurrence of Bosmina
                                                    longirostris throughout the entire range of pH (3.60-
                                                    7.25). These results are consistent with Sprules' (1975)
                                                    study of 47 Ontario lakes where B. longirostris occur-
                                                    red throughout  the range of pH from < 4.0 to  7.0.  The
                                                    present study also reports the occurrence of three ad-
                                                    ditional species,  Daphnia catawba,  D. ambigua,  and
                                                    Tropocyclops prasinus, throughout  the  range of pH
                                                    (4.5-7.2) reported  in Confer et al (1983). In addition, the
                                                    present study  did not  find Polyphemus  pediculus
                                                    below pH 5.00, while Confer  et al  (1983) found  this
                                                    species to  occur  throughout the pH range of 4.5-7.2.
                                                    Sprules (1975) reports Polyphemus pediculus occurr-
                                                                                        RozortacK Pond cluster

                                                                                       39 40
                                                                               Woods Lake cluster
                                                         Q~mode Ordination   Crustacean Community Composition
                                                                           {In standing crop)       3

                                                       Figure 7.—Q-mode ordination of Adirondack crustacean zoo-
                                                       plankton community composition. Cluster patterns are same
                                                       as in Figure 4. Lake numbers on ordination correspond to
                                                       numbers used in Figures 1 and 4.
                                                          Q-mode Ordmotion  Crustacean Community Composition
                                                                           (In standing crop)
    I
    y,  so
    CO
O
       TO
       60

       50
    ¥>
    a;  40
    o>
    o

    1  30
    £  20

    §  10

       0
                   Mesocyclops edax,
             Bosmina longirostris, Diaptomus minutus

               «450  451-550  551-650


                           PH

Figure 6.—Relationship between  mean percentage of
Mesocyclops edax, Bosmina longirostris, and Diaptomus
minutus in crustacean zooplankton standing crop of pH. N
=  number of surveyed waters.

                                                     PH

                                                    •  -50
                                                    •  50-599
*                                                      60-699
                                                      *70
                                                   Figure 8.—Q-mode ordination showing pH.
                                                    383

-------
 LAKE AND RESERVOIR MANAGEMENT
 ing at pH  values as low as 4.2. Differences in  the
 results  of  these  studies can be attributed  to  dif-
 ferences in sampling methodologies and lake sample
 size.
   The relationship between the number of crustacean
 zooplankton species and pH reported  in this stud/ is
 consistent  with  the results of  other  studies:  the
 number of  species declines with decreasing pH (see
 Haines, 1981; U.S. Environ. Prot. Agency, 1983 for
 reviews). In general, acidified  waters have  simple
 curstacean zooplankton communities.  In addition, the
 relationship between pH and the number of crusta-
 cean  zooplankton species  in Adirondack waters is
 consistent  with the relationship between pH and  the
 number of  rotifer (Siegfried et  al. 1983) and  phyto-
 plankton species (Siegfried et al. in prep.). The rotifer
 community assemblage of  humic acidified  waters
 was virtually identical to the assemblage of clear-
 water acidified waters (Siegfried et al. 1983). This v/as
 not the situation for the crustacean zooplankton com-
 munity reported  in this  study.  Clearwater acidified
 lakes—the  Woods Lake cluster—generally contain
 only  three  dominant  crustacean  species, while  the
 humic acidified lakes generally fall within the Ra;:or-
 back Pond  cluster, indicating a  more diverse crusta-
 cean community  assemblage. These results are con-
 sistent with the findings of Raddum (1978) that some
 species of crustacean zooplankton are  absent or have
 low abundance in clearwater acidified  lakes buy may
 develop substantial  populations  in humic  acidified
 lakes. The organics in the humic lakes may complex
 with the dissolved metals, mitigating metal toxicity.
   The  structure  of  crustacean  zooplankton  com-
 munities  in Adirondack waters with varying pH may
 represent only an  indirect relationship to this faclor.
 As pH decreases, other chemical changes also occur
 that may affect the crustacean community, for exam-
 ple, mobilization of heavy metals. Changes  in phyto-
 plankton, rotifer,  or fish  communities directly or in-
 directly related to pH may have significant impacts on
 the crustacean community.  The  survey approach to
 pH: aquatic community characteristics will provide a
 basis  for developing  experimental studies  to del er-
 mine the role of various processes—toxicity, competi-
 tion, and predation—in structuring the communities
 of acidified  waters.
   Neutralization  of acidified waters  with  hydraled
 lime (Ca[OH2]),  agricultural  limestone (CaCO3), and
 soda ash (Na2CO3) is a technique that has been us.ed
 to reclaim fisheries in New  York State. Preliminary
 analysis of  neutralization impacts on the plankton of
 an Adirondack ponded water indicates an immediate
 decrease of several  orders  of magnitude in  phyto-
 plankton and rotifer densities, but little noticeable ef-
 fect on  crustacean  zooplankton  densities (unpubl.
 data). The perturbation represented by  the rapid shift
 in pH from 4.30 to 7.30 eliminated most of the phyto-
 plankton and rotifiers in the  system. The persistence
of the crustacean zooplankton, primarily Diaptomus
minutus, Bosmina longirostris and Daphnia catawna,
 following  treatment  indicates an impressive
tolerance, not only to  a wide range of pH but also to
 rapid changes in pH. The practice of stocking fish in
 neutralized  waters soon after treatment should con-
sider the recovery requirements of phytoplankton and
zooplankton communities as a basis for fish produc-
tion.
 REFERENCES

 Colquhoun, J., J. Symula, and R. Karcher, Jr. 1982. Report of
   Adirondack sampling for  stream  acidification studies.
   1981 supplement. Tech. Rep. 82-3. N.Y. State Dep. Environ.
   Conserv., Albany.

 Confer, J.L, T.  Kaaret, and G.E. Likens.  1983. Zooplankton
   diversity and biomass in recently acidified lakes. Can J
   Fish. Aquat. Sci. 40:36-42.

 Gibson, J., and R. Linthurst. 1982. Effects  of acidic precipita-
   tion on the North American  Continent. In T. Schneider and
   L. Grant, eds. Mr Pollution by Nitrogen Oxides. Elsevier
   Sci. Publ. Co., Amsterdam, The Netherlands.

 Haines, T.A. 1981. Acidic precipitation  and its consequences
   for aquatic ecosystems:  a  review. Trans. Am.  Fish Soc
   110:669-707.

 Hall,  D.J.,  and  G.G. Waterman. 1968. Zooplankton  of the
   Adirondacks. N.Y.  Fish Game J.  15:186-90.

 Hendry, G.R. 1980. Effects of acidity on primary productivity
   in lakes: phytoplankton.  Pages  357-371 in D.  Shiner, C.
   Richmond,  and S. Lindberg, eds.  Atmospheric  Sulfur
   Deposition—Environmental  Impact and  Health Effects.
   Ann Arbor Science, Ann Arbor, Mich.

 Omernik, J., and C. Powers. 1982. Total alkalinity of surface
   waters—a national map.  EPA-600/D-82-333. U.S. Environ.
   Prot. Agency, Corvallis, Ore.

 Park,  R.A.  1968. Paleocology of  Venericardia sensu law
   (Pelecypoda) in the Atlantic  and Gulf coastal province: an
   application of paleosynecologic methods. J.  Paleontol
   42:955-86.

 Pfeiffer, M., and P. Festa. 1980. Acidity status of lakes in the
   Adirondack  region of New  York  in  relation to fish
   resources. FW-P168 (10/80).  N.Y.  State Dep. Environ. Con-
   serv., Albany.

 Raddum, G.G. 1978.  Invertebrates: quality and  quantity of
   fish food. In G. Hendry ed., Limnological Aspects of Acid
   Precipitation.  Brookhaven Natl. Lab., Upton, N.Y.

 Schofield, C.L 1978. Guidelines for survey  and monitoring of
   acid sensitive waters in New York State.  N.Y. State College
   Agric. Life Sci., Cornell Univ., Ithaca,  N.Y., for the N.Y.
   State Dep. Environ. Conserv. Unpubl. rep.

 Shannon, C.D.,  and  W.  Weaver. 1949. The Mathematical
   Theory of Communication. University of Illinois  Press, Ur-
   bana.

 Siegfried, C.A., J.W. Sutherland, S.O. Quinn, and J.A. Bloom-
   field. 1983. Lake acidification and the biology of Adiron-
   dack lakes. I Rotifer communities. Verh.  Int. Verin. Limnol.
   23 (in press).

 Siegfired, C.A., J.W.  Sutherland, J.A.  Bloomfield, and S.O.
   Quinn. (in prep.). Lake acidification and  the biology  of
   Adirondack lakes. III. Phytoplankton communities.

Sorenson, T. 1948. A method of establishing groups of equal
   amplitude in  plant sociology based  on similarity  of
   species content and its application to analyses  of the
  vegetation on  Danish commons. Biol. Sci. 5:1-34.

Sprules, W.G. 1975.  Midsummer crustacean zooplankton
  communities in acid-stressed lakes. J.  Fish.  Res  Board
  Can. 32:389-95.

U.S. Environmental Protection Agency.  1983.  The  acidic
  phenomenon and its effects. Critical assessment  review
  papers. Public review draft. 2 volumes. EPA-600/18-83-016
  A and B. Washington, D.C.
                                                   384

-------
THE LITTORAL ZOOPLANKTIC COMMUNITIES
OF AN ACID AND A NONACID LAKE IN MAINE
MIKE BRETT
Maine Cooperative  Fishery Research  Unit
Department of Zoology
University of Maine
Orono,  Maine
             ABSTRACT

             The littoral zooplanktic communities of an acidic and a nonacidic lake in Hancock County,
             Maine, were studied. Although the lakes are less than 200 m apart and similar in character and
             physical habitat, average annual pH differs in the two lakes. Fish are absent in the acidic lake
             while the nonacid lake contains golden shiners (Notemigonus crysoleucas),  brook trout
             (Salvelinus fontinalis) and rainbow smelt (Osmerus mordax). The zooplanktic community of the
             acid lake is dominated by the adult and nauphlii of the copepod, Diaptomus minutus, the large
             cladoceran, Diaphanosoma brachyurum, and the acid water rotifer, Keratella taurocephala. The
             nonacid lake is dominated by four zooplankters, D. minutus, the caladoceran, Bosmina cor-
             egonis, and the rotifers, K, taurocephala and Keratella cochlearis. The density of three species
             of zooplankton was greater in the acid lake and the density of ten species was greater in the non-
             acid lake. Notably, the large cladocerans, D. brachyurum, Polyphemus pediculus, and Acroperus
             harpae, were more dense in the acid lake. In the nonacid lake, the density was higher for the
             copepods, D, minutus  and  a cyclopoid, the cladoceran, 6.  coregonis, and  the rotifers, K.
             taurocephala, K. cochlearis, Keratella sp., Polyarthra remata, Asplanchna sp., and Trichocerca
             sp. A. Three factors are probably responsible for these differences, (1) biotic changes caused by
             the absence of fish  in  the acid lake, (2) oligotrophication resulting from acidification and a
             shorter flushing time in the acid lake, and (3) toxic effects of acidification.
INTRODUCTION

The zooplanktic community is an integral part of most
lentic ecosystems. Several authors have studied how
zooplanktic  communities are affected  by acidifica-
tion, and it is generally agreed zooplankton  com-
munities decline in diversity  and density (Raddum et
al.  1980; Roff and  Kwiatkowski, 1977; Sprules 1975).
Eriksson et al. (1977) removed fish from a nonacid lake
and found changes in the planktic community which
were attributable to the absence of fish. Grahn et al.
(1974)  stated  that  acidification  can  have  oligo-
trophicating effects on freshwater ecosystems. Tox-
icity is  another factor that could affect zooplanktic
communities. To  what extend these  three  factors
(biotic changes, oligotrophication, toxicity) affect the
zooplanktic communities was addressed in a study of
an acid  and  an nonacid lake  in Maine.
DESCRIPTION OF STUDY SITE

The study site consists of two lakes, Mud Pond and
Salmon Pond, in Hancock County, Maine. One lake is
acidified (Mud Pond) while the other serves as a con-
trol (Salmon Pond). The lakes are  less than 200 m
apart. Mud Pond has a surface area of  2 ha and a
maximum depth of  18 m, while the surface  area for
Salmon Pond is 4 ha and the maximum depth is 12 m.
The major difference between  the  lakes is the acid
lake (Mud Pond) has a average pH of 4.5 while the non-
acid lake (Salmon Pond) averages pH 6.2.
  The acid lake  has three inlet streams and one outlet
stream. The nonacid lake is spring fed with no inlets
and one outlet.  The  shorter flushing time of the acid
lake causes the pH  of the  lake water to approximate
the pH  of precipitation for the northeastern United
States (pH 4.0-4.4) (Atmos. Environ. Serv. 1979). In the
nonacid lake, with its slower flushing rate, there is
more time for the rain to be buffered, thus the lake pH
is significantly higher than that of atmospheric input
(S. Kahl, pers. comm.).
  There are no fish in the acid lake; however, it does
have  high densities of potential planktivorous  in-
vertebrates, mainly the backswimmer, Buenoa sp. The
nonacid lake has three species of fish: golden  shiners
(Notemigonus  crysoleucas),  brook trout (Salvelinus
fontinalis),  and  rainbow  smelt (Osmerus mordax).
Golden shiners are the most abundant; a Schumacher
and  Eschmeyer  estimate  (Ricker,  1958)  based  on
multiple mark-recaptures puts numbers of shiners at
1,098  (95 percent confidence  interval of 997 to 1,224)
for fish greater than or equal  to 70 mm (C.W. Fay, un-
publ. data).
METHODS

The littoral zones of both the acid and nonacid lakes
were sampled on four dates, June 21, July 7, July 21,
and Aug. 8, 1983. On each date, zooplanktic samples
were taken from eight fixed points around each lake.
Vertical samples were taken with aim plexiglass
tube at a lake depth of 1 to 2 meters. Each sample con-
sisted of 8.34 I of lake water. The zooplankters were
removed from  the  water with the filter  house of  a
Rohde  sampler with a 64 /^m screen  and  all samples
were preserved with 1 percent lugols. Species were
identified  using the keys of  Edmondson  (1959) and
Pennack (1978). A two-way analysis of variance, with
Duncan's multiple comparison, was used to determine
differences in zooplanktic densities in the two lakes.
                                                385

-------
 LAKE AND RESERVOIR MANAGEMENT
 RESULTS AND DISCUSSION

 During the sampling period, 18 species of zooplankton
 were identified from the acid lake and 21 from the non-
 acid lake. Sixteen species were found in both lakes.
 The two zooplankters found only in the acid lake were
 the cladocerans, Acroperus harpae and Chydorus up.
 The five found only in the nonacid lake were the clado-
 cerans,  Holopedium  gibberum  and  Bosmina
 longirostris, and the rotifers, Kellicottia longispina,
 Filinia longiseta, and Polyarthra spp. (Table 1).
  The acid lake had three species that were dominant
 (greater than 10 percent of total counts) and four that
 were common (greater than 1 percent of total counts)
 on any one sampling date. The dominant species were
 Diaptomus minutus, Diaphanosoma brachyurum, aid
 Keratella taurocephala, and the common species were
 a  cyclopoid,  Bosmina  coregonis,   Polyphemus
pediculus, and A. harpae. The nonacid lake had four
 dominant species,  D.  minutus,  B.   coregonis,  K.
 taurocephala, K. cochlearis, and six common species,
 a cyclopoid D.  brachyrum, Keratella  sp.,  Polyarthra
remata, Asplanchna sp., and  Trichocerca sp. A.
  In the 32 samples taken from the acid lake, sped as
diversity ranged from  a low  of four crustacean  and
zero rotifer plankters to a high of seven crustaceans
and three rotifers per sample. For the same number of
samples in the nonacid lake, the low was three crusla-
ceans and two rotifers to a high of five crustaceans
and 10 rotifers.  Sprules (1975) stated that the crusla-
cean zooplanktic communities of acid lakes (pH 5.0)  ranged
from nine to 16 species. The acid lake ranged from
four to  seven crustaceans per  sample and  is well
within the confines of Sprules' estimate.
  The nonacid lake ranged from three to five species
and is well below the estimate of nine to 16. This could
be because of  the  small  size of  the  nonacid  lake.
Sprules (1975) also found a correlation between lake
area and community diversity, or the difference could
be due to other factors (that is, a qualitative difference
between the  zooplanktic  communities  of Ontario,
Canada, and Maine).
  The fact that the rotifer community was much less
diverse  in the acid lake follows the findings of Rad-
dum et al. (1980) that several common rotifers are rare-
ly found in acid lakes. In addition, Pejler (1957, 1983)
stated  that planktic  rotifer species diversity suc-
cessively decreases as lakes tend toward oligotrophy
and  ultraoligotrophy. This  is important inasmuch as
oligotrophication has been cited as an effect of acidi-
fication (Grahn et al. 1974).
  Zooplanktic densities of all the dominant and com-
mon species were different in  the acid and nonacid
lake.  In addition, the rotifer, Kellicottia longispina,
which was neither dominant nor common in either of
the lakes was found to have different densities  in the
two lakes. Three zooplankters were more abundant in
the acid lake (Fig. 1). These were the large cladoceran
zooplankters (>.51 mm), D. brachyurum, and A. harpae
and  P.  pediculus. Higher numbers of  large  clado-
cerans could possibly be explained by the absence of
fish in the acid lake.
  Eriksson et al. (1979) stated that in  Lilla  Stock-
elidsvatten, a nonacidified lake from which fish were
removed, smaller zooplankters were replaced by  larger
zooplankters.  The  predacious   zooplankter, P.
pediculus, is commonly found in acid waters (Roff and
Kwiatkowski,  1977;  Sprules (1975).  Sprules  (1975)
speculated that this was caused by reduced fish
species diversity and reduced competition from other
predacious plankters (eg. Leptodora, Epischura).
  Raddum et  al.  (1980)  found  a similar  situation in
Norway with the large predacious plankter, Hetero-
cope saliens. This species was common in acid lakes
and  probably increased  in density in the absence of
fish. The nonacid lake had 10 species more abundant
                              Table 1.—Average density per 10 I for 24 samples.

Copepoda
Diaptomus minutus
Cyclopoida
Nauphlii
Cladocera
Bosmina coregonis
Diaphanosoma brachyurum
Polyphemus pediculus
Acroperus harpae
Daphnia sp.
Bosmina longirostris
Holopedium gibberum
Chydorus sp.
Rotatoria
Keratella taurocephala
Keratella cochlearis
Keratela sp.
Kellicottia longispina
Polyarthra remata
Polyarthra spp.
Asplanchna sp.
Trichocerca sp. A
Trichocerca sp. B
Lecane sp.
Brachionus
Monostyla sp.
Flllnia longiseta
' p < 001
' p < 01
Acid
(Mud Pond)

25.70
2.54
126.33

5.83
17.96
10.25
1.25
0.04
0
0
0.08

41.62
0.04
0.08
0
0.04
0
0.16
0.21
0.08
0.04
0.29
0.08
0


Nonacid
(Salmon Pond)

79.37
25.83
107.33

118.33
4.21
.33
0
0.04
0.04
0.13
0

345.63
91.62
35.08
1.50
10.50
1.87
7.08
9.75
0.17
0.12
0.04
0.12
0.04


Significance

* *i
„*
NS

,.
* 2
* *
* *
NS
NS
NS
NS






NS


NS
NS
NS
NS
NS


                                                386

-------
                                                                                        ACIDIC PRECIPITATION
 than the acid lake. Two of these species, D. minutus
 and B. coregonis, inhabit waters of pH values ranging
 from  4.0 to 7.0 (Sprules, 1975). Another species, K.
 taurocephala, is considered  an acid water  species
 (Pennack,  1978)  (Fig. 2). That these three  species
 which are tolerant of low pH values are more abun-
 dant in the nonacid lake than the acid lake suggests
 that   acidification  has  significant effects  besides
 direct  toxicity. Since these species are pH  tolerant
 they should  be in  equal  numbers in  both lakes, all
 other  factors being  equal.  D.  minutus   and  K.
 taurocephala  are probably less abundant in the acid
 lake  because  it is more oligotrophic than the nonacid
 lake, an effect of flushing time and acidification.
   Eriksson   et   al.  (1979)  stated  that  Bosmina
 longirostrls were  replaced  by  larger calanoids  (eg.
 Eudlaptomus gracilis) in Li I la  Stockelidsvatten. There
 was  no correlation between Bosmina coregonis  and
 larger  filter feeding zooplankters, D. minutus and D.
 brachyurum, in either the acid or the nonacid lake. It is
 therefore doubtful that the lack of fish in the acid lake
 is the cause for the lack of abundance of B. coregonis.
    18


    16


    14


    12


    10


     8


     6


     4


     2
 1 2

 1 1

 1 0 '

 09 '

 08

 0.7 •

 0.6 •

 0.5

 0.4 .

 03 •

 02

 01
11

10
 7 •

 6 •

 5 •

 4 •

 3 •

 2 •

 1 -
         MUD   SAL
            POND
        Diaphanosoma
          brachyurum
      MUD   SAL
         POND
      Acroperous
        harpar
     MUD   SAL
       POND
     Polyphemus
      pediculus
                                    This difference is also probably attributable to oligo-
                                    trophication of the acid lake.
                                      That seven  species of planktic rotifers are more
                                    abundant  in the nonacid lake is attributable to two
                                    factors. First, as previously mentioned, the acid lake
                                    is more oligotrophic than the nonacid lake, and se-
                                    cond,  toxicity. The toxic effects of acidification have
                                    been well documented for fish (Leivestad, 1982, Baker,
                                    1982;  Peterson et al. 1982; Haines, 1981; Drablos and
                                    Tollan, 1980) and to a  lesser extent for amphibians
                                    (Tome and Pough, 1982; Pough, 1976), but  much less
                                    so for invertebrates. Several authors have  mentioned
                                    that planktic  rotifers  are reduced in species  and
                                    numbers in acidified lakes (Roff  and Kwiatkowski,
                                    1977;  Raddum et al. 1980; Muniz, 1982), but the ques-
                                    tion of whether the direct toxic effects of acidification
                                    (for  example, low pH, high  metal concentrations) or
                                    the  secondary effects (such as oligotrophication or
                                    biotic  changes) are  primarily  responsible for  the
                                    decrease in planktic rotifer communities has not been
                                    answered. It is most certainly a synergistic combina-
                                    tion of the two, but the data from this study  would sug-
                                    gest that toxicity is quite important.
                                      Whereas both lakes had viable,  although different
                                    planktic crustacean communities,  the rotifer com-
100


 90


 80


 70
   I

 60^


 50 |


 40


 30


 20


 10
Figure  1.—Three large  zooplankters which seem to be
favored by absence of fish in acid lake.
                                              MUD  SAL

                                                POND

                                         Keratella cochleans
36

33

30

27

24

21

18

15

12

9

6

3
                                                                                  MUD  SAL
                                                                                    POND
                                                             Keratella sp
11

10

 9

 8 -

 7 -

 6

 5

 4 -
                                                             MUD  SAL
                                                               POND

                                                          Polyartha remata
    80


    70 -


    60


    50 -


    40


    30


    20


    10
120

110

100

 90

 80

 70

 60

 50

 40

 30 •

 20

 10
          MUD  SAL
            POND

      Diaptomus minutus
      MUD   SAL         MUD   SAL

        POND              POND

   Bosima coregonis   Keratella taurocephala
Figure 2.—Three species of low pH  tolerant zooplankters
which are more abundant in the nonacid lake. This is pro-
bably due to the oligotrophication of the acid lake with
sublethal toxicity possibly playing a lesser role.
                       4


                       3




                       2 J
                            MUD   SAL
                              POND

                           Asplanchna sp
                  8 •


                  7 '

                  6 .


                  5


                  4


                  3

                  2
                       MUD  SAL
                         POND
                1.6 n



                1.4



                1 2


                1.0



                0.8-



                0.6


                0.4:



                02 •
                     MUD   SAL
                        POND
                                                                                Trichocerca sp   Kellicottia longspina
                                   Figure 3.—Six species of zooplanktic rotifers which appear
                                   to be affected by direct pH toxicity.
                                                    387

-------
 LAKE AND RESERVOIR MANAGEMENT
 munity of the acid lake was characterized by the one
 acid-adapted rotifer, K. taurocephala, while that of the
 nonacid  lake was characterized  by the  rotifers, K.
 taurocephala, K. cochlearis, Keratella sp, Polyarthra
 remata, Asplanchna sp., and Trlchocerca sp. A. All
 these rotifers, except K. taurocephala, were nearly nb-
 sent in the acid lake (Fig. 3). Oligotrophic lakes are
 characterized by low diversity  and numbers of all
 types of  zooplankters (Wetzel, 1975; Pejler, 1983) but
 they generally retain several species  of  rotifers and
 are not limited to only one species making up a signifi-
 cant part of the planktic community.
 SUMMARY

 Acidification probably has three strong effects on zoo-
 planktic communities, each in its own way affecting
 species diversity and abundance: (1) Biotic changes,
 changes caused in the lake due to species disappear-
 ances (such as fish) related to acidification; (2) oligo-
 trophicating  effects  of  acidification which  reduce
 diversity;  and  (3) toxicity,  which  eliminates those
 animals not tolerant of low pH values and high  con-
 centrations of heavy metals. In  the study lakes, the
 fact that D. brachyurum,  P. pediculus, and A. harpae
 were more abundant in the acid lake is attributable to
 the absence of fish predation in the acid lake. That the
 pH tolerant plankters, D. minutus, B. coregonis, and
 the acid-adapted K. taurocephala are more abundant
 in the nonacid lake is attributable to the fact that the
 acid lake is more oligotrophic than the nonacid lake.
 That  the rotifers, K.  cochlearis, Keratella sp., Polynr-
 thra remata, Asplanchna sp., and  Trichocerca  sp. A.,
 are more abundant in the nonacid lake is probably at-
 tributable  to the toxicity of  acidification.

ACKNOWLEDGEMENTS: I am  indebted to Steve Kahl  wro
explained the probable cause of Mud Pond's acidity, lo
Clem  Fay who provided population estimates for  golden
shiners in Salmon  Pond, to Lisa Tabak  who helped  wilh
identification,  to  Joan Trial who provided much  needed
help with statistical analyses, to Joan Trial and John  Mcr-
ing  who  critically  read  the manuscript, and  to  Noreen
Modery who typed it. The project was supported by the U.!5.
Fish and  Wildlife Service and  the U.S. Environmental  Pro-
tection Agency.

REFERENCES
Atmospheric Environment  Service.  1979. CANSAP dala
  summary. Downsview, Ontario.
Baker, J.P. 1982. Effects on fish of metals associated  wilh
  acidification. Pages  1265-176 in R.E. Johnson, ed. Acid
  Rain/Fisheries,  NE Div., Am. Fish. Soc., Bethesda,  Md.
Drablos, D., and A. Tollan, eds. 1980. Proc.  Conf. on the  Eco-
  logical Impact of Acid Precipitation. Acid Precipitation Ef-
  fects on Forest and Fish Proj., Aas, Norway.
 Edmondson,  W.T., ed.  1959. Fresh-water Biology. 2nd  ed.
   John Wiley and Sons Inc.. New York.

 Eriksson, M.O.G.,  et  al.  1979. Important  for the biotic
   changes in acidified lakes. Ambio 9: 248-9.

 Fay, C.W. Unpubl. data. Maine Coop. Fish,  Res. Unit, Univ.
   Maine, Orono.

 Grahn, O., H. Hultberg, and L. Landner. 1974. Oligotrophica-
   tion - a Self-accelerating process in Lakes  subjected to ex-
   cessive supply of acid substances. Ambio 3: 93-4.
 Haines, T.A. 1981. Acidic precipitation and its consequences
   for aquatic ecosystems: a review. Trans. Am. Fish Soc
   110:669-707.

 Kohl, S. Pers. comm. Univ. Maine, Orono.

 Leivestad, H. 1982.  Physiological effects  of acid stress on
   fish.   Pages  157-164  in   R.E. Johnson,  ed.  Acid
   Rain/Fisheries, NE Div., Am. Fish. Soc.,  Bethesda, Md.
 Muniz, I.P. 1982. The effects of acidification on Norwegian
   freshwater ecosystems. Pages 299-322 in Ecological  Ef-
   fects of Acid Deposition. Natl. Swedish Environ. Prot.
   Board, SNV PM 1636, Berlings, Arlov, Sweden.

 Pejler, B. 1957. Studies on the taxonomy  and ecology of
   planktonic rotatoria. Thesis. Uppsala, Sweden.

 Pejler, B. 1983.  Zooplanktic indicators  of trophy and their
   food. Hydrobiologia 101: 111-13.

 Pennak, R.W. 1978. Fresh-water Invertebrates of the United
   States. 2nd ed. John Wiley and Sons, New York.

 Peterson, R.H., P.F. Daye, G.L. Laroix, and E.T. Garside. 1982.
   Reproduction in fish experiencing acid and  metal stress.
   Pages 177-196 in R.E. Johnson, ed. Acid Rain/Fisheries, NE
   Div., Am. Fish. Soc., Bethesda, Md.

 Pough, F.H. 1976. Acid precipitation and embryonic mortality
   of spotted salamanders Ambystoma maculatum. Science
   192: 68-70.

 Raddum, G., A. Holbaek, E. Lomsland, and T. Johnsen. 1980.
   Phytoplankton and zooplankton in acidified lakes in south
   Norway. Pages 3321-33 in D.  Drablos and A. Tollan, eds.
   Proc.  Int.  Conf.  Ecol.  Impact  of  Acid Precip.,  Acid
   Precipitation Effects on  Forest and Fish  Proj., Aas, Nor-
   way.

 Ricker, W.E.  1958. Handbook of compilations for biological
   statistics of fish populations. Bull. Fish. Res. Board Can.
   119.

 Roff, J.C., and R.E. Kwiatkowski. 1977. Zooplankton and zoo-
   benthos communities of selected northern Ontario lakes
   of different acidities. Can. J. Zool. 55: 899-911.
Sprules,  W.G. 1975. Midsummer  crustacean  zooplankton
   communities in acid-stressed lakes. J.  Fish. Res.  Board
   Can. 32: 389-95.

Tome, M.A., and F.H. Pough. 1982. Responses of amphibians
  to acid precipitation. Pages 245-254 in R.E. Johnson, ed.
  Acid Rain/Fisheries,  NE  Div., Am. Fish. Soc., Bethesda
   Md.

Wetzel, R.G. 1975. Limnology. Saunders, Philadelphia.
                                                     388

-------
 SOIL  LIMING AND RUNOFF ACIDIFICATION  MITIGATION
 PER WARFVINGE
 HARALD SVERDRUP
 Department of Chemical  Engineering
 Lund Institute of Technology
 Lund, Sweden


            ABSTRACT

            A measure often taken and frequently discussed in Sweden is to try to restore the runoff quality
            from acidified watersheds by soil liming. In Scandinavian acidified soils, where the base satura-
            tion often is below 20 percent, the dissolution of calcite will proceed without significantly im-
            proving runoff water quality until the base saturation is close to 100 percent. The dissolution rate
            of the calcitic minerals is governed by the particle size of the limestone used. Particles with a
            diameter over 0.3 mm will be used to a very small extent. As nearly all the runoff percolates
            through the soil column, a high base saturation is needed to get a stable and long lasting effect
            on the runoff water. The amounts needed to get a satisfying base saturation are estimated to be
            35 to 50 ton/ha. This implies that soil liming is one order of magnitude more costly than standard
            lake liming techniques for surface water. Although vast efforts may soon be needed in soil and
            forestry management in Scandinavia, the costs involved in soil liming emphasize the importance
            of defining the goal of every liming project.
 INTRODUCTION

 Thousands of lakes and rivers in Scandinavia and
 North America located on low weathering bedrock or
 surrounded by soils with low buffering capacity are
 suffering from the effects of acid precipitation.
  ,A major part of the acidification of surface waters
 arises from the fact that the surrounding soils have
 become acidified. After years of acidified precipitation
 and deposition the weathering of minerals cannot
 keep up with the neutralization need, and as a result,
 the reservoir of exchangeable  bases in the soil has
 slowly been depleted.
   Runoff from acidified soils also has elevated con-
 centrations of dissolved metals,  causing  biological
 problems  in the  soil and the surface waters. In par-
 ticular, high concentrations of aluminum may reduce
 growth rates for plants and forests, the latter being of
 economic importance  in regions  depending on  the
 forestry industry.
   A measure frequently taken  in Sweden is to  try to
 restore the buffering capacity in the soil to an accep-
 table level by liming. In this paper we will discuss the
 chemical  reactions  involved,  possible  effects  on
 runoff and surface water quality, and the costs.


 CATION  EXCHANGE REACTIONS

 One of the most important chemical properties of
 soils is the participation in cation exchange reactions.
The nature of cation exchange  is shown  in Figure 1.
 Cations, such  as H + ,  Na + , K + ,  Ca2 + ,  Mg2 + , and
A|3+ in the soil solution replace each other. For in-
 stance, the exchange between dissolved Ca2 + and ex-
 changeable hydrogen can be expressed as

     R-H2 + Ca2+ =  R-Ca + 2H +

where R represents the exchanging specie.
  The distribution of a certain cation  between the li-
quid and the solid phase depends on the  concentra-
tion in  the soil solution and the selectivity of the ex-
 changer for the species.  This can  be expressed in
 terms of equilibrium constants, as  postulated by
 Gapon (1933). For this reaction, the selectivity is defin-
 ed as

     Ca     R-H»(Ca)y*
        K = 	
     H       V2 R - Ca »(H)

 From this expression one can clearly see that an in-
 crease in hydrogen ion concentration, i.e., a pH drop,
 drives off Ca2+ from the exchange sites.
   Selectivity coefficients for  pairs  of  ions may vary
 from soil to soil but can fairly easily be determined in
 the lab (Robbins et al. 1980).
   The cation exchange capacity (CEC) varies within a
 large range, but can usually be determined with 25 per-
 cent with the formula

     CEC = 0.5«%C|ay + 2.5«%organics[meqv/100g]

   In  Scandinavian  soils,  the  main  part  of CEC
 originates from organic matter, especially decomposi-
 tion products such as humic and fulvic acid. Thus, the
 CEC varies considerably with depth, ranging from 250
 meqv/100g in the upper few centimeters to 5 meqv/100
 g soil at a depth of 50 cm.
   In acidified Scandinavian soils, less than 20 percent
 of the exchange sites are occupied  by basic cations,
 while the rest have hydrogen or aluminum ions attach-
 ed. This is  generally expressed in terms of the base
 saturation, defined as
    BS = 100»
S exchangeable bases

        CEC
or
               I exchangeable acids-CEC
    BS = 100*  	      [%]
                         CEC
                                                389

-------
 LAKE AND RESERVOIR MANAGEMENT
 The base saturation can be interpreted as the buffer-
 ing capacity of the soil since basic cations are readily
 replaced by the more strongly bonded H+ and Al:' + ,
 thus neutralizing acidic soil solutions.

 ON THE DISSOLUTION  OF CALCITIC
 MINERALS IN SOILS

 The dissolution of calcite in water solutions occur ac-
 cording to the reactions
CaCO3 + H+ ^
CaCO3
CaC03
                             HCO3
                                +  2HCO3
                             HCO3 + OH-
 Dolomite dissolves analogously but incongruently in a
 manner that implies that the  calcite part dissolves
 more rapidly than the magnesia part.
   Data support that dissolution in  the soil  can  be
 described  with the kinetics for a stagnant medium.
 The amount dissolved will then depend on the  mineral
 surface area available,  A,  and the chemical  driving
 force, F:
       dM
       dt
                                           [ke/sj
 where M is the amount of calcite and kM is a constant.
 This can also be expressed in terms of the particle
 radius r:
  dM        M
 •     = kM~rT"
  dt
            AP+
                                           [kfi/s]
WATER
                                 SOIL
Figure 1.—Examples of cation exchange reactions on a soil
surface.
                                               The validity of these kinetic expressions is shown in
                                               Figure 2 where relative dissolution rates in a soil lim-
                                               ing  experiment  (Meyer  &  Volk,  1952) are plotted
                                               against r2. In  calcareous or heavily limed soils the
                                               reaction
                                                   Ca2+
                  5£ CaCO
                                                                            H
will slow down the dissolution as the ion activity pro-
duct of Ca2+ and Hco3 approaches the solubility pro-
duct.
  Even if the kinetics of limestone dissolution is the
same in all type of systems, the paths of the chemical
reactions are different in pure water and in  soil sys-
tems. In a soil with less than 100 percent base satura-
tion the following reaction path will dominate:
CaCO3
R-H2
H +  + HCO3
             H+ ~»
                                                                         HCO3
                                                                   R-Ca + 2H
                                                                 H2O +  C02
                                                According to Ihese reactions, the calcium released
                                              to the soil  by the dissolution will replace hydrogen
                                                    ;: 20   -
                                                      16   •
CO


-------
                                                                                    ACIDIC PRECIPITATION
 ions in the exchange complexes. This will delay the
 dissolution of the soil solution as well as limit the in-
 crease in the ion activity product for calcite. Thus, the
 dissolution will proceed until the base saturation  is
 close to 100 percent. This is presented graphically in
 Figure 3.
   Runoff water from limed soils will consequently get
 its alkalinity from two  resources:  the dissolution of
 residual  limestone  and  leaching of exchangeable
 bases.  The  dissolution of  residual  limestone  will,
 however, decline considerably because of inactivation
 of the surface. The quality of the runoff will therefore
 depend on the buffer capacity provided by exchange-
 able bases.
 IMPACT OF HYDROLOGY ON SURFACE
 WATER QUALITY

 To understand the dynamics of the water quality in
 lakes one has to take a look at the flowpaths'of  the
 groundwater in moraines (Bergstrom and Sandberg,
 1983).
   The water  in the lake originates  either from dis-
 charge of ground water from  the areas surrounding
                                 Discharge from

                                 buffering soil
                     Fig.4a
                                 pH-spates on

                                 buffering soil
                     Fig.4b
                                Discharge from

                                acidified soil
                    Fig.4c

Figure 4.—The effect of base saturation and flow rate on run-
off water quality.
 the lake or from direct precipitation. The smaller the
 lake area is in relation to its watershed, the more sen-
 sitive it is to  changes in the buffering capacity of the
 soil.
   In nonacidified soils, the pH in the ground water will
 vary as shown in Figure 4a. In general, the longer the
 residence time in the soil is, the higher the pH of the
 discharge. This  is caused  by  the impact of  ion ex-
 change  reactions   and  mineral  weathering. The
 weathering reactions are, however, very slow, and will
 only affect the deep ground water.  In periods with high
 runoff, pH spates will occur according to Figure 4b. In
 acidified soils, however,  the precipitation will not be
 neutralized as it  flows through the ground, and the pH
 in the surface waters will go down to very low levels.

 SOIL LIMING  TO MITIGATE  SURFACE
 WATER ACIDIFICATION

 When' soils are to be limed, for agricultural purposes
 or to neutralize  the  runoff, the base saturation level
 has  to  be  concerned  as well  as the  deposition
 neutralization need. An estimate of the amount need-
 ed to raise the base saturation of an acidified Swedish
 forest soil from  20 to 100 percent would be 35 to 50
 ton/ha. This would be enough to neutralize a soil with
 CEC eq. 20 meqv/100g to  a depth of 50 cm, ensuring a
 long lasting  and stable  improvement  in  the  runoff
 quality. As  a comparison,  the theoretical  amount
 needed to neutralize the acid precipitation amounts to
 only 70 kg/ha a year.
   How  different liming  rates  affect the  runoff are
 shown in Figure 5.
   In actual liming all the material may not be used,
 depending on the grain size and  reactions that may
 deactivate  the calcite mineral surfaces.  In  mathe-
 matical terms, this means that  the dissolution equa-
 tion will have to be completed with a deactivation term
 for long dissolution times
       dM
                 M
      dt
or for calcite powder
      dM         n
                     M,
                                                           dt
                 i = 1 r2
                                                                            -•F«v(t)
                                                    where the summation is performed over the different
                                                    particle size fractions in the calcite powder. Deactiva-
                                                    tion will not be significant for very small particles, but
                                                    of importance to the dissolution of the coarser ones.
             Acidified soil
             Acidified runoff
             Less  than 207. base  saturation


             Partially limed soil; S-15 ton/ha
             Possible preacidified condition of soil
             Slight acidic runoff
             40-50% base saturation

             Modestly limed soil; 15-20 ton/ha
             Possible preacidified condition of soil
             Partially neutralized runoff
             50-60% base saturation

             Heavy limed soil; 35-50 ton/ha
             Neutralized runoff
             80-100% base saturation
Figure 5.—The effect of different lime rates and base satura-
tion on run-off water quality.
                                                 391

-------
 LAKE AND RESERVOIR MANAGEMENT
 In general, particles larger than 0.3 mm diameter will
 be used to a very low degree.
   An estimate of the efficiency of different calcite
 meals may be taken from a chart made by Scholkm-
 berger  and Salter  (1943). This chart,  Figure  6, ex-
 presses the efficiency as a function of particle size
 and time. This chart is based on large amounts of li-n-
 ing  data.  It  is  interesting to note  that dolomite
 dissolves substantially slower than calcite at first, but
 dissolves relatively faster at higher pH levels.
   To manage water quality in a lake, three options are
 available:
     • Soil liming in the watershed
     • Dosing in the tributary
     • Lake liming
 In addtion to restoring the water quality to an accep-
 table level, liming a large fraction of a watershed with
 adequate amounts of calcite or dolmite will raise the
 base saturation level in the soil and  immobilize harm-
 ful metals.
   Liming  of surface waters directly may precipitate
 the  metals to an acceptable level, but only in lakes
 with retention times over half  a year.  Recent data
 (Brown, 1982; Scofield, 1982) do, however,  show that
 the  presence of calcium ions may reduce the toxicity
 of theA|3+ ion to fish.
   To illustrate the large differences in effort and cost
 involved in the different strategies let us consider an
 example:
     A lake with a volume of 1O10-6 m3 has
     a retention time of 1 year. This is
     equivalent to a flowrate of 0.3 m3/s.
     In an area with a runoff off  10 1/km2«s
     this demands a watershed of 30 km2.
     Exchange rate: 8 sek = 1 US$

 Soil Liming

 It is assumed that it is necessary to lime 50 percent of
 the watershed, and  bring the base saturation level up
 to at least 80  percent.
   For the total amount of calcite needed we get:

     Ms = ABSN-WA-XL                      [ton]
 where
 MS    = The total amount of calcite needed     [ton]
 ABSN = Area specific soil neutralization need [ton/ha]
 WA   = Watershed area                      [ha]
 XL    = Fraction of the watershed limed
 and for the total cost, TCS:
     TCS = MS«SPS                          [sek]
 where
 S P s  = The specific cost per ton spread    [sek/ton]
            U.S. Standard sieve no.
 The specific cost is composed of the material cost at
 100 sek/ton, plus the transportation cost: 100 sek/ton,
 plus the cost of spreading, 200 sek/ton.
   It is assumed that the treatment will have to be
 repeated when  40 percent of the total amount has
 been leached out. This gives an estimated duration of
 70 years. We then get:

     Ms = 35 ton/ha»3000 ha»0.5 = 52 500 ton
     TCS = 52 500 ton»400 sek/ton  =  21  mill sek


 Dosing in the Tributary

 It is assumed that the water will be neutralized from
 pH 4.7 to 6.7 during two-thirds of the year for 70 years.
 The  total amount needed will be:
     MD = ON»ApH»T/xd
where
[sek]
 MD    = The total amount calcite needed        [ton]
 Q     = Flow rate                           [m3/s]
 N     = Specific neutralization need      [g/pH»m3]
 T     = Total time                           [yrs]
 xd     = Dissolution efficiency
 The total cost is:
    TCD = MD»SPD + ICD                    [sek]
 where
 ICD    = Investment cost                      [sek]
 SPD   = The cost per ton spread           [sek/ton]
 The SPD is calculated using 200 sek/ton calcite, and
 100 kr for the transportation. The investment  cost is
 based on today's price for a unit, 250 000 sek. A total
 of four units is assumed.
 We then get

    MD =  107m3/yr«5«10-6ton/pH»m3
           2 pH«(70 yrs»2/3)/0.7 = 6 330 ton
    TCD = 6 330 ton»300 sek/ton +
           4 units»250 000 sek/unit = 2.9 mill sek

 Lake  Liming

 It is assumed that the lake will be limed from pH 4.7 to
 pH 7.0. For a lake with a retention time of 1 year, the
 time for it to reacidify  to pH 6 will be approximately 2
 years.  Liming is then performed from pH 6 to pH 7.
  The total amount needed to keep the lake above pH
 6 will  be:
    ML = V»N»ApH»T/tL«xL
where
V      = Lake volume
tL      = Duration of one treatment
XL     = Dissolution efficiency
The total cost will then be
    TCL = ML»SPL
[m3]
[yrs]
                                                                                                 [sek]
Figure 6.—Dissolution efficiency of different commercial
limestone powders.
                                                    The cost per ton spread is calculated using 100 kr/ton
                                                    as material cost, 100 sek/ton transportation cost and
                                                    100 sek/ton labor cost. This adds up to:

                                                         ML = 107 m3«5»10-6 g/pH»m3 .2.3 pH»
                                                              (70 yrs/2 yrs)/0.4             = 10 100 ton
                                                    and, finally
                                                         TCL = 10 100 ton. 300 sek/ton      = 3.0 mill sek
DISCUSSION

From the example given it can be seen that the costs
and  efforts involved  in soil liming are one order of
magnitude  larger than  in  liming  the  target water
                                                 392

-------
directly. This emphasizes the importance of clearly
defining the target of interest. If the water quality and
fishery are the  main interest, than it is a far  more
economical way to treat the lakes directly than  to go
into soil liming.  Because of the costs involved in soil
liming, this mitigation strategy is only worth consider-
ing in cases  where the forestry and groundwater re-
sources are the  targets.
ON THE SOLUTION OF THE PROBLEM
OF ACID PRECIPITATION

It  must be  entirely clear that liming represents no
satisfactory solution to the problem of acid precipita-
tion,  it merely partially repairs  some of the worst
damage. In Scandinavia, some areas now show severe
forest damage caused by acid precipitation and soil
acidification.  In a  near future  it may therefore be
necessary to lime vast areas in order to save this in-
valuable natural resource.
   With the costs previously outlined, let us speculate:
Year

1985
1995
Affected area
 in Sweden

 50 000 km2
 100 000 km2
   Cost

 70»102 kr
 70«109 kr
140*102 kr
  Similar estimations  can  be  made for  Norway,
Canada, Central Europe, and the northeastern United
States.  In comparison, the total amount listed equals
the cost of more than a hundred stack gas scrubbers
at 1 TX  power plants.
                                                                      ACIDIC PRECIPITATION

                                        The only complete solution to the problem of acid
                                      precipitation is to reduce the emissions of acidifying
                                      agents.
                                      ACKNOWLEDGEMENTS: We want to express our gratitude to
                                      the  Swedish  Environmental Protection  agency, SNV-F,  for
                                      supporting this study, and the North American Lake Manage-
                                      ment Society for supporting the presentation  of this paper.
REFERENCES

Bergstrom, S., and G. Sandberg. 1983. Simulation of ground
  water response by conceptual model- three case studies.
  Nordic Hydrol. 71-84.

Brown, D.J.A. 1982. The effect of pH and calcium on fish and
  fisheries. Water Air Soil Pollut.  18.

Gapon, Y.E.N. 1933. On the theory of exchange adsorbtion in
  soils. J. Gen. Chem. USSR. 3:144-60.

Meyer, T.A., and G.W.  Volk.  1952. Effect of particle size on
  soil reactions, exchangeable ions and plant growth. Soil
  Sci. 71 (1).

Robbins, C.W., J.J. Jurinak, and R.J. Wagenet. 1980. Calcu-
  lating cation exchange in a salt transport model. Soil Sci.
  Soc. Am. J. 44:1195-1200.

Schofield, C.L, and  J.P. Baker. 1982. Aluminum  toxicity to
  fish in acidic waters. Water Air Soil Pollut.  18.

Schollenberger, G.J., and R.M. Salter. 1943. A chart for
  evaluation of agricultural  limestone. J. Am. Soc. Agron.
  35:955-66.

Sverdrup, H.U., and  I. Bjerle. 1982. Dissolution  of calcite
  and other related minerals in acidic aqueous solutions in a
  pH-stat. Vatten 38 (1): 59-73.
                                                  393

-------
                                                     Case  Studies
          of  Water  Quality Improvements
THE IMPROVED WATER QUALITY OF LONG LAKE FOLLOWING
ADVANCED WASTEWATER TREATMENT BY THE CITY OF
SPOKANE, WASHINGTON
RAYMOND A. SOLTERO
DONALD G. NICHOLS
Department of Biology
Eastern Washington University
Cheney, Washington
           ABSTRACT

           Long Lake, Wash., an impoundment of the Spokane River, has experienced high algal standing
           crop, low water clarity and extensive hypolimnetic anoxia during summer stratification The City
           of Spokane's primary sewage treatment plant was shown to be the primary contributor of
           phosphorus to the reservoir and the major cause of its eutrophic state. To reduce influent
           phosphorus (P) loading and improve Long Lake's water quality, the city provided advanced
           wastewater treatment (AWT) with chemical (alum) phosphorus removal in 1977. Monthly mean P
           load from the AWT plant has decreased approximately 90 percent and the overall load to the res-
           ervoir during the growing season (June-October) has declined about 74 percent. Mean post-AWT
           reservoir algal biovolumes and chlorophyll a concentrations are approximately 60 and 45 per-
           cent, respectively, less than pre-AWT values. A phosphorus load-chlorophyll a relationship,
           based on 5 years each of pre- and post-AWT data, was developed and provided excellent pre-
           dictions of mean reservoir chlorophyll a concentrations for the growing season. As a result of
           AWT, Long Lake has changed from a eutrophic to a mesotrophic body of water. In studying the
           effects of reduced P loading, it was determined that seasonal chemical phosphorus removal
           (April through October) could be as effective in reducing algal growth in Long Lake as year-
           around removal. This conclusion was based on the premise that temperature was the primary
           limiting variable outside the growing season. The city was granted a change in their AWT plant
           discharge permit and has implemented seasonal chemical P removal with no detrimental effects
           on the improved water quality of Long Lake.
INTRODUCTION

Long Lake, Wash., an impoundment of the Spokane
River, was formed by a concrete dam  in 1913 and is
used primarily for power generation. The reservoir is
approximately 18 km downstream from  the city of
Spokane and is 36 km in length with a mean depth of
14.6 m (Table 1). Maximum inflows generally occur be-
tween March and May with water retention times being
approximately 5 days. Minimum inflows usually occur
in August with water retention times in the order of 60
to 80 days. Nutrients, in particular phosphorus, dis-
charged to the Spokane River from Spokane's primary
sewage treatment facility were determined to be the
cause of excessive summer algal growth, low water
clarity,  and extensive hypolimnetic  anoxia during
summer stratification in Long Lake (Soltero et al. 1973,
1974, 1975, 1976, 1978). In compliance with directives
from the Washington State Department of Ecology to
reduce phosphorus (P) loading and improve the water
                                          395

-------
LAKE AND RESERVOIR MANAGEMENT
quality of the reservoir, the city constructed and put
"on-line" (1977)  an advanced wastewater treatment
(AWT) facility capable of removing at least 85 percent
P using aluminum sulfate (alum).
  This manuscript presents the impact of reduced P
loading on the reservoir's trophic status as a result of
the new AWT plant by comparing the water quality of
the reservoir over a 5-year period prior to AWT begin-
ning in 1972 with data covering the 5  years following
AWT starting in 1978.
MATERIALS AND METHODS

Midstream grab samples were usually collected semi-
monthly along the Spokane River at the effluent of the
treatment plant, Nine Mile, at the mouth of Little Spo-
kane River, and at the base of Long Lake Dam from
June through September and monthly for other months
(Fig. 1A). The phosphorus load at Nine Mile plus  thai
at the Little  Spokane River constituted the total load
to the  reservoir. Flow and discharge data were ob
tained from  the city of Spokane, the U.S. Geologica
Survey, and Washington Water Power Co. Water quali
ty data collected on June 2 and 16,1980, reflected the
effects of the May 18 eruption of Mt. St. Helens  anc
the subsequent area-wide volcanic ash cleanup. These1
values were therefore eliminated from all calculations
  Sample frequency for the  reservoir was  generally
the same as that for the river stations. Samples were
taken at 3 m depths throughout the water column al
each station (Fig.  1B). A euphotic zone composite fot
phytoplankton and chlorophyll a  analyses  was alsc
taken at each station by combining equal volumes ol
water usually taken at 2 m intervals from the surface
to the lower  limit of the euphotic zone as determined
by a submarine photometer.
  All analyses were usually conducted within 24 hours
after collection as described by Soltero et al. (1983).

RESULTS AND DISCUSSION

Mean daily total phosphate (TP) loads from the treat
ment plant effluent have decreased from a mean  pre
AWT value of 1.61 metric tons to a post-AWT value ol
0.20 metric tons for the period of June through Oc
tober, the growing season for Long  Lake  (Table 2)
Mean orthophosphate (OP) loads have also declinec
from 1.12 to 0.09 metric tons day-1.  Post-AWT loac
values for both phosphorus (P) fractions were found tc
be significantly lower (P = 0.01) than their respective
pre-AWT values. These P load reductions at the planl
have effected a 74 percent reduction in the TP load tc
the reservoir. In addition, the reduced treatment plant

  Table 1.—Morphometric data for Long Lake, Wash, at
         maximum capacity (elevation 468.3 m)
Maximum length
Maximum effective length
Maximum width
Maximum effective width
Mean width
Maximum depth
Mean depth
Area
Volume
Shoreline length
Shoreline development
Bottom grade
 35.4 km

  5.8 km

  1.1 km

  1.1 km

571.8m   -

 54.9m

 14.6 m
208.4 x 105 m2

3049 x 106 m3

 74.3 km

  4.6

  0.15%
                  effluent loads have caused  a  shift  in  the  principal
                  source of phosphorus to the reservoir from the treat-
                  ment plant during the pre-AWT years to the Spokane
                  River during  all of the post-AWT years (Soltero et al
                  1982).
                    Total Kjeldahl  and  total  inorganic  nitrogen
                  (TIN = NO3 - N + NO2 - N + NH3 - N)  concentrations
                  for treatment plant effluent have declined 87 and 25
                  percent, respectively, with  AWT (Table 3).  These
                  changes correspond to a 37 percent  reduction in the
                  total nitrogen load and a 14 percent  reduction in the
                  TIN load to the reservoir.
                    The  effect of reduced  nutrient  loading  on  the
                  phytoplankton community of Long Lake is shown in
                  Figures 2 and 3. Mean reservoir values for two time
                  periods are shown for each study year: June through
                  October (solid bar) and July through September (stip-
                  pled  bar) when recreational  use of  the reservoir is
                  generally the highest. Mean reservoir chlorophyll  a
                  values  for both seasons declined approximately 45
                  percent following AWT (Fig. 2). The mean post-AWT
                  values  for  June-October (8.03  mg  m-3) and
                  July-September (7.83  mg 3)  are  significantly lower
                  (P = 0.01) than their respective pre-AWT values. Also,
                  mean summer chlorophyll a concentrations were less
                  than 10 mg m-3, a value which has been suggested to
                  be the lower limit of eutrophy (U.S. Environ. Prot. Agen-
                  cy, 1973; Jones and Lee, 1980).
                    The mean  pre-AWT (1972-1977) phytoplankton bio-
                  volumes for both periods of time were 8.44 (June-Oc-
                  tober) and 9.74 (July-September) mm3 1-1 (Fig. 3). Max-
Figure 1.—Map of the lower Spokane River system detailing
the study area.
                                                396

-------
imum values for both periods occurred during low flow
years (1973 and 1977). In 1978, an unexplained  large
pulse of Microcystis aeruginosa occurred (confined to
the upper end  of the reservoir) distorting  calculated
phytoplankton  standing crops. Since then, the mean
biovolume  for both periods  decreased  to ap-
proximately 3.0 mm31~1 . This represents a decline in
excess of 60 percent from pre-AWT values.
   The response of the four major algal classes in Long
Lake to reduced nutrient loading is shown in Figure 4.
Diatoms dominated  the  phytoplankton community
during all pre-AWT years with a mean contribution of
62 percent. The greens and cryptomonads each contri-
buted  16 and 9.5 percent, respectively, while the blue-
greens contributed 5.4 percent. Since 1978,  the diatom
contribution has increased to 75 percent. The percent
contribution by the greens and cryptomonads (9.3 and
5.9 percent, respectively)  declined to approximately
one half their  pre-AWT values while the  blue-green
contribution (5.9 percent) has remained essentially un-
changed.
   The reductions in phytoplankton standing crop have
coincided with an increase in the euphotic zone TIN:OP
ratio brought about by AWT. The mean euphotic zone
ratio has increased from a pre-AWT value  of 6.0 to a
post-AWT value of 17.2. Miller et al. (1975) have shown
                             CASE STUDIES OF WATER QUALITY IMPROVEMENTS


                     through algal assay that water having a TIN:OP value
                     greater than 11.3 can be considered to be phosphorus
                     limiting to algal growth. Algal assay data prior to AWT
                     showed  Long Lake was  primarily nitrogen  limited
                     (Soltero et al. 1976,1978) while post-AWT data showed
                     the reservoir had changed to a phosphorus limited
                     system (Soltero et al.  1979).


                     The Blue-green Algae: Past and Present
                     Most of the recreational use and home  site develop-
                     ment along Long Lake has taken place in the upper
                     end (at or near  stations 3 and 4). Also, most of the
                     blue-green blooms that have occurred since 1972 have
                     been confined to one or both stations  (Soltero and
                     Nichols, 1981). A 1976 toxic bloom of Anabaena flos-
                     aquae occurred  primarily at station 4 and a  smaller
                     pulse of the same  species occurred  throughout the
                     reservoir in 1977. Only one mid-reservoir bay sample of
                     the 1977 bloom was shown to be toxic. In 1978, an ex-
                     tensive nontoxic bloom of Microcystis aeruginosa
                     also occurred in  the upper end of the reservoir (Soltero
                     et al. 1979). Following 1978, less severe pulses of non-
                     toxic blue-green forms, primarily Anabaena spiroides,
                     Anabaena circinalis, and Anabaena sp. have persisted.
                     These late summer pulses have generally decreased
Table 2.—Sewage effluent mean daily load (metric tons) of total Kjeldahl nitrogen (TKN) and total inorganic nitrogen
      (TIN = NOJ.N + NO2 + NH3-N) before (1972-1977) and after (1978-1982) AWT for June through October.

1972
1973
1974
1975
1977
X
1978
1979
1980
1981
1982
X
June
0.97
208
1.30
1.45
1.11
1 38
061
007
075
0 18
051
042
July
1.84
205
2.58
201
1 77
205
094
0 10
0 18
022
0.11
031
Aug.
TKN
1 47
2.53
1.92
257
1 35
197
029
011
007
007
011
0 13
Sept.
1.64
2.59
1.68
2.02
080
1 75
054
009
0.07
021
012
021
Oct.
195
3.24
2.09
3.50
0.21
2.20
0.55
0.09
0.10
011
017
0.20
June
097
1.55
1.03
1.07
1 31
1.19
1.21
064
0.83
1 43
0.83
099
July
1.36
1.46
1.53
1.54
1.09
1.40
1.48
067
074
1.02
0.61
0.90
Aug.
TIN
1 27
1.52
1.42
205
1 11
1 47
1.15
097
1.03
1 12
1.00
1.05
Sept.
1.27
1.67
1.31
1.63
1 02
1 38
1 14
1 23
1 14
1 11
083
1 09
Oct.
1.52
2.03
1.53
255
0.89
1 70
1 97
1 34
077
1 57
099
1 33
 Table 3.—Sewage effluent mean daily load (metric tons) of total phosphate (TP) and orthophosphate (OP) before (1972-1977)
                              and after (1978-1982) AWT for June through October.

1972
1973
1974
1975
1977
X
1978
1979
1980
1981
1982
June
1 36
2 11
085
1.27
1.27
137
0.06
0.15
0.78
0.14
018
July
1 79
202
1 31
1.76
1 00
1.58
004
0.25
0.35
0.14
0.17
Aug.
TD
1 66
1 86
159
2.14
1.15
1 68
0.05
023
0.15
0.12
0.17
Sept
1 84
1 77
1 53
1.87
1 53
1.71
012
0.25
0.28
0.15
020
Oct.
1.82
1.86
1.70
215
1.11
1 73
0.21
0.45
0.17
011
0.18
June
085
1 39
0.56
070
0.63
0.83
001
0.07
0.46
0.02
0.05
Julv
1 23
1.45
0.85
097
0.59
1 02
0.01
009
0 16
003
0.04
Aug.
OP
1 05
1.44
0.95
1 36
091
1 14
002
0.12
006
003
007
Sept.
1 38
1 33
092
1 34
1 35
1 26
009
0.11
0.18
0.06
0.10
Oct.
1 46
1 34
1 10
1 65
1 09
1 33
0.15
0.16
007
004
0.07
              0.26
                       0.19
0 14
0.20
0.22
0.12
0.07
                                             006
0.11
                                                      010
                                                  397

-------
 LAKE AND RESERVOIR MANAGEMENT
 water clarity and aesthetic appearance of the water
 especially in bays and on windward beaches.
  A reduction or elimination of the blue-green pulses
would increase the recreational potential of the reser-
voir, especially  in the upper end of the reservoir (sta-
tions 3  and 4) where most of  the houses  have been
built. Table 4 shows that  the combined  blue-green
contribution values  for stations 3 and 4 exceeded
those for stations 0 through 2 for 6 out of the 10 stud/
years. In addition, the mean combined blue-green cor-
tribution for stations 3 and 4 during the pre-AWT years
was approximately 6 percent. Since the 1978 pulse of
M. aeruginosa, the blue-green population has declined,
but not to pre-AWT levels. The mean combined contri-
bution for stations 3 and 4 between 1979 and 1982 for
June-October was 14.2 percent. This increase occur-
red despite the  74 percent decrease in TP  loading to
the reservoir.
  Soltero and Nichols (1981) explain the blue-green in-
crease as a result of reduced heavy metals (primarily
zinc)  entering the reservoir from  discharges  of the
north Idaho mining district. The  zinc reduction resulted

 Table 4.— Percent contribution by blue-greens to the total
  phytoplankton biovolume of each Long Lake sampling
     station for all study years, June through October.

                      Station
Year
1972
1973
1974
1975
1977
1978
1979
1980
1981
1982
0
4.4
13.0
0.7
1.5
16.2
2.4
9.1
0.9
0.5
87
1
3.9
1.2
1.2
1.8
10.3
1.1
3.2
0.7
2.6
4.1
2
3.3
1.3
0.6
1.2
15.8
20.4
4.6
4.3
10
8.2
3
4.3
0.7
02
29
12.9
89.9
7.3
6.5
5.9
12.4
4
8.8
3.5
0.1
1.4
27.1
98.9
35.9
7.5
12.8
24.7
 from federally imposed pollution abatement programs
 in the mining and smelting district near Kellogg, Idaho.
 Trend analysis of zinc concentrations in the Spokane
 River has shown  a 50 percent reduction near  Post
 Falls, Idaho, between 1973 and 1978 (Yake, 1979).  Prior
 to this  cleanup, zinc concentrations  in the  reservoir
 were algicidal  and/or algistatic. Algal assays using
 Selenastrum capricornutum (Green et al. 1975, 1976)
 and A. flos-aquae (Shiroyama et al. 1975) showed  max-
 imum growth yields from Long Lake water only  after
 heavy metals were removed (EDTA chelation). These
 yields were shown to be excellent predictors of reser-
 voir phytoplankton biovolumes and chlorophyll a  con-
 centrations.

 Areal P  Loading and  Phytoplankton
 Productivity

 Areal TP  loading  to the reservoir  has significantly
 (P = 0.01) declined  with reduced loads from the treat-
 ment plant effluent (Table 5). In-lake OP  concentra-
 tions have declined 73 percent. Chlorophyll  a values
 have also changed with post-AWT mean values being
 significantly (P = 0.01) less than the  pre-AWT mean
 values.
  A Vollenweider (1976) mean in-lake TP/influent  TP-
 mean depth to hydraulic residence time diagram  was
 used (Fig. 5) to determine the reasonableness of  esti-
 mated TP load values entering the reservoir. Load
 values plotted  well  within  the acceptable limits of
 reasonableness (±2X) as described by Past  and  Lee
 (1978). This suggests that the calculated  values for
 Long Lake during the June-October growing season
 can  be considered to be both reasonable and conser-
 vative. Values for in-lake TP concentrations from 1972
 through 1980 were not available. Therefore, TP values
 at Long Lake Dam were used to represent mean in-lake
TP values  during these study years. A comparison be-
tween monthly in-lake TP values with those at the Long
 Lake Dam when both were  available (1981 and 1982)
showed  no significant (P = 0.01) differences.
  A  Vollenweider (1976) normalized P load-chlorophyll
a relationship (r = 0.96) was developed (Fig. 6) to evalu-
   c
   o
   o
   ml
   ~>-
   £
   a.
   o
   _o

   6

   'o
   
-------
                                                             CASE STUDIES OF WATER QUALITY IMPROVEMENTS
ate the trophic response of Long Lake to reduced  P
loading. The relationship was also developed to create
a  water  quality  management  tool  to assess  the
response of the reservoir to future possible changes in
influent P load. The  adoption of Vollenweider's nor-
malized P load-trophic response approach was based
on its good  predictive success and applicability to a
wide variety of lake types (Lee et al. 1977; Jones and
Lee, 1978; Lee and Jones, 1980; Newbry et al. 1980;
Horstman et al. 1980; Ciecka et  al.  1980; Past et al.
1983; Ortiz et al. 1981). The regression equation for the
Long Lake P load-chlorophyil a  relation is given by:

     log Y = 0.503 log (X) + 0.298  r2 = 0.92  (1)

where, Y is the mean  euphotic zone chlorophyll a con-
centration  (mg  m  3)  and X   is  the expression
(Lp/qs)/(1 + \fZ/qs)  for the June-October growing
season. Predicted chlorophyll a values from Eq. 1 were
not significantly (P = 0.01) different from the actual
June-October values (Soltero et al. 1983).
   Several researchers have found a good correlation
between mean summer and maximum seasonal chlor-
ophyll a values (Ryding, 1980; Fricker, 1980; Jones et
al. 1979). A similar one was found for Long Lake during
the June-October growing season over all study years.
This relationship is defined by:

          Y = 3.75 (X)-7.14   r2 = 0.77  (2)

where, Y is the maximum and X is the mean euphotic
zone chlorophyll  a concentration,  respectively. This
equation points out the need for a potential user of the
Long Lake P load-chlorophyll a  relationship to recog-
nize that a maximum seasonal chlorophyll a value oc-
curs along with a chosen mean chlorophyll a  value.
For example, if a mean summer chlorophyll a concen-
tration of 10 mg m-3 is chosen to be the upper limit of
acceptable water quality, then a maximum seasonal
chlorophyll a value of approximately 30 mg m-3 can be
expected to occur. Seasonal  maximum values ex-
ceeding 32 mg m-3 occurred in  the reservoir during
1973,  1975, and 1977 (Soltero et al. 1974, 1976, 1978).
These pulses occurred in the upper end of the reservoir
(stations 3 and 4) and were associated with high phy-
toplankton  standing  crops (>40 mm-3 1-1) and low
water clarity (Secchi disk < 1m).
  A TP load-hydraulic load curve (Fig. 7) was used to
determine if reduced  P loading to the reservoir had ef-
fected a change in trophic state.  The permissible and
excessive loading lines that are normally based on in-
lake P concentrations of 10 and  20 m-3, respectively,
were changed  to values (15 and 30 mg  m-3) which cor-
responded to the simplified 95 percent confidence in-
terval values of in-lake P for  a chlorophyll a of  10 mg
m-3 (Fig. 6). A similar type of  change in the Vollen-
weider load curve was made by Kratzer (1979) to better
describe the trophic condition of the Florida National
Eutrophication Survey Lakes. The Vollenweider load
curve (Fig. 7) shows that Long Lake has changed from
a eutrophic state during the pre-AWT years to a more
mestrophic condition during  the post-AWT years. The
largest vertical distance above the excessive load line
occurred during the flow years of 1973 and 1977. Rast
and Lee (1978)  used the vertical distance above the ex-
cessive line to measure the degree of eutrophy.
  The estimated trophic state of Long Lake for a given
P load under certain flow  conditions (hydraulic load)
using Vollenweider's load  curve does not provide any
measure of reliability of the predicted trophic state.
Knowledge of this reliability would allow one to weigh
the value of a  predicted  trophic state.  Figure 8 shows
the results of  a trophic  state probability analysis for
          1972
                                      1975
                                                1977
                                                         1978
                                                                   1979
                                                                                     1981
                                                                                              1982
Figure 3.—Mean daily phytoplankton biovolume (mm-3 1 ~1) for the periods June through October (solid bars) and July
through September (stippled bars) for all study years, Long Lake, Wash.
                                                  399

-------
 LAKE AND RESERVOIR MANAGEMENT
  Long Lake following Reckhow (1979) and Federico et
  al. (1981). It  was assumed  that the prediction errors
  were normally distributed. Also, since P load values
  were determined from measured P concentrations and
  river flows, uncertainty associated with the load values
  could be considered to be zero (Reckhow, 1979). Theie-
  fore, the total prediction uncertainty would be equal to
  the error associated with the model (Eq. 1). The stan-
  dard model error using the Vollenweider normalized P
  load expression was determined to be 11.5 mg P m-s.
    The estimated mean in-lake P concentration for the
  pre-AWT study years during the growing season as
  determined from the Vollenweider P  load expression
  was 55.3 mg m-3. This value corresponds to a 99 per-
  cent probability that Long Lake was eutrophic during
  these  years (Fig.  8). Conversely,  the  post-AWT
  estimated in-lake P concentration of 15.3 mg m-3 is
  associated with a 10 percent probability of eutrophy
  and a 41  and 49  percent of mesotrophy and oligo-
  trophy, respectively.
    Using  this uncertainty analysis along with the P
  load-trophic response  relationships  (Figs.  6  and 7),
  Long Lake can be classified with reasonable certainty
  as being eutrophic for all pre-AWT years during the
               Bacillanophyci
Q.  100-
  100_
         T	1	T
1	1	1
               Chlorophyceae
                   1	1	1    I    I	1	1
              Cryptophyceae
              Cyanophyceae
~T
 72
I
73
 I
74
I
75
 1 - 1
 77   78
Year
 1
79
        1
        80
                                             \
                                            81
Figure 4.— Percent contribution to the total estimated phyto
plankton biovolume (mm-3 1~1) by the four major alga
classes found in Long Lake over all study years for the perioc
June through November.
                        June-October growing season. Reduced P loading as
                        a result of the AWT plant has improved the overall
                        water quality of Long Lake to a more mesotrophic con-
                        dition.
                          Housing and industrial development along the Spo-
                        kane  River  upstream  from Long Lake in an  area
                        bordered by the city of Spokane and the State of Idaho
                        (the Spokane  Valley)  is  increasing.  In addition, the
                        Spokane-Rathdrum aquifer flows westward through
                       la.
                       Figure 5.—An evaluation of Long Lake TP loading estimates-
                       Vollenweider  mean in-lake  TP/influent TP-hydraulic
                       residence time relationship. Mean in-lake TP concentrations
                       based on Long Lake Dam values are designated by the solid
                       dots, whereas the triangles represent measured in-lake TP
                       values.
                                                                                       tn'3>
                      Figure 6.—Phosphorus load, normalized by mean depth and
                      hydraulic load, versus mean chlorophyll a concentration for
                      Long Lake, Wash. (June through October). The dashed line
                      represents the 95 percent confidence interval.
                                                  400

-------
the valley and has been designated by the U.S. Envi-
ronmental  Protection  Agency  as  the  sole  source
domestic water supply for the metropolitan Spokane
area (population of approximately 340,000). Protecting
this aquifer from septic tank drainfield contamination
has been of prime concern. A step towards protecting
the aquifer would be to collect and treat domestic and
industrial wastewater from the valley, and eventually
discharge the treated water to the Spokane River.
  However, concern has also been raised as to  how
much more P can be discharged to the Spokane River,
over and above that currently being  added, without
seriously affecting Long Lake's improved water quali-
ty.  The Long Lake P load-chlorophyll a relationship
was used to compare actual and estimated load values
for  given flow conditions to determine if additional P
could be added to the reservoir. Table 6 presents the
information needed to make this determination.
  The mean chlorophyll a values chosen for this anal-
ysis were based  on values considered by some to be
indicative of eutrophic conditions (Jones et al. 1979).
The hydraulic load values of 43  and 64 m yr-1 repre-
sent mean daily  river flows during the June-October
season that can be expected  to be exceeded 95 and 80
percent of the time, respectively (Soltero et al. 1983).
Estimated  P load values were determined from the P
load-chlorophyll a relationship (Eq. 1) for given chloro-
phyll  a and hydraulic load (qs) values. The  residual
                                                CASE STUDIES OF WATER QUALITY IMPROVEMENTS

                                       value is  the  difference between the  estimated and
                                       mean post-AWT load value and, therefore, represents
                                       the amount of P, if any, that could be added to Long
                                       Lake.
                                         For a qs value of 64 and a mean chlorophyll a of 10
                                       mg m-3, 168 Ibs. P day-1 could be added to the reser-
                                               I   ]  I I  J I I M     I   I I  I I I I I
                                                          10
                                                            Hydraulic Load {m yr'1)
                                       Figure 7.—Vollenweider (1976) P loading curve for all study
                                       years and changes in trophic status of Long Lake as a result
                                       of reduced P input for the period June through October.
 Table S.—Specific areal total phosphorus loading (L_), mean orthophosphorus, and chlorophyll a concentrations for all study
                            years for the period June through October, Long Lake, Wash.
Year
1972
1973
1974
1975
1977
X
1978
1979
1980
1981
1982
X
(gPm-2)
7.49
6.03
6.87
7.01
438
6.36
1.58
1.77
1.39
2.03
1.63
1.68
Lp
(Ibs. P day-i)
1879
1513
1723
1758
1099
1594
396
444
349
509
409
421
Orthophosphorus
(M9P1-1)
43.0
97.8
32.3
31.6
83.8
57.7
17.6
14.3
17.6
15.0
14.3
15.8
Chlorophyll a
(mg m-3)
12.90
19.86
11.72
12.84
15.23
14.51
9.54
10.09
6.56
7.97
6.00
8.03
 Table 6.—Predicted specific total phosphorus loads influent to Long Lake for a given mean chlorophyll a concentration and
 hydraulic load using the Long Lake phosphorus load chlorophyll a relationship (Eq. 1) and the reservoir's capacity for addi-
                                tional phosphorus loading, June through October.
    'Mean
 chlorophyll a

   (mg m-3)
      10
      12
       8
      10
      12
(m yr'1)

  64
  64
  64
  43
  43
  43
(gPm-2)
1.51
2.35
3.38
1.09
1.69
2.43
(myr day-1)
379
589
848
273
424
610
Pm-2)
1.68
1.68
1.68
168
1.68
1.68
LP
(Ibs P day
421
421
421
421
421
421
 'Residual
(gPm-2)

  -0.17
   0.67
   1.70
  -0.59
   0.01
   0.75
    an values that represent lower limits of eutrophication (Jones et al  1979)
 ^Hydraulic load, 43 and 64 represent low flow years
 ^Estimated P load using the Long Lake regression eguation (Eq 1) for a given mean chlorophyll a and hydraulic load value
 4Overail mean post-AWT specific load value (Table 5)
 Difference between estimated and mean post-AWT Lp value
(Ibs P day-i)

     -42
     168
     427
   -1.48
       3
     189
                                                    401

-------
  LAKE AND RESERVOIR MANAGEMENT
  voir over the mean post-AWT load (Table 6). Likewise,
  an additional 427 Ibs. P day-1 could be added to the
  reservoir for a mean chlorophyll a concentration of 12
  mg m-3. A lower river flow (qs = 43) resulted in smaler
  residual P load values: only 3 additional Ibs. P day-1
  for a chlorophyll a concentration of 10 mg m-3 and 139
  Ibs. P day-1  for a chlorophyll a value of 12 mg m-3.


  Seasonal Chemical P Removal

  Spokane's AWT plant provided year-around chemical
  P removal from December 1977 to late 1981. Results of
  two investigations on the effect of reduced P loading
  to the reservoir (Gasperino and Soltero,  1977; URS,
  1981) concluded that seasonal chemical P removal at
  the AWT facility could be as effective in maintaining
  acceptable water quality in Long Lake as would year-
  around removal. Prompted by these findings, Spokane
  applied  for and  received  permission (1981) from the
  Washington  State Department of  Ecology to modify
  its  National  Pollution Discharge  Elimination  System
  (NPDES) discharge  permit  to  allow  for seasonal
  chemical P removal to begin on April 1 and terminate
  by Nov.  1. The April 1 startup date was chosen by the
  State Department of  Ecology to  be a  conservative
  estimate of the beginning of the growing season for
  Long Lake based on phytoplankton  biovolumes  and
  chlorophyll a values observed during the low flow
  years of  1973  and  1977.  The  seasonal  removal
  schedule began Nov. 1, 1981.
   The discharge permit change required additional
  monitoring of  the reservoir (2 years) to detect  any
 adverse  effects from  this operational change at  the
 AWT plant on  Long Lake's improved water qualit/.
       Estimated In-Lake Phosphorus Concentration
       (mg P rrf3)
Figure 8.—Trophic state probability plot for Long Lake based
on estimated in-lake phosphorus concentration.
  Nearly 2 years of monitoring following the initiation of
  the seasonal P removal schedule has revealed essen-
  tially no impact  on  the  reservoir's  water  quality
  (Soltero et al. 1983, unpubl.). The mean in-lake TP con-
  centration for the period  when chemical P removal
  was turned off (November-March) was similar to that
  found during year-around  removal and substantially
  less than the pre-AWT value for the same time  period.
  The euphotic zone TIN:OP ratio  resembled all other
  post-AWT values.  Phytoplankton succession  and
  biovolumes  for  the   growing  season  during  the
  seasonal  P removal periods did not appear to differ
  from other post-AWT years.
    It has been suggested (URS, 1981) that if a method-
  ology for predicting spring  runoff to the Spokane River
  was developed, a variable startup date for seasonal P
  removal might be  possible. Mires and Soltero (1983)
  showed that a recently updated U.S. Soil Conservation
  Service  equation  (Beard,  pers. comm.)  to  forecast
  runoff to the Spokane River at Post Falls, Idaho, pro-
  vided good estimates of flows in  the Spokane River
  and to  Long  Lake.  Using a relationship  between
  hydraulic and P residence times (Sonzogni et al. 1976),
  Mires and Soltero developed a methodology to deter-
  mine a startup date for chemical P removal other than
  the fixed date of April  1. This same basic approach
  along with the consideration of  other factors that
  might limit algal growth during the latter part  of the
  growing season was used  to develop a methodology
  to vary the Nov. 1 termination  date (Mires et al. 1983).
   Neither the startup nor termination date modifica-
 tion models have been  put into practice  at the AWT
 plant as of yet. However, Spokane  will be applying to
 the State Department of Ecology for another change
 in their NPDES discharge permit that would allow for
 these  modifications.  If a   change is  granted,  the
 Department may  require continued monitoring  of the
 reservoir to determine if a varying P removal scheme
 will adversely  affect Long Lake's improved  water
 quality.
 CONCLUSIONS

 Since  Spokane's AWT facility  began  chemical P
 removal the TP load to Long Lake has been reduced by
 74 percent. Less P loading from the treatment plant
 has caused the Spokane River to become the primary
 contributor of P to the reservoir during all post-AWT
 years. These  reduced P loadings have coincided with
 a 60 and 45 percent reduction in mean reservoir phyto-
 plankton biovolurne and chlorophyll a concentration,
 respectively. Diatoms have contnued to be the primary
 algal class during all study years. The percent con-
 tribution to the total algal biovolurne by the greens
 and cryptomonads has declined approximately 50 per-
 cent since AWT while the blue-green contribution has
 remained essentially unchanged. Late summer pulses
 of blue-greens continue to occur in the upper end of
 the reservoir.  Since 1978 these pulses have not been
 comprised  of  potentially toxic species, whereas toxic
 species occurred during some pre-AWT years.
  A good correlation (r- 0.96) was found between nor-
 malized P loads and chlorophyll a concentrations us-
 ing  10  years  of data for the June-October growing
 season. Chlorophyll a values calculated from this rela-
tionship were  not found to significantly differ from ac-
tual values. A  modified Vollenweider P load-hydraulic
load  curve  and a trophic  state probability analysis
showed that Long Lake has changed from eutrophic
to  mesotrophic srnce  the  initiation of  AWT. The P
                                                402

-------
                                                                  CASE STUDIES OF WATER QUALITY IMPROVEMENTS
load-chlorophyll a  relationship was used along with
conservative flow characteristics of the Spokane River
to show that additional P could be added to the reser-
voir, over and above the mean post-AWT load levels to
attain a mean chlorophyll a concentration of 10 mg
m-3 for the growing season.
   The seasonal P removal schedule (April 1-Nov. 1) at
the AWT plant does not appear to have affected the
improved water  quality of Long  Lake. Models have
been developed  to  make the startup and termination
dates of the seasonal P removal schedule more flexi-
ble  and  dependent  upon river  flow and  reservoir
physical/chemical characteristics, but as of yet, have
not implemented.

ACKNOWLEDGEMENTS: The work upon which this report is
based was supported in  part by a grant from the  city of
Spokane,  Wash.  Glen A. Yake,  Project Coordinator, John
Swanson, Director of Utilities, and Dan Robison, Director of
Environmental  Programs  for  the  City,  are  gratefully
acknowledged. The cooperation  extended by Washington
Water Power Company in making available daily reservoir
discharge records for Nine Mile and Long Lake dams is ap-
preciated. Thanks are given to Dale Arnold, Chief Chemist of
the city's Advanced Wastewater Treatment Plant, for making
available plant records. Thanks are given to Greg Ruppert,
Hydrologist, Water Resources Division of the U.S. Geological
Survey,  Spokane,  for  making  available  daily  discharge
records for  the Spokane  and Little  Spokane rivers, and
Hangman Creek.  Sincere thanks are extended  to Mary
Cather and Kim McKee for their assistance in the field and
the laboratory and to Carol  Harmon for her typing  of this
manuscript.
REFERENCES

Beard,J. Pers. comm. Snow Surv. Supervisor, U.S. Soil Con-
  serv. Sen/., Boise, Idaho.
Ciecka, J., R.  Fabian, and D. Merilatt. 1980. Eutrophication
  measures for small lake water quality management. Water
  Resour. Bull. 16: 681-9.
City  of  Spokane. Wastewater Treatment  Plant.  N. 4401
  Aubrey White Parkway, Spokane, Wash.
Fedenco, A.C., K.G. Dickson, C.R.  Kratzer,  and F.E. Davis.
  1981. Lake Okeechobee water quality  studies and eutro-
  phication assessment.  Resour. Plan. Dep. S.  Fla. Water
  Manage. Dist., West Palm Beach. Tech. Publ. 81-2.
Fricker, Hj. 1980. OECD eutrophication programme regional
  project: Alpine lakes. Swiss Fed.  Board Environ. Prot.,
  Bern.
Gasperino, A.F.,  and R.A. Soltero. 1977. Phosphorus reduc-
  tion and its effect on the recovery of Long Lake reservoir.
  D.O.E Proj. No. 77-108.  Final Progr. Rep.
Greene, J.C., et al. 1975. The relationship of  laboratory algal
  assays to measurements of indigenous phytoplankton in
  Long Lake, Wa. Pages 93-126 in E.J. Middlebrooks, et al.,
  eds. Biostimulation and Nutrient Assessment Workshop.
  Ann Arbor Science, Mich.
	. 1976. Use of algal assays to assess the effects of
  municipal and smelter  wastes upon phytoplankton  pro-
  duction. Pages 327-36 in  R.D. Andrews,  et al. eds.  Ter-
  restrial and Aquatic Ecological Studies of the Northwest.
  Eastern Washington State College Press,  Cheney.
Horstman, H.R., G.F. Lee,  and R.A. Jones. 1980. Eutrophica-
  tion-related  water quality in Dillon reservoir. Dep. Civil
  Eng. Colo. State  Univ.,  Fort Collins. Proc. Environ. Eng.
  Conf., New York.  Mimeo
Jones, R A., and G.F. Lee. 1978. Evaluation of the impact of
  phosphorus removal in the Danbury, Conn, sewage treat-
  ment plant on  water quality in Lake Lillinonah. Dep. Civil
  Eng Colo. State Univ., Fort Collins. Occas. Pap. No. 31
	. 1980. Recent advances in management of eutrophi-
  cation in water bodies. Dep. Civil Eng. Colo. State Univ.,
  Fort Collins. Mimeo.
Jones, R.A., W.  Rast, and G.F.  Lee. 1979. Relationship be-
  tween mean and maximum chlorophyll a concentrations in
  lakes. Environ. Sci. Technol. 13: 869-70.
Kratzer,  C.R.  1979. Application of input-output models to
  Florida lakes. M.E. Thesis, Univ. Florida, Gainesville.
Lee, G.F., and R.A. Jones. 1980. Recent advances in manage-
  ment of eutrophication in impoundments. Dep.  Civil Eng.
  Colo. State Univ., Fort Collins. Occas. Pap. No.  49.
Lee, G.F., W. Rast, and R.A. Jones. 1977. Recent advances in
  assessing  aquatic  plant  nutrient  load-eutrophication
  response for  lakes and  impoundments. Dep.  Civil Eng.
  Colo. State Univ., Fort Collins. Occas. Pap. No.  14.
Miller, W.E., J.C. Greene, and T. Shiroyama.  1975. Applica-
  tion of algal assays to define the effects of wastewater ef-
  fluents upon algal growth in  multiple use  river systems.
  Pages 77-92 in E.J. Middlebrooks, et al., eds. Biostimula-
  tion and Nutrient Assessment. Ann Arbor Science, Mich.
Mires, J.M., and R.A.  Soltero.  1983. Methodology  for as-
  sessing  the  initiation  date  of  chemical  phosphorus
  removal at Spokane's advanced wastewater  treatment
  plant  according  to  predicted  spring runoff. City  of
  Spokane,  Contr.  No. 414-420-000-534.40-3105. Comple.
  Rep. Eastern Washington Univ., Cheney.
Mires, J.M., R.A. Soltero, and D.G. Nichols. 1983.  Methodol-
  ogy for  assessing the fall  termination date of chemical
  phosphorus removal at Spokane's advanced wastewater
  treatment plant.  City  of  Spokane,   Contr.  No.
  4140-420-000-534.40-365.   Comple.   Rep.  Eastern
  Washington Univ., Cheney.
Newbry, B.W., R.A. Jones, and G.F. Lee. 1980. Assessment
  and analysis of eutrophication of Tennessee river system
  impoundments. Presented at Am. Soc. Civil  Eng. Symp. on
  Surface-Water Impoundments, Minneapolis,  Minn. June.
  Mimeo.
Ortiz, J.L., G.F. Lee,  and R.A. Jones. 1981. Development of
  basis  for establishing eutrophication  management pro-
  gram for Spanish impoundments. Pages 435-46  in H.G.
  Stefan, ed. Proc. Symp. Surface Water Impoundments.
  Am. Soc. Civil Eng., New York.
Rast, W., and  G.F. Lee. 1978. Summary analysis of the North
  American  (U.S. portion) OECD eutrophication  project:
  nutrient loading-lake  response relationships  and  trophic
  state indices. EPA-600/3-78-008. U.S. Environ. Prot. Agency,
  Corvallis, Ore.
Rast, W., R.A. Jones, and G.F. Lee. 1983. Predictive  capability
  of   U.S   OECD  phosophorus  loading  eutrophication
  response  models.  J. Water  Pollut.  Control  Fed. 55:
  990-1003.
Reckhow, K.H. 1979. Uncertainty analysis applied  to Vollen-
  weider's  phosphorus loading  criterion. J.  Water Pollut.
  Control Fed. 12: 2123-8.
Ryding, S. 1980. Monitoring of inland waters. OECD eutrophi-
  cation  programme: The  Nordic  project. Secretariat En-
  viron. Sci. Publ. 1980: 2. Helsinki, Finland.

Shiroyama, T, W.E. Miller, J.C.  Greene, and C.  Shigihara.
  1975. Growth response of Anabaena flos-aquae (Lyngb.)
  De  Brebisson in waters collected from Long Lake reser-
  voir, Wash. Pages 267-75 in R.D. Andrews,  et al. eds. Ter-
  restrial and Aquatic Ecological Studies of the Northwest.
  Eastern Washington State College Press, Cheney.
Soltero,  R.A., and D.G. Nichols. 1981. The recent blue-green
  blooms  of  Long Lake, Wash. Pages 143-59 in  W.W. Car-
  michael, ed. The Water Environment: Algal  Toxins  and
  Health. Plenum Publ. Corp., New York.

Soltero, R.A.,  A.F. Gasperino and W.G. Graham. 1973. An in-
  vestigation of the cause and effect of eutrophication in
  Long   Lake,   Washington.   O.W.R.R.  Project
  143-34-1OE-3996-5501. Comple. Rep. Eastern Washington
  State College, Cheney.
                                                     403

-------
LAKE AND RESERVOIR MANAGEMENT
	1974. Further investigation as to  the cause  Eind
  effect of eutrophication in Long Lake, Wash. D.O.E.  F'ro-
  ject 74-025A. Comple. Rep. Eastern Washington State  Col-
  lege, Cheney.

Soltero,  R.A., A.F. Gaspermo,  P.M.  Williams, and  S R.
  Thomas. 1975. Response of the Spokane River periphylon
  community to primary sewage effluent and continued in-
  vestigation of Long Lake. D.O.E.  Project 74-144. Comple.
  Rep. Eastern Washington State College, Cheney.
Soltero, R.A., D.G. Nichols, G.A. Pebles, and  LR. Singleton.
  1978.  Limnological investigation of eutrophic Long Lcike
  and its tributaries just prior to advanced wastewater treat-
  ment with phosphorus removal by Spokane, Wash.  D.C.E.
  Proj. 77-108. Comple.  Rep.  Eastern Washington  Univ.,
  Cheney.

Soltero, R.A.,  D.G. Nichols, G.P. Burr, and  LR. Singleton.
  1979. The effect of continuous advanced wastewater treat-
  ment by the City of Spokane on the trophic status of Long
  Lake, Wash. D.O.E. Proj 77-108.  Comple. Rep.  Eastern
  Washington Univ., Cheney.

Soltero, R.A., D.G. Nichols, and M.R. Gather. 1982. The effect
  of seasonal  phosphorus removal by the City of Spokane's
  advanced wastewater treatment plant on the water quality
  of  Long  Lake,  Wash. City of  Spokane,   Contr.  No.
  414-420-000-534.40-3105. Comple.  Rep. Eastern Washing-
  ton Univ., Cheney.

	. 1983. The effect of seasonal alum addition (chem-
  ical phosphorus removal) by the City of Spokane's advanc-
  ed  wastewater treatment plant  on the water quality of
  Long  Lake, Wash.  1982.  City  of  Spokane, Cont.  No.
  414-420-000-534.40-3105. Comple.  Rep. Eastern Washing-
  ton Univ., Cheney.
Soltero, R.A., et al. 1976. Continued investigation of eutrophi-
   cation in Long Lake, Wash.: Verification data for the Long
   Lake model.  D.O.E.  Proj. WF-6-75-081.  Comple.  Rep.
   Eastern Washington State College, Cheney.
Sonzogni, W.C., PC. Uttormark, and G.F  Lee. 1976. A phos-
   phorus  residence time  model:  Theory and application
   Water Res. 10: 429-35.

URS Co. 1981.  Spokane River wasteload allocation study
   Phase I. Comple. Rep. Seattle, Wash.

U.S. Environmental Protection Agency. 1973. Water quality
   criteria, 1972. Rep. Comm. Water Qual. Criteria. Nat. Acad
   Sci.  Eng. EPA R3.73/033, Washington, D.C.
U.S. Geological Survey. Federal Bldg.  Spokane, Wash.
Vollenweider, R.A. 1976. Advances in defining critical loading
   levels for phosphorus in lake eutrophication. Mem 1st Itai
   Idrobiol. 33: 53-83.

Washington Water Power Co. East 1411 Mission  Spokane
  Wash.

Yake, W.E. 1979. Water quality trend analysis—the Spokane
  River basin. Wash. State Dep. Ecol. D.O.E. Proj. 6 Olympia
  Wash.
                                                     404

-------
ECONOMIC RETURNS AND  INCENTIVES OF  LAKE
REHABILITATION:  ILLINOIS  CASE STUDIES
KRISHAN P.  SINGH
V. KOTHANDARAMAN
Illinois State Water Survey
Champaign, Illinois

DONNA F.  SEFTON
ROBERT  P. CLARKE
Illinois Environmental  Protection Agency
Springfield, Illinois
             ABSTRACT

             Studies were conducted on three eutrophic Illinois impoundments to determine practical and
             economically viable measures to reduce water quality related use impairments and evaluate the
             recreational benefits of those measures. As is typical in Illinois, these relatively shallow im-
             poundments were constructed in fertile floodplains draining croplands, have large watershed
             area to lake surface area ratios, and relatively short retention times. Frequent inflows of nutrient
             and sediment laden water contribute to problems of  hypolimnetic oxygen depletion, algal
             blooms, dense macrophyte growth, inorganic turbidity, and siltation. Even where the watershed
             has been controlled to the best practicable extent, runoff waters still carry nutrients in excess of
             eutrophic loading rates and significant amounts of sediment, and the lake remains eutrophic.
             Constraints imposed by morphologic, hydrologic, and watershed factors on lake quality were
             considered  in developing an integrated lake/watershed management strategy for each lake to
             preserve and maximize its use at minimal cost. In-lake techniques such as aeration/destratifica-
             tion, groundwater/surface water blending (dilution), lake shoreline stablization, weed harvesting,
             and algal control were not found to be palliative measures but essential management tools to
             preserve the lakes and allow their full development as  recreational resources. A  recreational
             benefit assessment was performed  for each project using the unit day value methodology
             recommended by the U.S. Water Resources  Council (18 CFR 713, Subpart K, App. 3. 1982).
             Benefit/cost ratios were determined in two ways: (1) by the ratio of the total discounted benefits
             to the requested Sec. 314 grant amount, and (2) by the annual recreational benefit divided by the
             sum of the  amortized capital costs and annual operation, maintenance, and repair costs. The
             recreational benefit assessment procedure appears to provide an excellent  tool for evaluating
             lake management strategies. The technical information  developed in these investigations may
             be applied to numerous other recreational and public water supply impoundments throughout
             the Nation.
INTRODUCTION

Illinois has over 2,900 lakes covering 77,298 ha and
more than 81,000 ponds (less than 2.4 surface ha)
covering 35,047 ha. Since 80 percent of Illinois water
bodies exhibit impaired use resulting from sediment,
macrophytes,  or algae,  and because resources for
lake protection/restoration  are very limited, special
studies were undertaken for three Illinois  impound-
ments to determine practical and economically viable
measures to reduce water quality related use impair-
ments and evaluate the recreational benefits of those
measures. These impoundments  were  selected for
their small, manageable size, relatively small  water-
sheds,  and  large  public  use  or benefits.  Each
impoundment  represented a different physiographic
region of the State (Fig. 1). Phase I diagnostic feasibili-
ty studies conducted in 1981-82  under  the  Clean
Lakes Program (Singh et al. 1983;  Kothandaraman et
al. 1983a,b) served as the basis for these case studies.
INTEGRATED MANAGEMENT STRATEGY
DEVELOPMENT

Constraints on the quality of Illinois lakes imposed by
hydrologic, morphologic, and watershed factors pro-
vided impetus for exploring integrated management
strategies to maximize the prime recreational useabili-
ty from May  through September. These constraints
are detailed in Sefton (1978), Sefton et al. (1980), and
Boland et al. (1979).
  Most Illinois lakes are artificially constructed in fer-
tile  floodplains  draining   croplands  and   exhibit
eutrophication symptoms soon after completion. Lake
quality is often governed by the physical bed of the
lake and the  short detention periods of incoming
watershed runoff containing high levels of eroded soil
and associated nutrients. Illinois lakes are also highly
productive, alkaline waters  capable of supporting
large populations of fish. Nuisance growths of aquatic
weeds and algal blooms are common as a result of fer-
                                                 405

-------
  LAKE AND RESERVOIR MANAGEMENT
  tile runoff inputs. Dissolved oxygen depletion in the
  bottom waters limits fish habitat, promotes release of
  nutrients and products of decomposition from the bot-
  tom  sediments, and  impairs  public  water supply
  usage. The better quality lakes are generally deeper
  and have longer water retention capacity.
    Because of the extensive row crop production in Il-
  linois and the fact that the watershed to surface area
  ratios of most lakes greatly exceed 20:1, sufficient
  control of watershed sources of  nutrients and sediment
  is often impossible. The practicality of changing la
-------
                                                            CASE STUDIES OF WATER QUALITY IMPROVEMENTS
water storage has decreased 13.5 percent since it was
constructed in 1948, representing an annual capacity
loss of 0.42 percent.  A major  portion of this sedi-
mentation occurred in the first  half of the lake's ex-
istence when the watershed was mostly agricultural.
  4. Algal  blooms.  Lake  of the  Woods  exhibits
periodic nuisance algal blooms. In 1981, the standing
crop of phytoplankton increased two-to-threefold in a
2-month period, achieving maximum standing crops of
23 to 27 million algal units per liter. Chlorophyll a con-
centrations ranged from Oto 136pg/l, with a mean con-
centration of 21.3 HQ!\.
  5. Excessive  aquatic  weeds.  Dense  growth  of
aquatic  macrophytes  (naiad,   pondweed,   elodea)
covers approximately 30  percent of the lake surface
area and interferes with boating, fishing, swimming,
and aesthetics. Plant growth generally follows  the
1.8-2.4 m bottom contour.
  6. Low water levels. Water level drawdown from the
use of lake water for golf course turf irrigation impairs
recreational use and adversely affects fish habitat in
the lake during droughts. A small capacity well near
the bath house helps  improve water quality in  the
swimming area during  normal years, but it is inade-
quate  to prevent water level drawdown during  dry
years.
  Goals to be Achieved. The following  goals were to
be achieved in the restoration of Lake of the Woods:
(1) dissolved oxygen of at least 5 mg/l throughout the
lake, (2) Secchi disk transparency of not less than 1.22
m,  (3) TP  less than 0.05 mg/l at all times, (4) total
suspended solids less than  10 mg/l, (5) limited aquatic
macrophyte growth, (6) no nuisance algae blooms, (7)
suitable turf irrigation system without lake level draw-
down, and (8) improved fishing  and recreational ex-
perience.
   Lake Protection/Management Plan. A critical review
of the efficiency and cost of feasible alternatives for
improving and maintaining the  water quality of Lake
of the Woods and maximizing its recreational poten-
tial resulted  in  recommending  the  following best
management  system.
   1.  Soil  conservation plan for 57 ha of cropland to
reduce sediment and nutrient input by  50 percent: no
cost to project.
   2.  Management plan to  minimize nutrient inputs
from other watershed sources (pasture,  residential
areas, and golf course). Although golf course fertiliza-
tion is only a minor contributor to nutrient input (7.5
percent of the TP and 6 percent of the TN), the golf
course fertilizer management plan should minimize in-
put from this source.  A  buffer strip of  unmowed
natural vegetation between the golf course and the
lake also  should be maintained: no cost to project.
  3.  Ground  water/surface  water  blending system
(600  gpm well) to dilute lake water with high quality
ground water (which has lower turbidity and nutrient
concentrations  than the inflowing streams) and to
help maintain water levels. Initial cost $57,000; opera-
tion, maintenance, and repair $1,800/year.
  4.  Aeration/destratification to improve dissolved
oxygen  levels with associated  benefits: initial cost
$15,000; operation, maintenance, and repair $560/year.
  5.  Periodic applications of chelated copper sulfate
followed by potassium permanganate for algal con-
trol:  $650.
  6.  Applications of Aquathol and Komeen twice per
year for aquatic weed control: $1,000.
  7.  Aquascreens to control nuisance weed growth in
the water slide and boat ramp areas and ensure their
safe operation: initial cost $2,670.
  Cost-Benefit  Analyses.  The  budget  for  imple-
menting  the chosen  techniques  for lake protec-
tion/management,  including water quality monitoring
for 1 year after implementation, and project manage-
ment and administration  is $114,500.
  The major water quality benefits expected from the
proposed protection/restoration plan  are:  (1)  fish
habitat will be increased to the full lake volume and
winter fishkills will be avoided; (2) release of nutrients
from the lake bottom sediments will be decreased by
90 percent;  (3)  total phosphorus and total  nitrogen
loading will be reduced by 64 and 70 percent, respec-
tively; (4) sediment loading to the lake will be reduced
by 50 percent; (5) transparency will be increased to at
least 1.2 m during summer;  (6) lake water temperature
will not exceed 21°C  in summer, allowing the develop-
ment of a year-round trout  fishery;  (7) nuisance algal
blooms will be prevented and aquatic weed growth
limited; and (8) a suitable turf irrigation system will be
maintained without lowering lake levels (Singh et al.
1983).
  Recreational  benefits  were calculated  using the
unit  day value method  recommended  by the  U.S.
Water  Resources  Council  (18 CFR 713, Subpart K,
App. 3, 1982), which gives  guidelines for assigning
points  for general recreation in terms of a matrix of
five criteria and five judgment factors, and for conver-
ting  points to unit day recreational  values (UDV) in
dollars (Table 2).
  Benefits from the  Lake of the Woods management
plan will be derived from improved recreational experi-
ence and environmental factors,  increased annual
users, and addition of a new use (trout fishery). Upon
implementation of the plan, the UDV will  increase
from $2.12 to $2.54 for 194,000 park visitors and water-
front users, and from  $1.88 to  $2.08 for 35,000 golf
course  users per year. The annual increase in park
               Table 1.—Morphometric and hydrologic characteristics of three Illinois impoundments.

Year constructed
Surface area (ha)
Volume (m3)
Mean depth (m)
Maximum depth (m)
Shoreline length (km)
Average retention time (yrs)
Total original capacity loss (percent)
Annual capacity loss (percent)
Watershed area (ha)
Lake of
the Woods
1948
9.4
2.78 X 105
3.0
6.70
2.89
0.534
13.5
0.42
245
Johnson
Sauk Trail L
1956
23.2
5.8 x 105
2.50
7.01
2.41
1.96
13.3
0.58
355
Lake
Le-Aqua-Na
1956
16.0
6.1 x 105
3.54
7.62
2.25
0.186
15.8
0.61
950
                                                 407

-------
 LAKE AND RESERVOIR MANAGEMENT
 visitors  and waterfront users is estimated at 27,COO
 and golf course  users ($8 fee each) at 5,000. The
 number  of museum visitors (26,000) will be unaffected.
 Thus,  the increase in recreational benefit is: $(254-
 2.12) 194,000 + (2.08-1.88)35,000 + (27,000 X 2.54) +
 (5,000 x 8) = $197,060/year.
   The total discounted benefit (at 7 1/8 percent and 10
 years) is  $1,376,100  and the  expected 314  grant
 amount  (Phase II) is 50 percent of $114,500, or$57,2!>0.
 Thus, the benefit/cost ratio for a Clean Lakes Phase II
 grant would be:
    Total discounted benefit
1,376,100
   Expected 314 grant amount     57,250
                                          = 24.0
   Using conventional cost-benefit analysis, the ar\-
 nual increase in recreational benefit  ($197,060) is
 divided by the annual cost ($24,270) obtained by amor-
 tizing the budget cost of $114,500 at 12 percent and 10
 years and adding the future operating costs of $4,0' 0.
 The benefit/cost ratio using this method is 8.1.
   Benefits in the property value of some homes  near
 Lake of the  Woods Park and use of the well as  an
 emergency water source for 3,360 persons served ay
 the Sangamon Valley Public Water District are not n-
 cluded  in this analysis.
CASE STUDY 2: JOHNSON SAUK TRAIL
LAKE

Pertinent Data. This 23.2 ha State park lake is located
in Henry County, III., part of the Rock Island-Daven-
port-Moline Standard  Metropolitan Statistical  Areia
(Fig. 1). It was formed in 1956 by the impoundment of
King Creek. Morphometric and  hydrologic character-
istics of the lake are shown in Table 1  and Figure 3.
The lake and surrounding park are managed by the Il-
linois Department of Conservation for recreational
purposes such as fishing, boating, hiking, camping,
and picnicking, and draw more than 379,000 visitors
annually. The State also  owns and manages 94 per-
cent of the 355 ha watershed, which is generally in e<-
cellent condition. Only 6 percent  of the watershed
area is devoted to agriculture.
   Water  Quality  Problems. A  diagnostic/feasibility
 study conducted on  Johnson  Sauk Trail  Lake in
 1981-82 identified the following problems:
   1.  High nutrient levels. Mean lake total phosphorus,
 dissolved phosphorus, and inorganic nitrogen concen-
 trations were 0.153, 0.07, and 0.61 mg/l,  respectively.
 Long term average gross loadings of TP, DP, and IN
 were 185, 149, and 14,334 kg/yr;  internal  regeneration
 accounted for 60.2, 75.0, and 93.5 percent of these
 loadings, respectively. The lake has  exhibited very
 high biological productivity requiring algicide and her-
 bicide applications since 1957 to control  algae  and
 macrophytes.
   2. Dissolved oxygen depletion. During  peak stratifi-
 cation periods, the lake was anoxic at depths below
 2.4 m from the surface. About 38 percent of the lake
 volume was devoid of  oxygen at that time.
   3. Lake turbidity  and sedimentation.  The Secchi
 disk transparency of Johnson Sauk Trail  ranged from
 0.15 to 2.6 m and averaged 1.3 m in 1981. The volume
 of water storage has  decreased 13.3 percent in 26
 years, a rate of 0.51  percent per year (Fig. 3). Prior to
 predominant State ownership of the lake's watershed,
 uncontrolled sediment transport  from the agricultural
 land  within  the watershed resulted in  the  sedi-
 mentation of the upper end of the lake. The tributaries
 now do not  convey unusual amounts of  sediment,
 even during storm events.
  4. Algal blooms. Algal growths of bloom propor-
 tions were observed during the summer months, with
 blue-greens  the  dominant  species. Chlorophyll a
 levels ranged from 20 to 80 ^g/l and averaged 40 ^g/l in
 I V?O I .
  5. Excessive aquatic weeds. About  27 percent of
 the lake surface area is covered by a dense growth of
 aquatic macrophytes  (predominantly  coontail  and
 pondweed).
  Objectives of Lake Management. The primary objec-
 tive of the proposed management program is to  im-
 prove the lake water quality and  maximize its recrea-
tional use. The specific objectives are:
  1. Improve fish habitat in the  lake during summer
and winter months by eliminating anoxic conditions in
the lake.
  2.  Minimize internal regeneration of nutrients in the
lake.
                 Table 2.—Summary of recreation benefit assessment (unit day value methodology).*
   GUIDELINES FOR ASSIGNING POINTS FOR GENERAL RECREATION

   Criterion                                               Judgment Factors
   a. Recreational experience                             # Activities supported
                                                      (general vs high quality)
   b. Availability of other opportunities                     # of other opportunities
                                                      (w/in 30 min vs 2 hours)
   c. Carrying capacity                                  Facilities present
                                                      (minimum vs ultimate)
   d. Accessibility                                      Access to site
                                                      (limited vs good)
   e. Environmental                                     Aesthetic factors
                                                      (low vs outstanding)
      TOTAL
 2. MATRIX FOR CONVERSION OF POINTS TO DOLLAR VALUES

 Activity category                                       Point Values

                    0       10      20     30      40      50     60

 General recreation   1.07    1.25     1.44     1.68     1.93    2.30     2.48
 General fishing
 and hunting         1.57    1.74     1.90     2.07     2.28    2.51     2.73

 •18 CFR 713, Subpart K, App 3,  1982
                                                                Points
                                         70

                                         2.67

                                         2.94
                           80

                           2.85


                           3.08
 90

3.04


3.17
100

3.22


3.20
                                                408

-------
                                                           CASE STUDIES OF WATER QUALITY IMPROVEMENTS
  3.  Improve the aesthetic quality of the lake waters
and enhance recreational opportunities in the lake.
  4.  Control  algal blooms and dense  macrophyte
growths in the lake that occur during the prime recrea-
tion period.
  5.  Enhance bank and open water fishing.
  Lake Management Plan. Based on technical, envi-
ronmental, and economic considerations, the follow-
ing in-lake management techniques were chosen for
Johnson Sauk Trail Lake.
  1.  Aeration/destratification  to  improve dissolved
oxygen  levels, with associated benefits:  initial cost
$15,000;  operation,  maintenance,  and  repair
$1,500/year.
  2.  Lake shoreline stabilization  (91  m  length)  to
reduce nutrient and sediment  input from shoreline
erosion: initial cost $10,000.
  3.  Harvesting of  aquatic  macrophytes twice  per
year to  reduce nuisance growths that interfere with
recreational  activities: $5,000.
  4.  Periodic applications of chelated  copper sulfate
followed by  potassium permanganate for algal con-
trol: $1,600.
  Cost-Benefit  Analyses.  The  budget  for  imple-
menting the chosen lake management techniques at
Johnson Sauk Trail Lake, water quality monitoring for
1 year after project implementation, and  project
management and administration is $74,990.
  Aeration/destratification  will  increase  the  fish
habitat to full lake volume in summer months, prevent
fishkills in winter, and improve aesthetic conditions in
the lake. It will also reduce the internal nitrogen and
phosphorus loading by 90 percent. The harvesting and
removal of macrophytes from the lake will export 113
kg of phosphorus per year from the lake. Lake shore-
line stablization  will also reduce nutrient and sedi-
ment inputs (Kothandaraman et al. 1983a).
  Recreational  benefits  were  calculated using the
procedure recommended by the U.S. Water Resource
Council (18 CFR 713, Subpart K,  App. 3, 1982). Upon
implementation of the management plan, the UDV will
increase from $2.22 to $2.51 for  379,581 visitors  an-
nually and number of visitors is expected to increase
to 530,000 annually, Thus, the increase in recreational
benefit  is $(2.51-2.22)  379,581   +  2.51  (530,000-
379,581) = $487,630/year. The total discounted benefit
(at 7 1/8 percent and 10 years) is $3,405,000  and the
U.S. EPA 314 grant amount (Phase II)  is 50 percent of
$74,990 or $37,495. Thus, the benefit/cost ratio for a
Clean Lakes Phase II grant would be:
                                                        Total discounted benefit
                                3,405,000
                                                       Expected 314 grant amount     37,495
                                                                                             = 90.8
Figure 3.—Johnson  Sauk Trail  Lake facilities and
bathymetric maps.
                                                      Using conventional  cost-benefit analysis, the an-
                                                    nual  increase  in  recreational  benefit ($487,630)  is
                                                    divided by the annual cost ($21,370) obtained by amor-
                                                    tizing the budget cost of $74,990 at 12 percent and 10
                                                    years and adding the future annual operating cost of
                                                    $8,100. The benefit-cost ratio using this  method  is
                                                    22.8.
CASE STUDY 3: LAKE LE-AQUA-NA

Pertinent Data. This 16-ha State park lake was formed
in 1956 by damming Waddams Creek in Stephenson
County, III. (Fig.  1). It is situated near the Illinois,
Wisconsin, and Iowa borders, and is within 40 km of
the Rockford  Standard Metropolitan Statistical Area.
Morphological and hydrological characteristics of the
lake are shown in Table 1 and  Figure 4. The lake and
surrounding State park are managed  by the Illinois
Department of Conservation and draw over 300,000
visitors annually  for recreational purposes such  as
fishing, boating, camping, and picnicking. Thirty-one
percent of the lake's 950 ha watershed is State owned,
and the rest is in small  private holdings, with 67 per-
cent of the area cropland.
   Water  Quality  Problems. The  water quality prob-
lems  exhibited by  Lake Le-Aqua-Na are  similar to
those discussed  previously  for Johnson Sauk Trail
Lake, although to a greater degree. The lake has ex-
hibited very  high biological   productivity,  requiring
algicide and herbicide applications since 1956 to con-
trol algae and macrophytes.  Algal  blooms are  fre-
quently recorded during summer months, with blue-
greens the dominant species. Chlorophyll a levels in
1981 ranged from 2 to 93 ng/l,  with an average of 46
^g/l. Over one third of the  lake surface area was
covered  with  a  dense  growth  of  macrophytes
(predominantly coontail and elodea). Although total
nitrogen/total  phosphorus   ratios  indicate   that
phosphorus is the limiting nutrient, there is an  abun-
                                                 409

-------
  LAKE AND RESERVOIR MANAGEMENT
  dance of phosphorus in the lake system at all times
  (average  total and dissolved phosphorus values  of
  0.373 and 0.247 mg/l, respectively, in 1981).
    Oxygen depletion  in the  hypolimnion has  been
  noted since the first year following impoundment. At
  peak summer stratification,  51 percent of the  kike
  volume is anoxic, with depths below  1.83 m totcilly
  devoid of oxygen. Several winter fishkills have also oc-
  curred.
    Based  on a sediment survey of the lake (Fig. 4), Ihe
  original lake volume was reduced 15.8 percent over a
  period of 26 years, representing  an annual capacity
  loss of 0.61 percent. Although the main creek and tri-
  butaries do  not carry abnormally heavy suspended
  solids loads during  normal rainfall conditions, they
  carry enormous loads during storm events. Two storm
  events with rainfall exceeding 10 cm each in June and
  August 1981 delivered 73 percent of the total phos-
  phorus loading to the lake. Long-term average gross
  loadings of total phosphorus, dissolved phosphorus,
  and inorganic  nitrogen were 1,802, 950, and 28,271
  kg/yr.  Internal  regeneration only  accounted for 7.3,
  13.8 and 55.7 percent of these loadings, respectively;
  most of the remainder came from Waddams Creek.
    Lake Protection/Management Plan. The objectives
  of the protection/management plan for Lake Le-Aqua-
  Na are  similar to those  discussed  previously for
  Johnson  Sauk Trail  Lake (to minimize  the influx of
  sediment and nutrients  and to improve the lake water
Figure 4.—Lake Le-Aqua-Na facilities and bathymetric maps.
 quality  and recreational experience). In developing
 this plan, constraints imposed on the quality of Lake
 Le-Aqua-Na due to morphological, hydrological, and
 watershed factors were considered. These constraints
 are typical of Illinois impoundments. Lake Le-Aqua-Na
 has a watershed area to lake surface area ratio of 59:1
 and it drains fertile cropland. This makes sufficient
 control  of watershed  sources of nutrients and sedi-
 ment very difficult. Even with 100 percent implementa-
 tion of watershed resource management systems, the
 contribution of phosphorus to the lake from the  water-
 shed alone (without even considering internal  regen-
 eration) will result  in a mean winter phosphorus con-
 centration of 0.066 mg/l (Kothandaraman et al. 1983b).
 And although cropland erosion rates will  be reduced
 by 42 percent, significant amounts of sediment will
 still enter the lake during major storm events.  Since
 the retention time of Lake  Le-Aqua-Na  is relatively
 short (0.186 year), frequent  inflows of nutrient and
 sediment laden water will result in nutrient loadings in
 excess  of eutrophic rates  and contribute to  water
 quality problems that must  be addressed by in-lake
 techniques.
   Based on these  considerations and a technical,
 environmental, and economic evaluation of the  feasi-
 ble  alternatives,  the  following  lake  and  protec-
 tion/management plan was recommended for Lake Le-
 Aqua-Na.
   1.  Adoption of conservation tillage practices with
 at least  40 percent residue left on the cropland  so as
 to  reduce soil  erosion  by 42  percent (that  would
 reduce  average  cropland  soil  loss  from  2.1
 tons/ha/year to 1.2 tons/ha/year): no cost to project.
   2.  Resource management  systems for lands  expe-
 riencing high soil loss (such as terracing and  strip-
 cropping practices on critical agricultural lands and
 erosion control measures along roadways): no cost to
 project.
   3.  Fencing along the main stem of Waddams Creek
 to exclude livestock and  implementation  of stream-
 bank stabilization measures to reduce soil erosion
 and  nutrient input from  this  source:  initial  cost
 $22,860.
   4.  Lake shoreline stabilization  (76 m  length) to
 reduce nutrient  and sediment input  from shoreline
 erosion:  initial cost $10,000.
   5. Aeration/destratification  to improve dissolved
 oxygen levels, with  associated  benefits:  initial cost
 $15,000,  yearly $1,500.
   6. Harvesting  of aquatic  macrophytes twice per
 year to reduce nuisance growths which interfere with
 recreational activities: $5,000.
   7. Periodic applications of chelated copper sulfate
 followed by potassium  permanganate for alqal con-
 trol: $1,280.
   Implementation  of watershed management  prac-
 tices (items 1-3) requires cooperation and  dedication
 of the private landowners in the watershed. As a result
 of this study, the Stephenson County Soil  and Water
 Conservation District received $38,000 in special Agri-
 cultural Conservation Program (ACP) funds in 1983 for
 implementation of the conservation tillage  plan in the
 watershed.  The  cost-share  funding  to  farmers is
 variable depending  upon the  amount of residue left.
 The Soil and Water Conservation District has received
 an excellent response from landowners in drawing up
 5-year contracts for conservation tillage (item 1) and in
 implementing resource  management  systems on
 critical areas (item 2).
  Cost-Benefit Analyses. Assuming that the cost of
watershed  measures (such as conservation tillage,
terracing, and stripcropping) will be borne by the  land-
                                                 410

-------
                                                               CASE STUDIES OF WATER QUALITY IMPROVEMENTS
owners and  other  agencies,  the budget  for imple-
menting the remaining measures (3-7), water quality
monitoring for 1 year implementation, and  project
management and administration is $112,320.
  The  implementation  of  watershed  management
practices (items 1-3) will reduce phosphorus inputs by
more than 40 percent,  but the  in-lake  restoration
measures (4 through 7) are necessary to improve water
quality and  lake usability. Aeration/destratification
will  reduce the  internal  phosphorus  and nitrogen
loading by 90 percent, thus  increasing fish habitat to
the full lake  volume  and  preventing  winter fishkills.
The harvesting and removal of macrophytes will result
in phosphorus export of 91  kg (Kothandaraman et al.
1983b).
   Most of the recreational benefits of the Lake Le-
Aqua-Na  restoration  project will  accrue  from better
recreation experiences, better environmental  factors,
increases  in the extent of use, and increases in the
number of annual visitors. After restoration, the unit
day recreational  value   using  the  unit  day value
methodology (18 CFR 713 Subpart K, App. 3,1982) will
increase from $2.22 to $2.51  for the present  301,587
visitors. The number of visitors is expected to increase
to 530,000 when the plan is implemented. Thus the an-
nual increase in recreational benefit equals: $(2.51-
2.22)301,587 + 2.51 (530,000-301,587) or $660,777. The
total discounted benefit (at 71/8 percent and 10 years)
is $4,614,000 and  the expected  314  grant  amount
(Phase II) would be 50 percent of $112,320 or $56,160.
Thus, the benefits cost ratio of a Phase II Clean Lakes
grant would be:
       Total discounted benefit   4,614,000

        Expected grant amount     56,160
= 82.2
   Using conventional cost-benefit analysis,  the  an-
 nual  increase in recreational  benefit ($660,777) is
 divided by the annual cost ($27,656) obtained by amor-
 tizing the budget cost of $112,320 at 12 percent and 10
 years and adding the future operating cost of $7,780.
 The benefit-cost  ratio using this method is 23.9.
 SUMMARY

 Studies  were conducted on  three  Illinois  impound-
 ments under the  Sec.  314 Clean Lakes Program to
 determine  practical   and  economically  viable
 measures  to reduce water quality related use impair-
 ments and preserve and enhance lake uses. Each im-
 poundment  represented a  different physiographic
 region of the State. These impoundments are eutro-
 phic with accompanying problems of hypolimnetic ox-
 ygen depletion, algal   blooms, dense  macrophyte
 growths, inorganic turbidity, and siltation.
   An optimal mix of remedial  measures and water-
 shed management techniques  was derived for each
 lake to  preserve  and maximize its  use at  minimal
 costs. Techniques such as  aeration/destratification,
 groundwater/surface  water  blending  (dilution),  lake
 shoreline  stabilization,  weed  harvesting, and  algal
           control were not found to be palliative measures but
           essential management tools to preserve the lakes and
           allow their full development as recreational resources.
             A recreational benefit assessment (of the expected
           increase in visitors and improvement in the quality of
           recreation) was performed for each  project using the
           unit day value methodology recommended by the U.S.
           Water Resources Council  (18 CFR 713, Subpart  K,
           App. 3. 1982). Benefit/cost ratios were determined in
           two ways: (1)  by the ratio of the  total discounted
           benefits to the requested Sec. 314 grant amount, and
           (2) by  the  annual recreational benefit divided by the
           sum of the amortized capital costs and annual opera-
           tion,  maintenance,  and  repair  costs.   Calculated
           benefit/cost  ratios for Phase II Clean Lakes grants
           were 24.0 for Lake of the  Woods,  90.8 for Johnson
           Sauk Trail Lake, and  82.2 for Lake Le-Aqua-Na. The
           recreational benefit assessment procedure appears to
           provide an excellent tool for evaluating lake manage-
           ment  strategies.  It demonstrates that management
           techniques to enhance lake use can  be very beneficial
           and reasonable in cost. Water quality values can  be
           enhanced without extensive  structural or watershed
           work.
             The technical information developed in these inves-
           tigations may  be applied to  numerous other recrea-
           tional   and  public  water   supply  impoundments
           throughout the Nation.
REFERENCES

Boland, D.H.P., et al. 1979. Trophic classification of selected
  Illinois water bodies: Lake classification through amalgama-
  tion of LANDSAT multispectral scanner and contact-sensed
  data. EPA-600/3-79-123. Environ. Monitor Sys. Lab. U.S. En-
  viron. Prot. Agency, Las Vegas, Nev.

Code of Federal Regulations. 1982. Title 18 (Conservation of
  Power and Water Resources), Subpart K (NED Benefit Evalua-
  tion Procedures: Recreation), Appendix 3 (Unit Day Value
  Method) revised April 1.

Kothandaraman, V., et al. 1983a. Clean Lakes Program phase 1
  diagnostic/feasibility study of  Johnson  Sauk Trail  Lake,
  Henry County, III. Prep, by coop, agree, between U.S. Environ.
  Prot. Agency, III. Environ. Prot. Agency, III. Dep. Conserv., and
  III. State Water Surv.
	1983b Clean Lakes Program phase 1 diagnostic/
  feasibility study of Lake Le-Aqua-Na, Stephenson County, III.
  Prepared by coop, agree, between U.S. Environ. Prot. Agency,
  III. Environ.  Prot. Agency, III. Dep.  Conserv., and III.  State
  Water Surv.

Sefton, D.F. 1978. Assessment and classification of Illinois
  lakes. Div. Water Po/lut. Control, III. Environ. Prot. Agency,
  Springfield.

Sefton, D.F., M.H. Kelly, and M. Meyer. 1980. Limnology of 63
  Illinois Lakes, 1979. Div. Water Pollut. Control, III. Environ.
  Prot. Agency, Springfield.

Singh,  K.P., et al. 1983.  Clean  Lakes Program phase  I
  diagnostic/feasibility study of Lake of the Woods, Cham-
  paign County, III. Prep, by coop, agree, between U.S. En-
  viron. Prot. Agency, III. Environ. Prot. Agency,  III. Nat. Hist.
  Surv., III. State Water Surv., and Champaign County Forest
  Pres. Dist.
                                                  411

-------
  AN HISTORICAL OVERVIEW OF A SUCCESSFUL LAKES
  RESTORATION PROJECT IN BATON ROUGE, LOUISIANA
 RONALD M. KNAUS
 Nuclear Science Center

 RONALD F. M ALONE
 Civil Engineering  Department
 Louisiana State University
             ABSTRACT
             The City Park Lake and the University Lakes complex have long been a source of pride to the citizens
             of Baton Rouge This five-lake system comDnses 300 acres. Over the past 50 years all aspects of
             the lakes have deteriorated with massive fishkills occurring on a regular basis during the warm sum-
             mer months. In 1977 the City-Parish Government of East Baton Rouge asked the Institute for En-
             vironmental Studies at Louisiana State University to draft an Environmental Impact Statement and
             a Section 314 grant application to the Environmental Protection Agency lo correct the hypereutrophic
             conditions in the five lakes  Then followed 2 years of negotiations gaining public support for the dredging
             project and obtaining areas for dredge spoil deposition. In September 1980 the project went out for
             bidding to the contractors, but because of the distance the spoil had to be pumped, bids far exceeded
             the $3 million available for the project. After several new dredging plans were presented, a large ma-
             jority of people affected by the dredging gave approval to a plan using 30 percent in-lake disposal,
             and 70 percent off-site disposal at a very close location. In June 1981, the low bid for $2 09 million
             was  accepted. Dredging commenced in November  1981. After the removal  of 490,000 m3
             of lake bottom material, dredging ended in May 1983. In the post-dredging monitoring program, a
             group of 15 interested citizens was  designated by the mayor to serve on the University Lakes Com-
             mission to oversee the monitoring of the lake waters, stabilization of shorelines and spoil banks, and
             to recommend funding for recreational use of the lakes and the immediately adjacent land areas.
 The restored lakes discussed in this paper are located
 within the city limits of Baton Rouge, La., about 3 km
 south of the downtown area. Collectively known as the
 University Lakes System, these five lakes are adjacent
 to the Louisiana State University (LSU) campus, with
 the Mississippi River less than 2 km to the west. U.S.
 Interstate 10 crosses the northernmost  lake of the
 system.
  The University Lakes System occupies what was
 once a Bald Cypress swampland. In 1925 the cypress
 were logged and the swamps and low relief bayous
 dammed by land developers to create waterfront lots.
 The surface area and shorelines of the various lakes
 formed  over the  stumps  and rubble of the logging
 operations  were donated  to the city of Baton Rouge
 "for the enjoyment of the  public into the future." Tie
 lakes  have  become  a source of beauty and pride for
 the Baton Rouge community.
  City Park Lake, deeded to the city of Baton Rouge in
 1925, is the northernmost  of the five-lake system and
 has a surface area of 24 ha. University Lake, deeded to
 LSU in 1933, is the largest  of the system and occupies
 a total of 84 ha. Crest Lake, 2 ha in size, was a branch
 of University Lake until it was truncated by a levee and
 a road.  It drains into University Lake and has been
 owned by LSU since 1933  (Knaus et al. 1977).
  The two remaining lakes of the system are close to
the other three, but in a different drainage basin. Cam-
 pus Lake is 3.7 ha, owned by LSU, and is located on
the LSU campus. College Lake is 1.4 ha and is owned
 by LSU, with one half of its shoreline in private owner-
ship.
 THE RESTORATION

 Over the last two decades, the lakes' continuing poor
 water quality culminated in the summer of 1976 with
 unprecedented massive fishkills. That same year the
 bottom of Campus Lake became anaerobic and sev-
 eral domestic ducks died of  Type C botulism. With
 public awareness heightened, faculty and  students
 petitioned (1,183 signatures) to rectify the poor condi-
 tion of the lakes. Baton Rouge city officials learned of
 the provision in the Federal Water Pollution Control
 Act Amendments of  1972 (P.L. 92-500) that  provided
 matching funds under Section 314, the "Clean Lakes"
 clause. The  officials  contacted the LSU Institute  tor
 Environmental Studies to provide the necessary docu-
 ments to the U.S. Environmental Protection Agency to
 apply for a matching grant under the Clean Lakes pro-
 vision.  In April 1977 a $100,000 initiation grant from Ci-
 ty revenue sharing funds was awarded to LSU to draw
 up a combined environmental impact statement and
 grant application (Knaus et al. 1977). In this proposal it
 was determined that  the best means  to improve the
 hypereutrophic  conditions exhibited  in four of  the
 lakes would be to  increase  the water volume  by
 hydraulic dredging. The grant application asked for $3
 million,  half  from the U.S. EPA, the other half from
 City-Parish sources.
  In October 1978 a letter of credit in the amount of
$1.5  million  from the EPA  was deposited with the
 Federal Reserve Bank  of New Orleans.  In  January
 1979 the City-Parish  of Baton Rouge officially ear-
marked  $1.5 million from revenue sharing funds  to
                                                412

-------
                                                              CASE STUDIES OF WATER QUALITY IMPROVEMENTS
 match EPA's grant (Fig. 1). LSU and City-Parish per-
 sonnel  formed the  Lakes  Restoration  Committee
 which  immediately  began to  assess  alternative
 methods that had been outlined in the Environmental
 Impact  Statement-Grant Application document  for
 the best way to improve water quality in the five-lake
 system.
   The  first  alternative  seriously considered was
 pumping hydraulic dredge spoils directly to the Mis-
 sissippi  River  batture, less than 2  km  away. This
 economically feasible alternative was rejected by EPA
 and Army Corps of Engineers for its  potential to
 pollute  the river.  Another alternative,  placing the
 dredge  spoils  on LSU agricultural lands,  was also
 disapproved because the spoils might make the land
 unsuitable for cattle grazing for several years. This left
 the Lake Restoration Committee with the alternative
 of pumping the dredge spoils to a site 11 km from the
 project area (Knaus et al. 1979).
   In September 1980 the project went out for bidding
 to several  contractors. Because of the distance the
 spoil  had to be  pumped,  bids far exceeded the  $3
 million available for the project. During the  next year
 the Lakes Restoration Committee wrestled with plans
 for in-lake dredge disposal methods: for instance, en-
 larging  existing  peninsulas and creating  islands.
 These plans met  with great resistance from adjacent
                                           land owners.  Using land owner resistance to in-lake
                                           disposal as leverage, a compromise of 30 percent in-
                                           lake dredge spoil disposal and 70 percent close off-
                                           site disposal was hammered out between land owners
                                           adjacent to the lakes and land owners adjacent to the
                                           close-in 10 ha disposal site.
                                             The revised dredge plan went  out to bid in  June
                                           1981. A low bid of $2.09 million was accepted.  After
                                           more  public  hearings and  clarifications,  hydraulic
                                           dredging commenced  in University Lake in November
                                           1981.
                                             After the removal of 460,000 m3  of lake bottom
                                           material from City Park and University  Lakes the
                                           dredging ended in May 1983.  The average depth of
                                           University  Lake  before dredging  was 0.6 m;  after
                                           dredging it was 1.3 m. Maximum depths in University
                                           Lake increased  from less than 1  ha at 1.6 m before
                                           dredging to 40 ha at 1.6 m or over after dredging (Table
                                           1). City Park Lake had an average depth of 0.9 m; after
                                           dredging this increased to 1.2 m (Table 1). Crest Lake
                                           with an average depth of 1.6 m was deemed to be deep
                                           enough, so no dredging took place.
                                             The numerous stumps and logging debris left in the
                                           lake bottoms since the lakes' formation, although well
                                           known by the bidding  contractors, did  in the end
                                           cause extensive delays in the dredging operation. Dur-
                                           ing  dredging the contractor  drained  the two small

ing
i
50-

Boton
Rouge
City -
Parish
I00,00(
Rev
Sharing


5


Baton Rouge
City - Parish
$1,500,000
(Less State Funds
$ 600,000)
State of Louisiana
$ 600,000
-^-
Revenue Sharing Funds

U S. Environmental Protection Agency
| 1,500,000
-<•
BR, C-P
$ 253,000
*-

USEPA
$ 253,000
Supplement
*.
                1977 I  1978
                      I   1979    I   1980    I
                                                          1981
                                                                      1982
T  1983
1984   '
     Funds
   Remaining
100

 50

  0
Figure 1.—Funding sources that initiated and supported the University Lakes Restoration Project in Baton Rouqe La  March
1977 through December 1984.                                                                    '   ''


 Table 1.—A summary of lake parameters, Lakes Restoration Project, Baton Rouge. Crest Lake, not shown, was not dredged.
 Lakes
                             City Park
                                                           University
  Campus
 'The surface areas of City Park and University Lakes were reduced by in-lake fill by 1 ha and 2 ha, respectively
 'Indicates only dredging effects on detention time Impact of the interceptor drains on these lakes could effectively double the detention times
    College
Area (ha)
Average depth pre-dredge (m)
Average depth post-dredge (m)
Cubic meters removed (in 1000's)
Detention time, pre-dredge (days)
Detention time, post-dredge (days)
24'
0.9
1.2
100
47
56
841
0.6
1.3
360
49
101
3.7
0.5
1.1
24
19
402
1.4
1.0
1.3
4
23
292
                                                  413

-------
 LAKE AND RESERVOIR MANAGEMENT
  lakes, Campus and College  Lakes,  and  removed
  28,500 m3 of silt by means of dragline dredge accom-
  pained by a fleet of trucks (Table 1). The data in Taale
  1 are newly computer-generated figures, not simale
  conversions of data found in Knaus et al, 1979.
  PROJECT FUNDING AND ANCILLARY
  EXPENDITURES

  The failure of the first bidding round in  1980 caused
  unforeseen  delays  in the project, already extenced
  beyond normal  budget  intervals. The  City-Parish,
  therefore, had to reassign monies held aside for Ihe
  lakes restoration project  to other much  needed civic
  obligations. At this juncture the State of Louisiana ac-
cepted  a request to restore the reapportioned City-
Parish funds with State monies, thus saving the pro-
ject (Fig. 1).
   In May 1983 the City-Parish was awarded a supple-
mental grant from EPA of $253,000 for erosion control
and shore stabilization. The local government is pre-
pared to match this money with funds left over from
1982 Federal revenue  sharing,  or from money gen-
erated by a Federal jobs bill (Fig. 1).
   During the time of the funded activities shown in
Figure 2, there was considerable expenditure of time
and money by local government, LSU personnel, and
the EPA administration which received absolutely no
compensation from grant  monies. These nonfunded
activities are depicted as bars in  Figure 2 along a time-
line from 1977 to 1985.





Fu nded

Activity
i

i
t
N o n- c
Funded
Acti v i t y











EIS

&

Grant

Appli'c.

I977
/
/
/
/
/#'
/
/Bid
/
A
/
/
/
M
/ i
/Bid I

D
R
E
D
G
I






N

o








Monitoring Activities

I978 I979 I980

I98I
I983

I983 I984 I985
City - Parish of E. Baton Rouge Personnel Input — »•
C-P, EBR Capital, Maintenance Support — *•
LSU Personnel Input — *-

•< 	 EPA Administrati Dn — >-

Univ Lakes Commission — »-

Figure 2.-University Lakes Restoration Project activities, March 1977 through December 1984. Activities funded by Federal
btate, and local governments are shown above the dateline. Activities funded entirely outside the project funds are shown
below the dateline.

u
C-P
BR EC
State Fish-Wildlf ?
C-P Erosion M cc/~inn/-\
SEPA Control » 550000
Road Improv. I 00000
Maintenance 40000/yr
C'P Jog6 Poths 1 50000
LSU Intercepto
c p Engr.
0 p Study
r 92000
43000
l/SEP^'8 Sewerage Rehabilitation 1 50000
City-Parish Sewerage Corrections
50000
City -Parish Staff Com pensotf on (Conservative est. ) 200000
EIS
Grant
I977 I978 I979 I960 I98I I982 1 983 I984
I 00 000
ft 1,475,000
Figure 3.—Local funding stimulated by the University Lakes Restoration Project are shown along the dateline  In case of
large ancillary projects, the portion of funds devoted to the University Lakes Project area have been prorated.
                                                 414

-------
                                                             CASE STUDIES OF WATER QUALITY IMPROVEMENTS
  Figure 3 approximates  in dollar  amounts expen-
ditures in time, money, and capital for projects an-
cillary to, but directly stimulated  by the main  lakes
restoration project.  As was stated earlier, the  City-
Parish awarded a $100,000 initiation grant for the En-
vironmental  Impact  Statement-Grant  Application
document. Additionally, from 1977  through the pre-
sent, personnel of the City-Parish government have
spent untold hours inside and outside of their regular
working  hours  meeting with  the  Lakes Restoration
Committee, at evening meetings with land owners and
civic associations, formal and informal public hear-
ings, Federal and State governmental agencies, and
LSD administrators. This project occupied hundreds
of hours of City-Parish staff time  in letter writing,
document and map preparation, and keeping abreast
of Federal regulations. None of these services was
charged  against the project.
   Many other benefits in the vicinity of the lakes pro-
ject were wholly or partly accomplished because of
the existence of the lakes restoration project. These
include  sewerage  correction   and  rehabitation,
engineering studies, road improvement, and recrea-
tional uses of the lakes area, such as bike and jogging
paths, park areas, and fish stocking (Fig. 3). Out-
standing among these ancillary projects is the LSD in-
terceptor drain project for $92,000. This project involv-
ed diverting silt- and  nutrient-ladened waters around
Campus and College Lakes, thus preventing deteriora-
tion after dredging. Had LSD not cooperated in the in-
terceptor program, original grant conditions would not
have been met and those two lakes would have been
struck from the project.
Several times the project could have been aborted as
past  lake  improvement  proposals  had  been  in
previous  years.  But fortuitous  circumstances and
dedicated  individuals  always  seemed  to appear
together  and carry the project forward. The expertise
and  support from Federal, State and local govern-
ments, together with LSD, have given Baton Rouge
revitalized lakes.
  Although the Clean Lakes provision of P.L. 92-500
operates on a 50-50 matching funds basis, it was
found  that  in this  project  ancillary expenditures
amounted to an additional $1.5 million. Therefore, the
local match was  actually $3 million compared  to the
EPA grant of $1.5 million. This fact should encourage
government officials to  back Federal  involvement in
helping local projects like Clean Lakes because such
projects  stimulate  far  more expenditures  and  ac-
tivities than their original mandate.
  Although the actual dredging process of the lakes
has  been completed, interest in the lakes project is
continuing  and plans are  being  made for the  lakes'
future. Interested townspeople,  with  city personnel
and  LSU consultants,  have formed  the  University
Lakes Commission to make recommendations on the
future of the revitalized lakes, including landscaping,
public access, and recreational  potentials. Ongoing
monitoring of the lakes' waters is being conducted by
LSU researchers to study changing conditions  of the
lakes and water quality. City officials and LSU person-
nel will continue to learn about the dynamics of these
lakes and publish their findings so lake and reservoir
managers in other communities will benefit from this
project.
 CONCLUSIONS

 A great deal  of  effort on the part of many people
 brought the dredging phase of the project to a suc-
 cessful conclusion. Before the restoration, the lakes
 were an unmanageable blight. The Baton Rouge com-
 munity now has  a lakes system that  can be molded
 and  integrated  into  an  important  functional role.
REFERENCES

Knaus, R.M., et al. 1977. Lakes Restoration Project. Dep. Pub.
  Works, City-Parish of Baton Rouge, La.
	. 1979. Lakes Restoration Project. 2nd ed.-Dep. Pub.
  Works, City-Parish of Baton Rouge, La.
                                                  415

-------
  DREDGING OF  CREVE COEUR LAKE, MISSOURI
  GREG KNAUER
  Booker Associates
  St. Louis, Missouri


               ABSTRACT
               Creve Coeur Lake, an oxbow lake located m the Missouri Bottoms area of St. Louis County Mo
               is decreasing in surface area by deltaic deposition. In the early 1900's, the main lake had a surface
               area of approximately 81 ha (400 acres) and a small upstream lake. The main lake was reduced to
               an area of 32.4 ha (180 acres) by 1974 and the small lake completely filled. The average water depth
               decreased from 3.85 m (10 feet) to less than .77 m (2 feet). A 1971 engineering report included soils
               information, hydrologic and hydraulic investigations, field surveys, laboratory analyses, alternate designs
               with estimated  costs, and economic analyses of benefits. Water quality problems noted in the 1971
               water quality sampling program indicated high pH (>9.0), excessive nutrients, influx of fecal coliforms
               and excessive algal growth (>lO,000/ml). The result of the analyses indicated an optimum lake with
               an area of 133.4 ha (330 acres) dredged to a depth  of 3.85 m (10 feet). St. Louis County initiated
               dredging in June 1974, completing it in December 1981.  Baseline data collected in  1978 indicated
               high concentration of solids, excessive nutrients, and high levels of some heavy metals, including
               mercury. Sediment samples contained excessive nutrients with total phosphorus ranging from 1 200
               to 2,200 ng/g and exceeded Missouri State Water Quality Standards for cyanide, mercury, phenols,
               iron, lead, nickel, and zinc. Routine water quality monitoring was performed monthly at the inflow!
               outflow, and five sampling stations in the lake. Sampling occurred during March-December 1981'
               while dredging was in operation, and during March-December 1982, after dredging was completed!
               Parameters routinely monitored included temperature, dissolved oxygen, Secchi transparency, pH,
               total phosphorus, ortho phosphorus, total nitrogen, fecal coliform, mercury, phenol, cyanide, and zinc!
               Fish flesh analyses were conducted on three species each quarter to monitor a bioaccuniulation of
               mercury in the food chain. Analysis of the parameters during the dredging operations and the year
               following completion of dredging did not indicate any significant differences attributable to the dredg-
               ing operations.
 INTRODUCTION

 Project Background

 Creve Coeur Lake, located in  St. Louis  County,  is
 situated in the floodplain of the south bank of the
 Missouri River approximately 56 kilometers (35 mileii)
 above its confluence with the  Mississippi River. The
 largest natural lake in St. Louis County, it was formed
 approximately 10,000  years ago as  a result of  a
 change in course of the Missouri River. The lake has a
 watershed area of approximately 71.6 square km (27.5
 square miles).
   The recorded history of the lake dates back to the
 late 18th  century. During  that time,  there was  ,a
 smaller, upper lake approximately 1  mile west of the
 present lake. Maps prepared as late as 1920 indicated
 the presence of both lakes; however, the upper lake has
 since been filled. With  the  arrival of  streetcars from
 the St. Louis area to the lake, it soon became a major
 recreational  area.  Resort cabins were constructed
 along  the  lake and an amusement  park  known  as
 "Electric Park" was built on a  bluff overlooking the
 lake. But, the area fell into disrepair during the prohibi-
 tion era and was not actively  used  again until the
 1940's.
   Since the early  1900's,  the  lake's  surface area
 dramatically declined as a  result of  excess erosion
 and sedimentation. Sediment deposition reduced the
 lake surface area from 162 ha (400 acres) to 99 ha (24(>
 acres)  by 1970, and to less than 73 ha (180 acres) by
 1971. The average water depth also decreased from 3
 meters (10  feet) to less than 0.6 meters (2 feet) during
this time (Fig. 1).
   In 1969,  a general obligation bond  was passed by
St. Louis County voters with a portion of the funds set
 aside for rejuvenation of Creve Coeur Lake. Additional
 matching  funds were provided  initially through the
 Heritage Conservation and Recreation Service and
 subsequently by  the U.S.  Environmental  Protection
 Agency Clean Lakes Program.
   A 1971 engineering study determined that if correc-
 tive measures were  not initiated, the lake would be
 completely filled within 10 to 15 years. To prevent this,
 a dredging program was  recommended  to restore
 Creve Coeur Lake to a 130 ha (320 acre) lake including
 a 4 ha  (10 acre) island. The desired depth  was iden-
 tified as 3  meters (10 feet).

 Dredging Operations

 Between June 1974 and December 1981 approximate-
 ly 3,700,000 cubic meters (4,800,000 cubic yards) of
 dredged material  were  removed from Creve Coeur
 Lake.  The dredging operation  continued  unless
 weather conditions or mechanical failures prevented
 successful  operation of  the dredge.  The  dredge
 operated at a rate of about 18,925-22,710 I per minute
 (5,000-6,000 gallons  per minute) and the  dredged
 material was  transported to several settling ponds
 near the lake  by means of a pipeline. The dredging
 material consisted of clay and clayballs which were
 carried  in slurry form to the settling basins.
  The settling ponds were located immediately north
 and west of the lake.  Fill areas were prepared prior to
the placement of dredging material by using surface
soils  in the construction  of containment dikes. The
surface soils  consisted of  organic  clays,  medium
                                                  416

-------
                                                            CASE STUDIES OF WATER QUALITY IMPROVEMENTS
plastic clays, and clayey and silty sands. The dikes ex-
tended 0.5 meters (1.5 feet)  above the expected fill
elevation to allow sufficient freebroad to contain the
dredged material.
  The soils used to construct the earthen dikes were
compacted to approximately 90 percent of the Stan-
dard Procter Density. Slopes on the outside perimeter
dikes were approximately 1-11/2 horizontal to 1 ver-
tical. Some embankments were constructed for the
dual purpose of serving as dikes and  roads.
  The sediment was  allowed to settle and the super-
natant liquid was drained back into the lake by adjust-
ment of a weir overflow structure. There was  a 3 to 6
month time lag  between the suspension  of dis-
charging dredged  material to a settling basin  and the
establishment of vegetation on the site. The disposal
areas are presently being graded and seeded to pro-
vide additional productive park areas which will be us-
ed for recreational activities associated with the park.

Water Quality Monitoring Program

Water quality sampling was conducted to provide re-
cent baseline water quality information, monitoring of
water quality during  dredging, and the monitoring of
water quality following the cessation  of dredging ac-
Figure 1.—Delta growth.
tivities. The baseline water quality data was collected
during the  initial study in 1971  and again in  1978.
Routine water quality sampling occurred at eight loca-
tions  including upstream, five in-lake stations, down-
stream, and at the effluent from the spoils area from
March 1981 through December 1982.
  Water quality samples were collected monthly and
analyzed  for  total  and  ortho-phosphorus,  total
nitrogen, fecal coliform, mercury, phenols,  cyanide,
zinc, and pH. Field measurements  were collected at
the time  of sampling for temperature, dissolved ox-
ygen,  and transparency.
  Flow monitoring was conducted  at the  upstream
and downstream stations. During certain storm events
the stream  monitoring also included testing for total
and ortho-phosphorous; nitrite, nitrate, and Kjeldahl
nitrogen; pH; suspended solids; and metals listed in
the Missouri Water Quality Standards for the protec-
tion  of aquatic  life, including  antimony,  arsenic,
beryllium,  boron,  cadmium,  chromium,  copper,
cyanide,  iron, lead, mercury, nickel, selenium, silver,
thallium,  and zinc.


Sediment Cores/Elutriate Tests

During the initial  project feasibility study completed in
1971,  sediment  characteristics  of the lake  were
evaluated by the collection and analysis of 25 test bor-
ings and  18 hand samples. Additional sediment cores
were collected during April 1979  to determine which
chemical constituents were present and which might
be released into the water column during the dredging
operation.
  Sediment samples were tested for the heavy metals
previously listed, Kjeldahl nitrogen, total phosphorus,
pH, phenols, and pesticides BHC,  DDT, and Endrin.
Elutriate  samples were analyzed for iron, lead, mer-
cury, nickel, zinc, phenol, and cyanide because these
parameters had exceeded the State of Missouri Water
Quality Standards for the protection of aquatic life in
impoundments in an initial analysis of supernatant ef-
fluents from the dredged spoil sites at Creve Coeur
Lake.
                                                   Fish Flesh Monitoring

                                                   Because mercury concentrations in the 1978 baseline
                                                   water quality samples exceeded established Missouri
                                                   Department of Natural Resources standards, samples
                                                   of fish flesh were routinely analyzed for mercury dur-
                                                   ing the project. The fish flesh analyses were used to
                                                   ensure that, although the results of the water quality
                                                   sampling indicated levels of mercury exceeding State
                                                   standards,  the amount in the fish  flesh remained
                                                   within levels prescribed by the Food and Drug  Ad-
                                                   ministration (FDA) as fit for human consumption.
                                                      Fish  were  collected  during  March,  June,  and
                                                   September 1981; December 1982; and May 1983. The
                                                   fish species collected  included gizzard  shad,  large-
                                                   mouth bass, white crappie, channel catfish, and bull-
                                                   heads. Functionally, these fish represent three  dif-
                                                   ferent types of feeders. The gizzard shad are primarily
                                                   filter feeders and, therefore, would be a good indicator
                                                   of any mercury absorbed to suspended particulate
                                                   matter. The largemouth bass and white crappie  are
                                                   primarily predators and would provide an indication of
                                                   bioaccumulation  upward  through the  food  chain.
                                                   Channel catfish and bullheads are bottom feeders and
                                                   provide  information   on  uptake  from   bottom
                                                   sediments.
                                                417

-------
LAKE AND RESERVOIR MANAGEMENT
 RESULTS
 Water Quality

Parameters were routinely monitored during dredging
in 1981 and  in  1982, the year following dredging.
Sampling occurred monthly at the inflow station, the
five in-lake stations, and the outflow station. Tables 1
and 2 present the results for the 12 parameters during
1981 and 1982. Figure 2 illustrates the mean values for
selected parameters. While there are  some dif-
ferences between 1981 and  1982, it is not apparent
that  the cause of the differences can be attributed
solely to the dredging.
  Before discussing the specific parameters, two con-
ditions that have a direct impact on the in-lake water
quality need to be considered. First, the sewage line
and pumping station upstream of the  lake malfunc-
tioned at times,  contributing to excessive levels of
fecal conforms in the lake samples. A second  influen-
cing  factor  is the  meteorological conditions im-
mediately  prior  to  sampling.  Daily precipitation
records from the U.S. Weather  Bureau and rainfall
events were compared with sampling results.  When
precipitation events were greater than 2.5 centimeters
(1.00 inch) in measurable rainfall and occurred within
the week prior to sampling, in-lake concentrations for
total phosphorus, total nitrogen, and zinc were higher
than during dry weather or lesser rainfall events.
  Analysis of much of the water quality data has been
prepared in three formats for comparison of ranges
(Tables 1 and 2), annual arithmetic mean values  of the
five in-lake stations  per parameter for each year (Fig.
2), and arithmetic mean values for each station for
each year (Fig. 3,  4,  and 5).
  Temperature. In both 1981 and 1982, Creve Coeur
Lake warmed rapidly during April to 14-20°C (57-68°F)
and remained above 24°C (75°F) for most of the sum-
mer and early fall. During September, the temperature
decreased to 21 °C (70°F) and then to 10°C (50°F) or less
during November.
  Thermal stratification was never firmly established
during   the   spring-summer  warming  cycle.
Temperature differences between  the  upper layers
and bottom waters were generally not more than 3°C.
              Table 1.—Creve Coeur Lake range of wtiter quality testing results March-December 1981.
Parameter
Temperature (°C)
Dissolved oxygen
(mg/l)
Secchi (cm)
pH (units)
Total phosphorus
(mg/l)
Ortho phosphorus
(mg/l)
Total nitrogen
(mg/l)
Fecal coliform
(colonies/100ml)
Mercury (mg/l)
Phenol (mg/l)
Cyanide (mg/l)
Zinc (mg/l)
2
8.20-20.0

6.9-8.3
15-60
7.2-9.0

0.128-4.05

0.010-0.138

0.480-13.03

< 100-25,000
< 0.0001 -0.0004
0.004-0.02
0.004-0.05
0.04-0.16
Lake Sampling
3
20.0-27.0

6.0-8,0
15-60
7.3-8.6

0.112-4.100

0.010-0.174

1.100-14.19

50-23,000
0.0001-0.0008
0.0006-0.31
< 0.01-0.06
0.04-0.38
Stations
4
21.0-26.0

7.1-8.1
15-75
6.4-9.5

0.117-3.800

0.01-0.112

1.160-12.99

2-31,000
0.0001-0.0003
0.003-0.01
0.008-0.03
0.04-0.17
5
21.0-26.0

7.6-8.0
30-75
7.5-8.2

0.115-4.75

0.010-0.112

1.106-14.24

< 100- 11, 000
0.0001-0.004
<0.01
0 008-0.01
0.03-0.12
	 6 	
21.0-28.0

6.9-8.2
30-90
7.0-10.0

0.107-4.100

0.010-0.200

1.150-12.40

40-7,000
0.0001-0.0003
<0.01
0.006-0.01
0.04-0.12
              Table 2.—Creve Coeur Lake range of water quality testing results March-December 1982.
Parameter
Temperature ( C)
Dissolved oxygen
(mg/l)
Secchi (cm)
pH (units)
Total phosphorus
(mg/l)
Ortho phosphorus
(mg/l)
Total nitrogen
Fecal coliform
(colonies/1 00ml)
Mercury (mg/l)
Phenol (mg/l)
Cyanide (mg/l)
Zinc (mg/l)
2
9.0-26.0

1.2-7.2
15-45
7.2-8.5

0.112-0.698

0.021-0.433
1.212-5.991

200-> 200,000
< 0.0001-0.0004
< 0.01-0.01
<0.01
0.02-0.27
Lake Sampling
3
9.0-25.0

3.8-11.0
15-46
7.4-8.I5

0.101-0.1598

0.026-0.460
1.949-5.830

14->200000
0.0001-0.0002
< 0.01-0.06
< 0.01
0.01-0.12
Stations
4
9.0-26.0

1.0-9.2
15-45
7.0-8.0

0.110-0.788

0.026-0.535
2.038-5.710

6- > 200,000
0.0001-0.0002
<0.01
<. 0.01
0.01-0.11
5
9.0-26.0

2.4-10.4
15-45
6.7-8.2

0.114-0.700

0.032-0.450
2.240-6.333

< 2- > 200,000
0.0001
<0.01
< 0.01 -0.01
0.01-0.10
6
9.0-26.0

1.2-11.7
15-45


0.83-0.755

0.030-0.530
1.654-4.500

2-200,000
0.0001-0.0002
< 0.01 -0.01
<0.01
0.02-0.10
                                                418

-------
                                                            CASE STUDIES OF WATER QUALITY IMPROVEMENTS
  Without the  establishment  of  definitive thermal
stratification, the lake should continue to mix during
most of the year. The depth of the lake at 3 meters (10
feet) is the  approximate depth  at which thermal
stratification  can be expected to occur in lakes in the
temperate climatic zone. The uniform depth and lack
of wind protection  on three sides will probably main-
tain the mixing  action of the lake.
   Dissolved Oxygen. The dissolved oxygen (DO) con-
centrations  followed  similar patterns  in  1981  and
1982. DO concentrations in the upper waters generally
were in the 10 to 12 mg/l range during March and April
of each year and decreased to a low 6 mg/l during the
summers.  During fall the DO concentrations increas-
ed to the 8 to 10 mg/l range. In 1982, the DO concentra-
        4.000


        3.000


        2.000


     c 1.000
     ~»
     o>
     3 0.900


        0.800
     c 0.700
     .0
     '^
     CO
     ~ 0.600
     0>
     u

     o 0.500
        0.400


        0.300


        0.200


        0.100
                                   13.03    < 5.99
tions remained at 10 mg/l during December. In 1981,
the DO concentrations decreased  to 5 mg/l in late
November.
  The bottom waters of Creve Coeur Lake managed to
maintain oxygen  throughout  the  March  through
December period of both years.  The DO  concentra-
tions did decrease to less than 2 mg/l for short periods
during the year but ranged up to 7 mg/l during much of
the summer.
  pH. Values for pH generally  were slightly alkaline,
ranging between 7.5 and 8.2 for the majority of 1981
and 1982. The exceptions to this pattern occurred in
June 1981, when  pH values were in the 8.0-10.0 range,
and in July 1982, when a value  of 6.7 was  reported at
one station.
0.14


0.13


0.12


0.11


0.10


0.09


0.08
   c   0.07
   _o
   '•^
   CO
   •jjj   0.06
   0)
   o

   o   0.05
       0.04


       0.03


       0.02


       0.01
                            , .270
               Total       Ortho        Total
            Phosphorus  Phosphorus   Nitrogen
          Cyanide
                   Zinc
Phenol
Figure 2.—Annual mean values and ranges for various parameters, Creve Coeur Lake.
                                                 419

-------
 LAKE AND RESERVOIR MANAGEMENT
    Cyanide,  Mercury,  and  Phenols.  These  throe
 parameters  remained essentially unchanged during
 1981 and 1982 with a few exceptions. For most of the
 sampling period, the cyanide values were at 0.01 mg/l.
    During June, July, and August 1981, mercury con-
 centrations varied from 0.0002 mg/l to 0.0004 mg/l. Far
 the remainder of  the sampling program,  mercury
 values were at or less than 0.0001 mg/l. While these
 levels  exceed  the State of Missouri Standards  of
 0.00005 mg/l, mercury  levels in the fish flesh remained
 well below the upper limits prescribed by FDA for edi-
 ble fish.
    Phenols  remained  unchanged during  the  water
 quality monitoring program at levels of 0.01 mg/l.
   Zinc. Concentrations of zinc at each station were at
 or below State of Missouri Standards of 0.100 mg/l
 based  on  arithmetic mean values for nine sampler
 from five stations each year. Annual mean values at
 each station were higher in 1981  than in 1982 (Fig. 3).
 The values ranged from 0.07 to 0.10 mg/l for 1981. For
 1982, values ranged from 0.05 to 0.07 mg/l.
   Total Nitrogen. Concentrations of total nitrogen
 followed similar trends for 1981 and 1982, although
 the peaks for  1981  were  much  greater. The general
 trend for both years was for total nitrogen to increase
 in May, decrease through the summer, increase again
 in late September or October, and then decrease. The
 separation between high  and low values was  much
 greater for 1981 than for 1982 (Fig. 4).
   The mean value was approximately 4.0 mg/l for 1961
 and 3.0 mg/l for 1982. The peaks both in 1981 and 1982
 occurred in the spring and fall,  and are probably in-
 dicative of activities in the watershed corresponding
 to the residential nature of the watershed—seasonal
 fertilizing and lawn preparation.
      0.20


      0.15


      0.10


      0.05
              1981  f~]

              1982  ||
rh   m   m
         Sta. 2    Sta. 3    Sta. 4    Sta 5


 Figure 3.—Zinc: mean annual concentration.
       6.0


       5.0


   ~.   4.0


   I   3'°

       2.0


       1.0
              1981 j  j

              1982 ||
         Sla. 2    Sta 3    Sta 4    Sta 5    Sta. 6

Figure 4.—Total nitrogen: mean annual concentration.
     0.50O


     0.400


     0.300-


     0.200


     0.100-
               1981 [~~1

               1982 |   |
         Sta. 2    Sta 3     Sta. 4    Sta. 5     Sta 6

Figure 5.—Total phosphorus: mean annual concentration.
   Total Phosphorus. During 1981 and 1982, the total
 phosphorus concentrations  were  erratic and  no
 definite pattern  was  apparent other than that the
 highest values occurred in late summer and early fall.
   The  total phosphorous  levels  for October  1981
 samples were unduly high compared with the concen-
 trations many other studies showed in  freshwater
 lakes and  were six times  greater  than  the other
 highest concentration  of total  phosphorus  within
 Creve Coeur  Lake.  These aberrant values were  at-
 tributed to  a break  in a sewer line upstream of the
 lake. When the October 1981 data are removed  from
 the calculations,  the arithmetic mean values are  as il-
 lustrated in Figure 5.
   Values for both years were in the 0.200 to 0.300 mg/l
 range, with a slightly higher mean concentration oc-
 curring in 1982. Comparing the phosphorus data  by
 using a t-test (Sokal and Rohlf, 1969) did not indicate a
 significant difference between 1981 and 1982.

 Comparison  With Water Quality Standards

 Comparison of the results of the routine water quality
 monitoring program with Federal water quality criteria
 and Missouri waler quality standards is presented in
 Table 3.
   The Federal  criteria are those established  in the
 U.S. Environmental Protection Agency publication en-
 titled Quality  Criteria for Water (U.S. Environ. Prot
 Agency, 1976).
   Missouri  Water Quality Standards (Mo.  Dep.  Nat.
 Resour., 1977) are based on stream flow classification
 and beneficial water uses. Creve Coeur Lake falls into
 the category of Class P1 Waters. This classification
 includes standing water reaches and impoundments
 of permanent flowing streams. Beneficial water uses
 assigned  to  Creve  Coeur  Lake  due to the  P1
 Classification include the following:
   •  Maintenance of conditions to support health  in
 livestock and wildlife watering;
   •  Maintenance of conditions to sustain warm water
 fish and  other warm  water  aquatic  life  including
 critical stages of  reproduction and early life; and
   •  Maintenance  of conditions  for boating  and
 canoeing, where very little  contact with the water  is
 assumed.
   Two levels of criteria in the Missouri Water Quality
 Standards apply  to Creve  Coeur Lake. The general
 criteria dictate that the waters of the State of Missouri
 shall  be:
   •  Free from substances that will cause  the forma-
 tion of putrescent or otherwise objectionable bottom
 deposits;
   •  Free from oil, scum, and floating debris in suffi-
 cient  amounts to  be unsightly or deleterious;
   • Free  from  materials that cause  color, odor, or
 other conditions  in such  degree  as to  create  a
 nuisance; and
   • Free from substances  or conditions that have a
 harmful effect on  human, animal, or aquatic life.
  Specific criteria that apply to Creve Coeur Lake are
 found in Table A of the Missouri Water Quality Stan-
 dards for Class  PI Waters, under the headings of  Pro-
 tection of Aquatic Life and  Livestock-Wildlife Water-
 ing. Table 3 includes criteria from  Table  A for the
specific parameters listed.
  As  noted in Table 3, only zinc is below the limit of
the established  criteria. The remaining parameters ex-
ceeded  their criteria limits. Differences among  the
 mean values for the  remaining four parameters  bet-
ween  1981 and 1982 showed an increase in total phos-
                                                 420

-------
                                                              CASE STUDIES OF WATER QUALITY IMPROVEMENTS
phorus but a decrease in mercury, cyanide, and zinc,
and no change for phenol.

Trophic Status

Carlson's Trophic State Index (Carlson, 1977) was us-
ed  to classify the trophic conditions of Creve Coeur
Lake  during 1981 and 1982. The Trophic State Index
(TSI)  is a  numerical classification  that retains the
original meaning of the nomenclatural trophic system
by  using major divisions that correspond  roughly to
existing conceptions of trophic  groupings. For Creve
Coeur Lake, TSI data points were  computed using
total phosphorus values  (Fig. 6).
  As  illustrated by  Figure  6,  the  lake  could be
classified as eutrophic or, as the scale approaches
the upper level,  hypereutrophic. The  trophic state did
not change between 1981 and 1982, and,  therefore,
dredging did not appear to have any major impact.
Comparison of  the  trophic state for both years in-
dicates similar seasonal  trends.

Fish  Flesh Samples

The analyses conducted  for  mercury in various fish
from Creve Coeur Lake determined that the levels in
the fish were below the Food and Drug Administration
(FDA) recommended levels of 0.5 ppm or less in edible
fish tissue. The majority of the samples were less than
this level by a factor of 10. The results are  in Table 4.
  It is reassuring to note that all three types of fish
show  similar levels of mercury. This tends to indicate
                          that based on these samples, mercury contamination
                          is not occurring through bioaccumulation at this time
                          and that  overall  this  presents no  problem for the
                          fishery resources of the lake.
                                    | *PR | MAY | MH \ MIL \ MtB \  SEP |  OCT | MOV [ DEC
                                    1981	
                                                             OCT  NOV  DEC
                          Figure 6.—Trophic state index, Creve Coeur Lake.
                           Table 3.—Water quality criteria versus routine monitoring.

Total phosphorus (/jg/l)
Mercury (^g/l)

Phenol (ng/l)

Cyanide (^g/l)
Zinc (fig/I)
Federal
(USEPA)
25
0.05

1.0

5.0
5.0 mg/l
Missouri
(DNR)
—
0.05

1.0

5.0
100 mg/l
Creve Coeur Lake
1981* 1982*
261.0
0.18

10.0

14.0
93.0
324.0
0.14

10.0

10.0
7.40
Comments
—
Often reported as
less than 0.1 ^g/l.
Often reported as
less than 0.1 ^g/l.
—
1981 range: 40-140 ngl\
Fecal coliform bacteria
  (/100ml)
                                                                                   1982 range: 20-270 ^g/l
200
200
                       < 10- > 36,000    < 10- > 200,000
"Arithmetic annual mean values for five (5) stations through nine (9) months of sampling
                          Table 4.—Results of mercury analysis in fish flesh (mg/gm).
Species
Channel catfish
and black bullhead


Gizzard shad


White crappie


Largemouth bass


Sample

A
B
C
A
B
C
A
B
C
A
B
C
Mar.

0.10
0.09
<0.05
<0.05
0.11
<0.05



0.05
0.13
0.20
1981
June

0.06
0.05
<0.05
<0.05
<0.05
<0.05
0.05


0.07
0.08

Sep.

0.05
0.08
<0.05
0.06
0.05
<0.05



0.07
0.05
0.06
1982

0.03


0.04
0.04
0.04
0.04
0.03
0.04



1983

<0.05
0.08
0.05
<0.05


0.08





Green sunfish
Carp
                                                                    <0.05
                                                                      0.07
                                                 421

-------
LAKE AND RESERVOIR MANAGEMENT
                                     Table 5.—Sediment trap efficiency.
  May 10, 1983
  July 23, 1983
  June 4, 1983
Maximum inflow
Maximum inflow
Maximum inflow
Flow
(cfs,)

 31
700
 6!)
 Maximum
Suspended
  Solids
  Inflow

  7,000
  5,000
  3,000
 Maximum
Suspended
  Solids
 Outflow

     25
    200
  1,000
AVERAGE
Percent
Trapped

  96
  96
  66
  86
Sediment Loading

The initial lake rehabilitation study identified much of
the sediment  loading from  the watershed of  Creve
Coeur Creek as a result of development activities. Part
of the monitoring program analyzed sediment loading
and sediment trap efficiency from three storm evenls
by comparing  maximum suspended solids and max-
imum flow at the inflow and outflow sampling stations
(Table 5). The average sediment trap efficiency was 86
percent resulting from the three storm events.
  This calculated trapping efficiency of 86 percent
compares favorably with the work of Brune (1953), who
related sediment trap efficiency to the capacity-inflow
ratio.  Using Brune's graph, the sediment trap efficien-
cy would be approximately 92 percent.
  The sediment production rate for Creve Coeur Lake
identifies the amount of sediment produced per year.
Gottschalk (1964) presented a formula for sediment
production for  the Missouri River Basin which is equal
to the watershed area raised to the 0.8 power.  For
Creve Coeur this would amount to:

    SPR = 27.508  = 1.748 x 10" m3/year

Using 1.750 x  104m3/yearas the sediment production
rate for the Creve Coeur Creek watershed and using a
90 percent sediment trapping  efficiency, expected
sediment accumulation in Creve Coeur Lake would be
at a rate of 1.575 x 104 m3/year. At this rate, the 3.7 :<
106 m3 lake would require over 100 years before one
half of the lake volume was lost through sedimenta-
tion and  over 50 years before one fourth of the lake
volume was filled in with sediment.
  A second approach involves the arithmetic average
of 205 measurements for annual sediment production
rate on watershed sizes ranging from 26-260 km2. The
sediment production  rate for watersheds this size was
7.61  x 102 m3/km2/year (Gottschalk, 1964). With  the
71.2 km2 watershed for Creve Coeur Lake, this would
amount to 5.42 x 10* m3/year. Using a 90 percent sedi-
ment  trap efficiency, the  sediment production rate
would be approximately 4.88 x 104 m3/year and  the
lake would be filled with sediment in approximately 75
years.
  Using an average of the  two rates  results in a sedi-
ment production rate  of 3.23 x 104m3/yearwiththe90
percent efficiency trap. At this rate, the lake would not
completely fill  in with sediment for more than approx-
imately 115 years. In actuality, this would probably be
                                even longer because the sediment  trap efficiency
                                becomes less as the lake capacity is reduced. Thus,
                                the time that it would take the lake to lose one half of
                                its capacity would be in the range of 60 to 80 years.
                                SUMMARY
                                The dredging of Creve Coeur Lake has  provided a
                                number of benefits for the users  of the lake and sur-
                                rounding park. The enlarged and deepened  lake pro-
                                vides excellent  boating and fishing areas. The park
                                setting has  improved overall from the  increase  in
                                defined shoreline and creation of an island for wildlife
                                habitat. Areas surrounding the lake that were used for
                                settling  basins  of  the  dredged material  have been
                                graded and seeded and will increase the usable park
                                area surrounding the lake.
                                  Water Quality. The water quality in Creve Coeur
                                Lake did not change during dredging operations and
                                following year. Only 1 year of monitoring has been
                                conducted following the dredging and some changes
                                may be  noticed  by more extended monitoring. The
                                lake is classified as eutrophic, based on the levels of
                                total phosphorus in the lake; however, as equilibrium
                                is attained following the removal of the nutrient-rich,
                                shallow  sediments,  a  reduction in the  amount  of
                                nutrients in the water may become  evident.
                                  The water quality parameters monitored during and
                                following  dredging did  not  demonstrate  any dif-
                                ferences  of  water  quality that  could  be  linked
                                specifically to  the dredging activities.  The initial
                                primary benefit of the dredging is increased usability
                                of the lake.
                                REFERENCES

                                Brune, G.M. 1953. Trap efficiency of reservoirs. Trans. Am
                                  Geophys. Union. 34: 407-18.
                                Carlson, R.E. 1977. A trophic state index for lakes. Limnol.
                                  Oceanogr. 22: 361-8.
                                Gottschalk, LC. 1964. Reservoir sedimentation. Pages 17-1
                                  to 17-34 in V.T. Chow, ed. Handbook of Applied Hydrology.
                                  McGraw Hill, New York.

                                Missouri  Department of Natural Resources. 1977. Missouri
                                  Water Quality Standards.
                                Sokal, R.R., and F.J.  Rohlf. 1969. Biometry. W.H. Freeman,
                                  San Francisco.
                                U.S. Environmental Protection Agency. 1976. Quality Criteria
                                  for Water. Washington,  D.C.
                                                 422

-------
RESERVOIR MANAGEMENT PLANNING:
AN ALTERNATIVE TO REMEDIAL ACTION
DONALD W. ANDERSON
Tennessee Valley Authority
Chattanooga,  Tennessee

            ABSTRACT
            Tellico Dam and Reservoir were constructed by the Tennessee Valley Authority (TVA) as a multipur-
            pose water resources project on the Little Tennessee River in eastern Tennessee. Purposes of this
            project include industrial and residential shoreline development, recreation, water supply, navigation,
            flood control, and power generation. To achieve these objectives without sacrificing the high water
            quality in Tellico Reservoir, TVA has undertaken a water quality management program to guide reservoir
            development. The resulting reservoir water quality management plan will recommend maximum
            allowable waste loads, industrial siting requirements, water use classifications and criteria, and fisheries
            management  This m conjunction with a reservoir land use plan, will provide comprehensive develop-
            ment guidelines for Tellico Reservoir. Implementation of the plan is to be accomplished through a
            unique blend of regulatory and nonregulatory measures carried out by TVA and the State of Ten-
            nessee. To facilitate incorporation of its recommendations into the water quality regulatory program,
            the reservoir water quality management plan will be formally adopted as a portion of the State Water
            Quality Management Plan. Through its authority as manager of reservoir lands, TVA will implement
            nonpoint source control recommendations for which the State lacks regulatory authority. This integration
            of environmental planning into Tellico Reservoir development will provide for intensive uses of the
            reservoir's land and water resources while maintaining high levels of water quality
INTRODUCTION

Restoration of lake or reservoir water quality to an ac-
ceptable level is difficult. The techniques are frequent-
ly expensive and of uncertain effectiveness. Patterns
of use  or  abuse,  once established, are difficult  to
change. Where water quality  conditions are accep-
table, a prudent course of action is to prevent water
quality degradation and thereby avoid the difficult and
uncertain task of restoration.
   Tellico Reservoir, on the Little Tennessee  River near
Knoxville, Tenn., offers the opportunity to develop and
manage a major water resources project in a manner
which will provide a variety of economic benefits while
maintaining a high level of water quality. Completed in
1979, Tellico Reservoir was designed as a multipur-
pose project with planned benefits from recreation,
navigation,  flood  control,  power  generation, water
supply,   and  industrial  and   residential  shoreline
development. Table 1  provides a summary of the prin-
cipal features of Tellico Reservoir. The nature of the
watershed, the high quality of inflowing waters, and
the absence of pollution sources combine to produce
a  reservoir with excellent water quality. To maintain
this while achieving the planned benefits of the Tellico
Reservoir project, the  Tennessee Valley Authority has
undertaken a program of water quality management.
Its primary  components are  a post-impoundment
water quality survey,  use of reservoir water quality
models, and various management recommendations
that guide  the continuing development and manage-
ment of the reservoir.
TELLICO LAND USE PLAN

In  addition  to  these  components,  a  separately
developed  land  use  plan  for  reservoir properties
served to identify future residential,  industrial, and
recreational development patterns that might impact
reservoir water quality (Blackburn et  al.  1981).  Adja-
cent lands amounting to approximately 22,000 acres
were acquired by the Tennessee Valley Authority for
planned development. Figure 1 identifies the reservoir
lands selected for industrial development, residential
              LITTLE TENNESSEE RIVER BASIN  TELLICO RESERVOIR LOCATION IN
             LOCATION IN TENNESSEE RIVER BASIN LITTLE TENNESSEE RIVER BASIN
Figure 1.—Location map of Tellico Reservoir.
                                                 423

-------
LAKE AND RESERVOIR MANAGEMENT

development, and recreation/open space uses. Lands
designated  for residential development  occupy the
western shore of the lower part of the reservoir from
Little Tennessee River miles 4.0 to 13.0. Between Little
Tennessee River miles 13.0 and 19.0 and occupying
both  sides  of  the  reservoir are  tracts  of land
designated as industrial  sites. The remainder of the
reservoir lands were allocated to a variety of low im-
pact uses collectively identified in Figure 1 as Recrea-
tion/Open Space.

POST-IMPOUNDMENT SURVEY
A  post-impoundment survey  documenting  baseline
water quality and biological conditions following im-
poundment of the new reservoir was conducted during
the 16-month  period from June 1980 to September
1981. Results show water quality to be good in almost
every respect and to meet U.S. EPA and Tennessee
water quality criteria and guidelines (Sagona et  al.
1983). Water flowing into Tellico  Reservoir from im-
poundments  and unregulated watersheds  upstream
retains many of  the characteristics of the  mountain
streams in the area. These basic characteristics make
Tellico Reservoir suitable for domestic and industrial
water supply, fish  and  aquatic  life,  swimming and
other forms  of  water-contact recreation,  livestock
watering and wildlife, irrigation, and where possible,
navigation.
  The waters of Tellico Reservoir are soft, poorly buf-
fered, slightly acidic to  near neutral,  and low in
dissolved solids, suspended solids, color, turbidity,
and  most metals.  Available  nutrients,  particularly
phosphorus,  are  very low  in the  reservoir, often
measuring at the analytical detection limit. Bacterial
concentrations are  very low and  meet generally ac-
cepted criteria for water-contact recreation. The con-
sensus of those who have studied the reservoir is that
Tellico is one of the more desirable and attractive im-
poundments in the  region from a  water  quality per-
spective.
  Thermal  stratification  of Tellico  Reservoir begins
during April and extends from Tellico Dam to about
Little Tennessee River  mile  30 (LTRM 30). By late
September deslratification  begins  and  isothermal
conditions  exist in the reservoir from  late October
through  March. During  thermal stratification, low
levels (less than 5  mg/l) of dissolved oxygen (DO) are
observed in the hypolimnion from the dam to between
LTRM 10 and 20. By September, a zone of very low DO
(less than 0.5 mg/l) in the lower hypolimnion extends
from Tellico Dam to near LTRM 18. This  DO depletion
results from  low levels of background biochemical
and chemical oxygen demands and a moderate sedi-
ment oxygen demand. This pattern of  stratification
and  DO depletion  will likely remain characteristic of
Tellico Reservoir and influence reservoir  processes
such  as  metal   solubility,  nutrient  cycling,  and
chemical and biological oxygen demands.
  Phytoplankton communities in Tellico  Reservoir are
quite diverse, partly because this is a new impound-
ment and partly because of the variety of habitats in
the reservoir. Genera found range from  those typical
of ponds or shallow pools on the Tellico River embay-
ment to those commonly found in mature lentic envi-
rons on the Little Tennessee River embayment. Gen-
erally, the lower and middle sections of  the reservoir,
below LTRM  21.0  and Tellico River mile 3.0 (TelIRM
3.0), contain higher algal  standing crops than the up-
stream sections. Spatial differences in phytoplankton
production  generally influence the fish population in
that  higher numbers of both  phytoplankton and fish
occur in the epilimnetic layers of the lower sections of
the  reservoir.  Overall,   Tellico exhibits   adequate
temperature  and  DO to  maintain viable  warm and
coldwater fisheries most of the year, although condi-
tions are not  ideal  for either group.
  The post-impoundment survey has established the
natural or baseline water quality against which future
changes in  waler  quality  from  reservoir aging or
management strategies may be judged.  It further pro-
                           Table 1.—Tellico Reservoir: summary of principal features.8

 Location

 Little Tennessee River 0.3 mile above confluence with the Tennessee River; 33.2 miles downstream from Chilhowee Dam.

 Streamflow

 Drainage area above dam:
   Total	2,627 sq miles
   Uncontrolled (below Fontana Dam)	1,056 sq miles

 Reservoir

 Operating levels at dam
   Maximum used for design	El. 817.5
   Maximum probable (126,000 cfs)	El. 817.8
   Top of gates (area 17,300 ac.)	El. 815
 Normal operation:
   Full pool (area 16,500 ac.)	El. 813
   Minimum for navigation, flat pool  (area 14,200 ac.);
     (same El. as Fort Loudoun Reservoir)	El. 807
   Backwater length at full pool	32.2 miles
 Length of shoreline at El. 813;
   Main shore	302 miles
   Islands	8 miles
   Total	310 miles

 Storage (flat pool assumption):
   Total volume at El. 815	447,300 ac.-ft

 Controlled flood storage
   (El. 815-807)	-126,000 ac.-ft

 aTenn Valley Auth Off Eng Design Construe
                                                 424

-------
                                                            CASE STUDIES OF WATER QUALITY IMPROVEMENTS
vided the data by which reservoir water quality models
could be calibrated and subsequently used to analyze
the reservoir's response to a variety of stimuli.

WATER QUALITY MODELING

Initial modeling efforts centered on application of an
adaptation of the  one-dimensional  (vertical)  water
quality model, CE-QUAL-RI, developed by the U.S. Ar-
my Corps of Engineers Waterways Experiment Sta-
tion. This model was used to investigate potential ef-
fects of reservoir aging on hypolimnetic dissolved ox-
ygen levels and to  determine what factors most in-
fluence the dissolved oxygen balance (Mauser et al.
1981). The  presence of pronounced longitudinal as
well as vertical water  quality  gradients  in Tellico
Reservoir limited the usefulness of the one-dimension-
al model to these relatively simple analyses.
   To more accurately assess the  impact of reservoir
development, a two-dimensional reservoir  model was
applied (Mauser, 1982). The model was applied first to
the entire reservoir and subsequently to the vicinity of
the  industrial sites at a higher level of  resolution.
Following model calibration, hypothetical waste loads
were applied. The  magnitude of the waste  loads as
well as their  vertical and longitudinal locations were
varied and reservoir  responses were compared to
baseline conditions.  The  primary  focus of  these
analyses was the dissolved oxygen response to waste
inputs over the stratified period of approximately April
through September.  Model output was screened to
identify reservoir segments exhibiting the most signif-
icant deviations from baseline conditions.

WATER QUALITY MANAGEMENT PLAN

The land-use plan,  post-impoundment survey and
mathematical modeling components described here
provided  the  basic information by which potential
water quality impacts could be  assessed. The  basic
reservoir information and assessment were then used
to develop  a water quality management  plan. This
stage of the planning process presents an opportunity
to favorably affect  the course of reservoir develop-
ment and protection. It is this water quality manage-
ment plan that integrates the various reservoir study
components into a unified approach designed to pro-
tect water quality  and to provide an alternative to
future remedial action in the reservoir. Recommenda-
tions m the management plan were proposed to ad-
dress the most common questions regarding appropri-
ate water use classifications, nonpoint source control
needs, effluent limitations for discharges, and siting
of water intakes and waste discharges. Since these
recommendations have not yet  been finalized, they
are presented here  in a general form.  Final recom-
mendations of the management plan will, however,
contain sufficient specific information to guide reser-
voir management decisions.
  •  Because  Tellico Reservoir exhibits a variety of
physical features including both riverine and reservoir
characteristics, the reservoir should be segmented in-
to three reaches for the purpose of identifying ap-
propriate use classifications and establishing appro-
priate water quality criteria.
  • Industrial waste discharges should  be restricted
to that portion of the reservoir adjacent to the lands
presently designated as industrial sites.
  • That portion of the reservoir  immediately  adja-
cent to and downstream of the designated industrial
sites should not be classified for domestic  water sup-
ply use to avoid potential use conflicts.
  •  Based on temperature and dissolved oxygen pat-
terns observed  in the lower portion of the reservoir,
this  portion of the reservoir should not be classified
and  protected as trout waters. The riverine portion of
the  reservoir above the designated industrial sites
should be considered for trout waters designation.
  •  Waste  discharges  into reservoir  embayments
should be prohibited.
  •  Wastewater discharges should be restricted to
specific maximum and minimum elevations selected
to reduce potential impacts.
  •  Maximum  daily loads  for  ultimate oxygen de-
mand, nitrogen, and phosphorus should  be  estab-
lished.
  •  Initial dilution requirements should be developed
for major discharges to prevent localized water quality
and  aesthetic impacts.
  •  In view of the uncertainty inherent in the predic-
tion  of water quality  impacts, point source control
recommendations should be reevaluated when waste
discharges reach a predetermined level.
  •  Nonpoint source impacts from development and
use of reservoir lands should be minimized by apply-
ing appropriate best management practices for agri-
cultural, forestry, and construction activities.

IMPLEMENTATION OF
RECOMMENDATIONS

The preparation of such recommendations remains of
little practical benefit unless proper actions are taken
to assure implementation. To implement the Tellico
Reservoir water quality management plan, a mixture
of regulatory and nonregulatory approaches has been
selected.
  The Tennessee  Valley Authority does  not have
direct regulatory authority over waste discharges into
its  reservoirs.  In  Tennessee  such authority  and
responsibility is entrusted to the Tennessee Depart-
ment of Health and Environment (TDHE). Therefore,
the recommendations imposing discharge limitations
must be  implemented primarily by the TDHE. To
facilitate this aspect of plan implementation, TVA will
request that the TDHE formally adopt the plan recom-
mendations as  a part of its Statewide Water Quality
Management Plan. Upon adoption, these recommen-
dations could then be used as the basis for regulatory
decisions.
  TVA does have  limited authority under Section 26a
of the TVA Act for activities such as the placement of
a wastewater discharge or  water intake  which could
potentially impact the operation of  a  reservoir.  Fur-
thermore, TVA  has  the right of a property owner to
regulate use of the lands under its ownership. TVA will
use  this  authority to assist in plan implementation
particularly with respect to  nonpoint source controls
for which the TDHE lacks regulatory authority.
  To facilitate  the development of  reservoir proper-
ties, TVA  has transferred large tracts of land  to the
Tellico Reservoir Development Agency (TRDA). Under
a contractual agreement, TRDA will be responsible for
day-to-day management of these properties. Develop-
ment standards contained in the TRDA/TVA contract
will  ensure that  plan  recommendations affecting
these transferred properties will be recognized.


SUMMARY

The Tennessee  Valley Authority, in  cooperation with
the State of Tennessee, has undertaken an extensive
program of water quality management planning aimed
                                                425

-------
LAKE AND RESERVOIR MANAGEMENT
at preserving the excellent water quality of Tellico
Reservoir.  Management  recommendations are being
developed  which, when implemented, should  mini-
mize the water quality impacts which would otherwise
result from the planned development of reservoir pro-
perties. While a water quality management plan can-
not guarantee that water quality goals will always be
met, it can provide a flexible tool for  systematically
assessing  potential  impacts and  identifying  control
measures.  In the  case of Tellico  Reservoir,  it is be-
lieved that  this approach provides an excellent oppor-
tunity to prevent water quality  degradation and thus
avoid  the  need for costly and  difficult  remedial
measures.
REFERENCES

Blackburn, W.W., et al. 1981. Tellico  land use plan. Tenn.
  Valley Author.

Mauser, G.E. 1982. Results of waste load allocation models
  for Tellico Reservoir. Rep. No. WR28-2-65-103. Tenn. Valley
  Auth. Water Sys. Dev. Br. Morris, Tenn.
Mauser, G.E., S.L McCarley, and R.T. Brown. 1981. Post-im-
  poundment  modeling of  Tellico Reservoir water quality.
  Rep.  No. WR28-1-65-102.  Tenn. Valley Auth., Water Sys.
  Dev. Br. Norris, Tenn.

Sagona, F.J., A.M. Brown, D.L Dycus, and W.L Poppe. 1983.
  Tellico  Reservoir postimpoundment water  quality.  Rep.
  No. TVA/ONR/WR-83/6. Tenn. Valley Auth., Water Qual. Br.
  Chattanooga, Tenn.
                                                 426

-------
                                                       Trophic  Status
THE TROPHIC STATE CONCEPT:
A LAKE MANAGEMENT PERSPECTIVE
ROBERT E. CARLSON
Department of Biological Sciences
Kent State University
Kent, Ohio
           ABSTRACT

           Trophic state is fast becoming a nonconcept because of the confusion over its proper definition.
           The lack of agreement to trophic definition stems from its early typological origins, the fusion of
           causal factors with the resulting biological condition, and the assumption of the complex nature
           of the concept. A medical analogy of obesity is used to illustrate that trophic state can easily
           and unambiguously be determined by the measurement of plant biomass alone.
The term "trophic state" and its many ancillary terms,
such as "eutrophic" and "oligotrophic," have been in
the limnological literature since Naumann first used
them in lake classification in 1919. Today there is still
no general agreement as to what the terms mean. This
lack of common definition would probably be accep-
table  if the concept of trophic state were in some
backwater of  limnological terminology and research.
These terms, however, are frequently used in both lake
classification and  theoretical limnology  and have
found their way into the language of the lay public,
engineers, and government officials. The manner in
which these terms  are defined and used can signifi-
cantly  shape the  underlying assumptions of  our
research, our attitudes toward lake restoration, and
even the direction of funding  by governmental agen-
cies.
  The problem with trophic state is not the lack of a
definition but the  overabundance of definitions.
Trophic state can measure potential nutrient inputs of
the watershed (Hutchinson,  1969),  or the  rate of
 nutrient input (Beeton and  Edmondson,  1972); it can be
defined biologically as primary  productivity (Aberg
and Rodhe, 1942) or algal  biomass (Carlson, 1977). It
can be the shape of an oxygen curve, the rate of hypo-
limnetic oxygen depletion, or the presence or absence
of a particular species of plant or animal. Often any or
all of these criteria are combined into a single defini-
tion with varying  emphasis in either nutrients or
biology. For many, trophic state has become a hybrid
concept,  in which nutrients and  biology are  inex-
tricably combined (Brezonik and Kratzer, 1982).
  Trophic state is fast becoming a nonconcept, in the
sense of  Hurlbert (1971), as it loses its distinctness
under a plethora of conflicting definitions. Perhaps it
will wither until terms like "eutrophic" will be used on-
ly to lend an air of scientific erudition when speaking
to laymen. Perhaps it is time the terms disappeared;
the terms are remnants of a long since discarded lake
typological classification.  Perhaps these  terms also
no longer serve any scientific purpose.
  It is my opinion that the problem is not so much that
the concept is old or useless as much as that we are
attempting to cling to the outdated typological defini-
tion of the  concept. We are attempting  to classify
lakes as though there are distinct lake types, probably
as distinct as the types that discriminate between
species.  In early typological classification it was
necessary only to identify  those aspects  of the lake
which allowed it to be distinguished from other lake
types. This produced a list of characteristics shared in
common  by that lake type, such as nutrient content,
                                              427

-------
 LAKE AND RESERVOIR MANAGEMENT
 productivity, shape of the oxygen curve, the bottom
 fauna, and so on. In these classifications one char-
 acteristic  was not necessarily more important than
 another; all the characteristics taken as  a whole
 classified  the lake. To preserve the original intent of
 the trophic state concept today, we would have to use
 a multifactorial approach to pigeonhole the lake into a
 classification unit.
   The lack of a clear distinction between the impor-
 tance of nutrients and  biological condition  in  the
 original  typologies has contributed to the confusion
 that exists today.  Some existing trophic state defhi-
 tions have fused  the cause of a condition with its
 resulting effect on productivity within  the  lake:  An
 oligotrophic lake has low productivity caused by low
 nutrient  inputs; a eutrophic one has high productivity
 caused by high nutrient inputs. If a lake had low pio-
 ductivity caused by some factor other than nutrienls,
 it would be classified as a different lake type, not as
 oligotrophic. With these definitions, it  would  be  im-
 possible for a lake to be eutrophic as the result of any
 factor other than high nutrient concentrations.
  These definitions are justifiable today only if it can
 be  demonstrated  that nutrient loading is  the scle
 cause of high productivities in lakes. Imagine for in-
 stance that,  in spite of high nutrient loadings, a lake
 remained clear because of intense zooplankton graz-
 ing.  It could not be labeled oligotrophic because, by
 definition,  oligotrophy  is caused by  low nutrients. If
 pesticide inputs  killed  the zooplankton community
 and there were a sixteenfold increase in algal biomass
 (Hurlbert, 1971), this could not be called eutrophica-
 tion: changes in nutrients were not involved. The dif-
 ficulty of situations such as these illustrates the prob-
 lem of insisting that cause and effect must be fused in
 trophic classification.
  Another   problem  hindering  the  clarification  of
 definition is that we tend to expect the term "trophic
 state" to  be an all-inclusive  explanation to every
 phenomenon  happening  within  the  lake  system.
 Naumann (1927) stated that the lake's  biology as a
 whole was related to the amount of nutrients in the
 water. The often-reprinted diagram of Rawson (1939) il-
 lustrating the factors affecting primary productivity
 served to reinforce the idea that the concept of trophic
 state is complex and therefore difficult to measure.
 The  idea that numerous factors related to  climate,
 geology,  hydrology, and human activity interact to pra-
 duce an  effect in numerous components of a lake's
 biological system reinforces the concept of the com-
 plex nature of trophic state. When  faced with a com-
 plex system to classify, the tendency is to resort to in-
 dices to simplify the undertaking.
  Attempts to extricate  cause from  effect  or  lo
 minimize the underlying complexity  of  the concept,
 have  resulted in single factor definitions of troplrc
 state. These attempts have generally met with limited
 success. Nutrient-related  definitions  cannot pleas.e
 those that consider the lake's biological condition lo
 be important, and as these indices almost universally
 use phosphorus as the standard,  the  classificaticn
 becomes useless in  lakes limited by other nutrienls
 such as nitrogen (Kratzer and Brezonik, 1981). To use
 these indices requires  the determination of what
 nutrient limits the particular system prior to classifica-
tion.
  Biotic  indices also meet  resistance.  Primary pro-
ductivity  has been suggested as a basis for trophic
classification (Aberg and Rodhe, 1942; Smith,  1979),
 but  productivity-based classifications  break  down
when there is high  biomass but low productivity as is
 found  in  arctic lakes (Kalff and  Welch, 1974).  As
 nutrients are, by definition, the cause of the biological
 condition of the lake, there is also an uneasiness that
 definitions or indices disregarding nutrients may mis-
 classify a lake. If a lake were to be clear because of in-
 tense zooplankton grazing, there might be an objec-
 tion to the classification of  that lake as oligotrophic
 (Osgood, this vol.). Fish productivity would probably
 be very high, in contrast to  what would normally be
 found in an  oligotrophic lake. Such a lake might be
 classified  as  oligotrophic  on   the basis of  algal
 biomass, and eutrophic on the basis of productivity
 and nutrient loading.


 A NEW DEFINITION

 It is clear  that trophic state must  be redefined in a
 simpler, clearer  manner than  presently exists before it
 loses all credibility as a useful classification system.
 Such a new  definition should (a) be clear and unam-
 biguous, (b) provide for easy evaluation in the field, (c)
 preserve the underlying assumption of complexity,
 and (d) separate cause from  effect, yet retain the im-
 plication that nutrient supply is the common driving
 force of trophic  state.
   Two possibilities exist.
   I have previously suggested  (Carlson,  1979) that
 system terminology could be used to  better define
 trophic state. Certainly we are working with a multi-
 component system. The term system state could  be
 substituted for  trophic state, and the  state of each
 component could  be  measured  in units of energy or
 biomass. Changes in the rate of nutrient or energy in-
 puts can cause certain system responses which could
 be classified according to the oligotrophic-eutrophic
 scale. Such an approach allows the trophic  classifica-
 tion of the system on the basis of the state of the
 system and, separately, classifies on  the basis  of
 energetic  or   nutrient  inputs.  Systems  theory
 recognizes state, productivity, and nutrient inputs as
 three separate measures of the nature of the system,
 and  does not attempt to force  them  into a  single
 classification scheme. Although  this approach solves
 some of the problems  associated with trophic defini-
 tion, the problem  of complexity remains.  You must
 still deal with the entire system and thus measure all
 components  of the system or resort to indices. The
 trophic concept  is relabeled and  updated, but  funda-
 mental problems remain.
   Cooper  (1978)  suggested  that  ecosystems are
 analogous  to organisms, and, as  such, ecosystem
 pathologies could  be considered in a manner similar
 to human disease. The idea that both the ecosystem
 and  the human body are systems subject  to and
 responding to perturbations  does provide  for  a new
 perspective as to the definition of trophic state.
  Consider that  the process of eutrophication may be
 analogous to the gain  in weight  leading  to the condi-
 tion termed obesity. It should not be an entirely unac-
 ceptable  analogy  as  "eutrophic"  does  mean
 "well-fed."  Eutrophy could be defined as that  condi-
 tion produced when the input of carbon or  energy of
the green  plants  is greater than  its  utilization  in
 respiration  and grazing, resulting in a net increase  in
 plant biomass. Trophic state could be measured as
the amount of plant biomass (in carbon or  energetic
 units) within the  lake.
  The  comparison of  eutrophication  with obesity
forces us to consider that eutrophication, like obesity,
 is a pathological  condition of the lake system. As with
obesity,  the  cause of the  condition  may not  be
                                                 428

-------
                                                                                        TROPHIC STATUS
necessarily a problem of high nutrient inputs; it can
also be the result of the degree to which the carbon is
used within the system. Such a definition would avoid
the present tendency to automatically assume that all
eutrophy is related to high nutrient inputs. Similarly, in
the sense that obesity  can be  controlled  by either
dieting or exercise,  control of nutrient inputs is only
one of a number of possible restoration techniques
available. It may be  that in some cases, manipulation
of the grazers may  be more feasible  (Shapiro, 1980;
Carlson and Schoenberg, 1983).
  The comparison of eutrophy with obesity also sug-
gests that the  notion  of complexity can still  be
preserved with this restriction of trophic classification
to plant  biomass alone. Obesity is a  specific, well-
defined condition, yet it  can have effects on the total
human system. A doctor,  however, does not have to
measure every aspect  of  the  human condition to
diagnose obesity. By analogy, eutrophy can be defin-
ed solely by the amount of plant biomass, yet changes
in plant biomass can have both direct  and indirect ef-
fects  on the total aquatic system. By simply recogniz-
ing that  the  state  of the  plant component  can in-
fluence the entire system, the essence of complexity
ascribed to trophic state is maintained. As such, lakes
can be classified without the necessity of measuring
every possible system component or  resorting to in-
dices.
  Perhaps most important, this redefinition of trophic
state in medical terms is  in accord with our present
view  of  the  lake  manager as  a practitioner who
assesses the condition  of lakes and  prescribes ap-
propriate measures to either protect or  restore a lake's
health. The power of this new trophic definition is that
classification  is  not inextricably intertwined  with
causation,  and therefore,  techniques  of restoration
may not  be necessarily  nutrient-related. This  trophic
state definition serves  to separate  diagnosis  into
three distinct steps: first, the identification of the
severity of  the condition (classification), second, the
search for causation, and finally,  the recommendation
of a technique for restoration.

IMPLICATIONS AND IMPLEMENTATION

Acceptance of a medically-based analogy for trophic
state  will not necessarily cause major  changes in our
methods  of  lake   classification  because  plant
biomass,  especially algal  biomass,   is already  a
popular basis for classification. Algal biomass  has
provided useful and accepted trophic  classifications
under the guise of  trophic state indices. Apparently
we  are willing to accept biomass as a trophic state
variable as long as the term index is used, thus preser-
ving the illusion that some greater concept is actually
being measured.
  The problem of complexity and the interconnected-
ness of system variables is already being addressed.
Empirical relationships have already been derived for
hypolimnetic  oxygen depletion (Walker, 1979; Welch
and Perkins, 1979) and for fish productivity (Jones and
Hoyer, 1982). Such empirical relations do not so much
underscore the complexity of trophic state as they il-
lustrate that changes in lake structure and metabo-
lism are related to changes in algal biomass, and that
algal biomass is generally a reliable predictor of sys-
temwide  changes in the  lake.
  A problem with existing biomass indices  is that
they ignore changes in macrophyte  biomass. This
reluctance  to  consider  macrophyte  biomass  is
reflected not  only in our indices, but  in our nutrient
loading  models  as  well. Prediction of trophic state
relies on the relationship between nutrient loading
and algal chlorophyll. The importance of macrophytes
in the estimation of trophic state and in the accelera-
tion  of eutrophication  has probably been far under-
estimated.  Carpenter  (1981) produced  a  eutrophica-
tion  model  that implied a  positive feedback between
macrophytes and both algal biomass and the rate of
filling-in of the lake basin. Such a model suggests that
factors that affect the extent of macrophyte growth,
such as the amount  of  silt  income  (Cooke,  pers.
comm.), may be  more important  to  eutrophication
than nutrient loading itself.
  A  practical problem  remains for a biomass-related
trophic  state definition.  Biomass  itself is an  am-
biguous term, and it can be measured or estimated by
using a number of variables, almost all of which have
some inherent difficulties. Chlorophyll pigments are
most commonly  used because  the  pigments are
specific to  plants, but chlorophyll concentration per
cell is known to change with algal species or environ-
mental  conditions. Carbon  or caloric units would
seem preferable  as a definitional base of  trophic
state, but their practicality for  use in  actual  trophic
state estimation is limited by the amount of  detrital
material found in the water and on the macrophytes.
Nicholls and Dillon (1978)  suggested that cell  volume
is the best  alternative to measuring carbon  or caloric
content  of  phytoplankton  biomass.  Macrophyte
volume also could be  easily accomplished.  It  may be
that  empirical relationships  would have to  be estab-
lished between a calorie-based trophic state definition
and  other,  more  easily  used variables,  such  as
chlorophyll or biovolume.
  The greatest remaining problem with  a single factor
trophic definition is the potential for  misclassifica-
tion. Although a biomass definition will generally cor-
relate well with either  nutrient loading or productivity,
it will be the deviant lakes with  low biomass but high
nutrient loading that will cause the most objections. If
we return to the obesity analogy,  a possible solution
to such misclassifications is apparent. If a person has
a high caloric intake, but exercises heavily, he may not
be classified as obese, but his  potential  to be obese
exists if he stops exercising. Lakes with high nutrient
inputs but equally high internal utilization of the  plant
biomass may be  considered oligotrophic, but with a
high potential for eutrophy. This could produce a dual
classification system, one based on the actual condi-
tion  of the  lake,  termed trophic  state, and another
term based on the nutrient loading or nutrient content,
termed  potential state. Such a  dual  classification
system might overcome some of the objections of en-
tirely biomass-based  classification system, yet still
allow for an unambiguous classification system.
 REFERENCES

 Aberg, B., and W. Rodhe. 1942. Uber die Milieufaktoren in
  einigen sudschwedischen Seen. Symb. Bot  Ups. 5:1-256.

 Brezonik, P.L., and C.R. Kratzer. 1982. Reply to discussion by
  Victor W.  Lambou, "A Carlson-type trophic state index for
  nitrogen in Florida lakes." Water Resour. Bull. 18:1059-60.

 Carlson, R.E. 1977. A trophic state index for lakes.  Limnol.
  Oceanogr. 22:361-9.
 	1979. A review of the philosophy and construction
  of trophic state indices. Pages 1-52 in T.E. Maloney, ed.
  Lake  and  Reservoir  Classification  Systems.
  EPA-600-3-79-074. U.S. Environ. Prot. Agency. Washington,
  D.C.
                                                 429

-------
 LAKE AND RESERVOIR MANAGEMENT

 Carlson, R.E., and S.A. Schoenberg. 1983. Controlling blue-
   green algae by zooplankton grazing. Pages 228-33 in Lake
   Restoration,  Protection,   and  Management.   EPA
   440/5-83-001. U.S. Environ. Prot. Agency. Washington, D C.
 Carpenter,  S.R. 1981. Submersed vegetation: an  internal
   factor in lake ecosystem succession. Am. Nat. 118:372-83.
 Cooke,  G.D. 1983. Pers. comm. Dep. Biolog. Sci. Kent State
   Univ., Kent, Ohio.

 Cooper, W.E.  1978. Systems prediction: the integration  of
   descriptive,  experimental and  theoretical approaches
   Ohio  J. Sci. 78:186-9.

 Beeton, A.M., and W.T. Edmondson. 1972. The eutrophication
   problem. J. Fish. Res. Board Can. 29:673-82.
 Hurlbert, S.H. 1971. The nonconcept of species diversity: a
   critique and alternative parameters. Ecology 52:577-86.
 Hurlbert, S.H., M.S. Mulla, and H.R. Willson. 1972.  Effects  of
   an  organophosphorus insecticide on the phytoplankton,
   zooplankton, and insect populations of fresh-water ponds
   Ecol.  Monogr. 42:269-99.

 Hutchinson, G.E. 1969.  Eutrophication,  past and present.
   Pages 17-26 in Eutrophication: Causes,  Consequences,
   Correctives. Publ. 1700. Natl. Acad. Sci. Washington,  D.C.
Jones, J.R., and  M.V. Hoyer. 1982. Sportfish harvest  pre-
   dicted by summer chlorophyll  a concentration  in Mid-
   western  lakes and  reservoirs.  Trans.  Am. Fish  Soc
   111:176-9.

Kalff,  J., and H.E. Welch. 1974. Phytoplankton production  n
  Char  Lake, a natural polar lake, and in Meretta Lake,  a
  polluted  polar  lake, Cornwallis  Island, Northwest  Ter-
  ritories. J. Fish. Res. Board Can. 31:621-36.
Kratzer, C.R., and P.L Brezonik  1981. A Carlson-type trophic
  state index for nitrogen  in Florida lakes.  Water Resour
  Bull. 17:713-15.

Naumann,  E 1927. Ziel  und Hauptprobleme  der regionalen
  Limnologie. Bot. Notiser. 1927:81-103.
Nicholls, K.H., and P.J. Dillon. 1978. An evaluation of phos-
  phorus-chlorophyll-phytoplankton relationships for  lakes.
  Int. Revue ges. Hydrobiol. 63:141-54.
Rawson, D.S. 1939. Some physical and chemical factors in
  the metabolism of lakes. Pages 9-26 in Problems in Lake
  Biology.  Publ. No.  10.  Arn. Ass. Advancement Sci.
Shapiro, J.  1980. The  importance of trophic-level interactions
  to the abundance and species composition of algae in
  lakes. Pages 105116 in J. Barica and LR. Mur, eds. Hyper-
  trophic Ecosystems. W. Junk, The Hague.

Smith, V.H. 1979. Nutrient dependence of primary productiv-
  ity in lakes. Limnol. Oceanogr. 24:1051-64.
Walker,  W.W.  1979. Use of hypolimnetic  oxygen depletion
  rate as a trophic state index for lakes. Water Resour. Res
  15:1463-70.

Weich, E.B., and M.A. Perkins. 1979. Oxygen deficit-phos-
  phorus loading relation in lakes. J. Water  Pollut. Control
  Fed. 51:2823-8.
                                                     430

-------
  WHO NEEDS TROPHIC STATE INDICES?
 RICHARD OSGOOD
 Metropolitan Council-Twin Cities
 Minneapolis-St. Paul, Minnesota
             ABSTRACT

             The degree of nutrient enrichment has classically indicated trophic state. Nutrients alone, how-
             ever, inadequately describe a lake's trophic-dynamic structure. Therefore, trophic state indices
             are simplifications of a more dynamic process. The specific utility of trophic state indices for
             their intended uses (classification and ranking by trophic status) is limited by deviant behavior
             among the indices. Secondary index information may be obtained by examining these devia-
             tions. Often, these deviations indicate the nature of nontypical trophic behavior.
 INTRODUCTION

 The oligotrophic-eutrophic nomenclature classically
 refers to a lake's state of enrichment with nutrients:
 oligotrophic being nutrient-poor and eutrophic being
 nutrient-rich. The level of nutrients alone, however, in-
 adequately characterizes a lake's trophic status. The
 qualitative and quantitative aspects of energy trans-
 formations in aquatic ecosystems  describe a lake's
 trophic-dynamic structure (Lindeman, 1941, 1942). The
 size,  structure, and  distribution of aquatic  com-
 munities  are   manifestations  of their  interrelated
 energy dynamics. Classically, autotrophic planktonic
 productivity has characterized lake  trophic-dynamics
 and is related  to nutrients (Smith, 1979). Trophic state
 indices are by necessity, then, simplifications of a
 more dynamic process. These simplifications include
 numerous attempts to produce  indices of the trophic
 status of lakes and reservoirs. The utility of these in-
 dices for their intended uses (i.e., classification and
 ranking)  is limited by their general  nature,  but this
 limitation may indicate deviant trophic behavior and
 be useful for lake management.
 UTILITY OF TROPHIC STATE INDICES

 Algal community characteristics (primary productivity,
 species composition, abundance) generally reflect a
 lake's trophic dynamics (assuming nearshore trophic
 processes are  insignificant). Phosphorus,  nitrogen,
 Secchi  disk,  chlorophyll,  oxygen  depletion,  and
 nutrient loading are all related (both functionally and
 statistically) to the lake's algal community dynamics.
 These relationships have permitted such parameters
 to be used alone and in various combinations to  in-
 dicate  trophic  state  (Carlson,  1977; Walker,  1979;
 Kratzer  and  Brezonik, 1981). Trophic state indices
 generally use a numeric nomenclature that represents
 trophic status as a continuum (Carlson, this vol.). This
 facilitates classifying and ranking lakes by trophic
 state and is useful for communication.
  Whether to  use  a single parameter  index  (e.g.,
 chlorophyll,  phosphorus or  Secchi disk) or a multi-
 variate index (Brezonik, this vol.) needs to be  con-
 sidered.  If  a  single  lake is   being   considered
 throughout all  phases of a  restoration  effort  or  if
 similar lakes in a small region are being classified,
 perhaps a single parameter index is appropriate. More
often, where there is more than one perceived problem
or the causes as well  as the  consequences  of
eutrophication  need  to  be  classified,  a multiple
parameter  index  may be  used. Choosing single or
multiple parameter indices requires the resolution of
both conceptual and practical limitations. These are
best exemplified by examining  the  evolution  of  a
trophic state index (TSI) introduced by Carlson (1977).
  The TSI of Carlson (1977) presents the trophic condi-
tion on a zero to 100 scale with a change of 10 units in-
dicating  significant differences  in trophic ecology.
The index was derived from Secchi disk (SD) trans-
parency values spanning 64 to 0.06 m  (TSI: 0-100) with
SD  =  4-2  m  being the  approximate oligotrophic-
eutrophic transition (TSI  = 40-50). Chlorophyll a (CHL)
and  total   phosphorus  (TP)  concentrations   were
statistically matched to TSI (SD) to yield the additional
index measures: TSI(CHL) and TSI(TP), respectively.
Carlson prefers that TSI (CHL) is the index choice, but
the other values can be  appropriate substitutes. The
three index values may not be averaged  for concep-
tual and mathematical reasons. In the past, however,
they have been averaged, often inappropriately.
  The  presentation of three indices  that apparently
equivalently (statistically) represent trophic state, in-
vites averaging and the addition of other  statistically
related parameters (Porcella et al. 1979;  Kratzer and
Brezonik, 1981; Osgood, 1982a). The nearly simultan-
eous presentation of Kratzer and Brezonik's nitrogen
index (1981) and  Osgood's  more general scheme
(1982a), both averaging the TSI's,  and  the literature
crossfire that followed (Osgood, 1982b, 1983a; Kratzer
and  Brezonik,  1982;  Lambou,  1982;  Brezonik  and
Kratzer, 1982;  Carlson, 1983) served to  elucidate the
utility of Carlson's approach. Regional  differences
among the index values are dealt with in two ways: by
creating a new index (Kratzer and  Brezonik, 1981) or
more generally by using the differences to explain the
trophic  behavior of dysfunctional lakes (Osgood,
1982a). Both approaches average  the index values,
which  is appropriate within  the assumptions  and
limitations presented.
  Carlson's TSI then, has the advantage of presenting
trophic state on a continuous numeric scale (Carlson,
1977; Reckhow, 1979) and can approximate the oligo-
trophic-mesotrophic-eutrophic nomenclature (Carl-
son, 1979). Although the index can be used to classify
and rank lakes  according to  trophic state, priority
ranking may be difficult. Square Lake, for  example, is
the clearest lake in the Twin Cities metropolitan area,
                                                 431

-------
 LAKE AND RESERVOIR MANAGEMENT
 Minnesota  (Osgood, 1981, 1982a),  but  phosphorus
 management to protect this  lake  is inappropriate
 (Osgood, 1983b, this  vol.). The trophic  state  inde*
 values suggest that difficulty (Osgood, 1982a; Fig. 1):
 TSI (TP) is greater than TSI (CHL) or TSI (SD). First, the
 lake's  phosphorus  is  primarily  from  groundwater
 sources (67 percent of the annual input) and cannot be
 controlled. Also, algal  abundance  (chlorophyll) in the
 lake  is not related to phosphorus; Daphnia grazing
 controls  chlorophyll. Fisheries management is  the
 primary  management  strategy   for  Square  Lake
 (Osgood,  1983b). Trophic state classification may cor-
 rectly rank Square Lake in the metropolitan area, but
 would not correctly indicate the appropriate manage-
 ment approach.
   Trophic state indices  may  also be used to  com-
 municate  lake  quality  to the  public.  The index
 presented must  adequately reflects the  perceived
 problems and can be easily related to the intended
 management approach. TSIfTP), for  example, would
 not communicate either in the case of Square Lake.


 TROPHIC STATE INDICES FOR LAKE
 MANAGEMENT

 Trophic state indices simplify environmental measurei-
 ments, seemingly  rendering  them  inadequate  for
 detailed ecological  analysis.   However,  differences
 among index values, such as Carlson's TSIs, can in-
 dicate dysfunctional or nonnormal trophic behavior
 This secondary index information may help to better
 understand and manage lakes.  Lake management in-
 volves correcting or protecting lakes from degrada-
 tion,  not  necessarily eutrophication. This  requires
 defining problems (existing or  potential), identifying
 causes, examining feasible management alternatives,
 and implementing remedial measures to achieve the
 desired condition. The relationship  of trophic state in-
 dices to one another or to other environmental para-
 meters may be helpful  for some of these aspects of
 lake management. The following illustrate nonnormal
 trophic behavior:
   1. Lakes/reservoirs with nonalgal turbidity.  Non-
 algal turbidity may be indicated with trophic state in-
 dices (Osgood, 1982a), but this condition is not neces-
 sarily related to the eutrophication process. Classifi-
 cation and  ranking  though, may  still be possible
 (Carlson, 1980; Walker, this vol.). A nutrient manage-
 ment strategy would not likely be the primary manage-
 ment approach.
  2. Nitrogen limitation of algal abundance. Trophic;
 state  indices may adequately  indicate the trophic
 character  of a  nitrogen-limited lake  (Kratzer  and
 Brezonik,  1981; Osgood,   1982a,b). This should  not
 replace a thorough  assessment of  nutrient hydro
dynamics  or a definitive limiting nutrient  determina-
tion.
  3. Aquatic  macrophytes are   the  predominant
 primary producers.  Abundant  growths  of  rooted
 vascular plants may represent  the most  significant
 autotrophic production in the lake. Indices that rely on
 limnetic conditions, then, poorly reflect trophic state.
 Lake George, Minn.,  exemplifies this situation. Lake
 George is  a fairly large  (197 ha), shallow (mean depth
 = 2.7 m) lake. The limnetic water quality is quite good
 (Table 1), but aquatic macrophytes grow from 1 to 5 m
 depth (Fig. 2; Osgood, 1983c). In this case, the trophic
 state  indices do not reflect an   important trophic
 aspect of  Lake  George, although another index  may
 be  satisfactory (Canfield  and  Jones,  1984). Thesei
  macrophytes  are  not  a particular problem in Lake
  George since they grow away from shore and below
  the lake's surface. The proposed management deals
  solely with the limnetic aspects of the lake (Osqood
  1983c).
    4.  Lakes  with  Aphanizomenon   flake  blooms.
  Aphanizomenon flos-aquae filaments aggregate into
  flakes  with  grass-blade morphology  when  large-
  bodied size Daphnia are present and there is an oxic
  sediment-water interface (Lynch, 1980; Lennon, 1981).
  Carlson's trophic state index values in this case are
  typically related as follows: TSI (TP) >TSI (CHL) due to
  nitrogen limitation, the "cost of coloniality" (Smith et
  al. 1982) or grazing Daphnia; and TSI (CHL) > TSI (SD)

    Table 1.—Trophic state index values of Lake George.1

  Year          TSIfTP)       TSI(CHL)        TSI(SD)
1980
1981
1982
49
48
47
50
50
50
44
42
42
  'Carlson's (1977) TSI using summertime average values of the respective
  parameters taken from the center of the lake at the surface Data from Osgood
  (1981, 1982C, 1983C)
  3
  *l

  I •-
  0 4-
                                               -30

                                               •20
       • n F.b I Mar I Apr ! May I June'July I Aug S.ptl Oet I Nov I Dec
                        1862

Figure 1.—Total  phosphorus, chlorophyll and Secchi
transparency, Square Lake. TSI from Carlson (1977)1.
disk
                                                 432

-------
                                                                                              TROPHIC STATUS
because the large flakes do not attenuate light as do
smaller particles (Edmondson, 1980). Aphanizomenon
flakes are nuisances related to the lake's nutrients as
well as the lake's  biology. Trophic state indices may
very  well  reflect  the  presence of  flakes  (Osgood,
1982a), but  inadequately describe the lake's ecology.
   5.  Daphnia controls algal abundance. An earlier ex-
ample illustrates  the  effect  of  efficient grazing by
Daphnia on algal  abundance (Fig. 1). Although algal
abundance is not primarily controlled by nutrients,
nutrients need to be considered (Osgood, this vol.).
   Abnormal trophic behavior may  be indicated by
deviations  among trophic state  indices. In  this way,
trophic state indices may be  useful for lake manage-
ment. The operating trophic mechanisms require veri-
fication  following the initial  direction given  by the
trophic state indices.
WHO  NEEDS TROPHIC STATE INDICES?

Trophic state  indices, as  well as  the oligo-meso-
eutrophic nomenclature, are useful to lake managers
for  classifying,  ranking,  and  communicating the
general trophic nature of lakes. Understanding detail-
ed trophic-dynamic  aspects  of lakes using  trophic
state indices alone  is not possible,  although the in-
dices may  indicate  the nature of  nontypical  trophic
behavior. Trophic state indices are useful,  indeed,
have become important for some functions of the lake
manager,  and  can  also  indicate a  lake's  trophic
character.
ACKNOWLEDGEMENTS:  Funding  for projects  involving
Lake George and Square Lake was provided at various times
by the U.S. Environmental Protection Agency, the State of
Minnesota, and the Metropolitan Council, with cooperative
funding  from  the  U.S.  Geological Survey   and  the
Metropolitan Waste Control  Commission.  I am indebted to
Bob Carlson for his trophic state index, for his discussions
(formal and informal) over the last few years, and for his com-
ments on this paper.
Figure  2.—Distribution  of  aquatic  macrophytes in  Lake
George, 1982.
REFERENCES

Brezonik, P.L. This volume. Trophic state indices: Rationale
  for  multivariate  approaches.  In  Lake and  Reservoir
  Management. Proc. Int. Symp. N. Am. Lake Manage. Soc.
  Knoxville, Tenn. Oct. 18-20, 1983.
Brezonik, P.L., and C.R. Kratzer. 1982. Reply to discussion by
  V.W.  Lambou: A  Carlson-type  trophic state  index  for
  nitrogen in Florida lakes. Water Res. Bull. 18:1059-60.
Canfield, D.E., Jr., and J.R. Jones. This volume. Trophic state
  classification of lakes with aquatic macrophytes. In Lake
  and Reservoir Mangement. Proc. Int. Sym. N.  Am. Lake
  Manage. Soc. Knoxville, Tenn. Oct.  18-20, 1983.
Carlson, R.E. 1977. A trophic state  index for lakes. Limnol.
  Oceanogr. 22:361-9.
	1979. A review of the philosophy and construction
  of trophic state indices. In T.E.  Maloney, ed.  Lake and
  Reservoir Classification Systems. EPA-600/3-79-074. U.S.
  Environ. Prot. Agency, Washington, D.C.
	1980. Using trophic state indices to examine the
  dynamics of eutrophication. In Lake Restoration and Pro-
  tection. Proc. Int. Symp. Inland Waters and Lake Restora-
  tion, Portland, Maine. Sept. 8-12. EPA-440/5-81-010. U.S. En-
  viron.  Prot. Agency, Washington,  D.C.

	1983.  Discussion:  "Using differences  among
  Carlson's  trophic state index values in regional water
  quality assessments," by R.A. Osgood. Water Res. Bull.
  19:307-8.
	This volume. Conceptual overview of trophic state.
  In Lake and Reservoir Management. Proc. Int. Symp.  N.
  Am Lake Manage. Soc. Knoxville, Tenn. Oct. 18-20, 1983.
Edmondson, W.T. 1980. Secchi disk and chlorophyll.  Limnol.
  Oceanogr. 25:378-9.

Kratzer, C.R., and P.L. Brezonik. 1981. A Carlson-type trophic
  state  index for nitrogen in Florida Lakes. Water Res. Bull.
  17:713-15.
	1982. Reply to discussion by R.A.  Osgood: "A
  Carlson-type  trophic state index for nitrogen  in  Florida
  lakes." Water Res. Bull. 18:543-4.
Lambou, V.W.  1982. Discussion:  "A Carlson-type  trophic
  state  index for nitrogen in Florida lakes," by C.R.  Kratzer
  and P.L Brezonik. Water Res. Bull. 18:1057-8.
Lennon,  H.J.  1981.  The  natural  history of a  bloom  of
  Aphanizomenon flos-aquae. M.S. Thesis. Univ. Minnesota.
Lindeman, R.L. 1941.  Seasonal food-cycle dynamics in a
  senescent lake. Am. Midi. Nat. 26:636-73.
	1942. The  trophic-dynamic aspect  of  ecology.
  Ecology 23:399-418.
Lynch, M  1980. Aphanizomenon blooms: Alternate  control
  and  cultivation  by Daphnia pulex.  Am.  Soc.   Limnol.
  Oceanogr. Spec. Symp. 3:299-304.
Osgood, R.A.  1981. A study of the water quality of 60 lakes
  in the seven county metropolitan area. Metro. Counc. Publ.
  No. 01-81-047.

	1982a. Using differences among Carlson's  trophic
  state  index values in regional water quality assessment.
  Water Res. Bull. 18:67-74.

	1982b. Discussion: "A Carlson-type trophic state
  index  for nitrogen in Florida lakes," by C.R. Kratzer and
  P.L. Brezonik. Water Res. Bull. 18:343.

	1982c. A 1981 study of the water quality of 30 lakes
  in the seven county Metropolitan Area. Metro.  Counc.
  Publ.  No. 10-82-005.
	1983a. Reply to discussion by R.E. Carlson:  "Using
  differences among Carlson's trophic state index values in
  regional water  quality  assessment." Water Res. Bull.
  19:309.

	1983b. Diagnostic-feasibility study of seven metro-
  politan area lakes:  Square Lake. Metro. Counc. Publ. No.
  10-83-093G.
                                                    433

-------
LAKE AND RESERVOIR MANAGEMENT
	1983c. Diagnostic-feasibility study of seven metro-
  politan area lakes- Lake George. Metro. Counc  Publ  No
  10-83-093D.

	. This volume. Long-term grazing control of algal
  abundance: A case history. In Lake and Reservoir Manage-
  ment. Proc. Int. Symp.  N. Am.  Lake Manage. Soc.  Knox-
  ville, Tenn  Oct 18-20, 1983.

Porcella, D B., S.A.  Peterson, and D.P. Larsen.  1979. Pro-
  posed method for evaluating the effects of restoring lakes.
  In Limnological and Socioeconomic  Evaluation of Lake
  Restoration  Projects.  Approaches  and  Preliminary
  Results.  EPA-600/3-79-005. U.S. Environ.  Prot.  Agency,
  Washington, D.C.

Reckhow, K.H. 1979. Quantitative Techniques for the Assess-
  ment of Lake Quality. EPA-440/5-79-015. U.S. Environ. Prot.
  Agency, Washington, D.C.
Smith, V.H.  1979. Nutrient dependence of primary produc-
  tivity in lakes. Limnol. Oceanogr. 24:1051-64.
Smith, V.H., J. Shapiro, and N.P. Holm. 1982.  Ecological
  costs of coloniality in Aphanizomenon flos-aquae. 45th
  Annu.  Meet.  Am. Soc. Limnol. Oceanogr., Raleigh N C
  June 14-17.

Walker, W.W. Jr. 1979. Use of hypolimnetic oxygen depletion
  rate as a trophic state index for lakes. Water Resour Res
  15:1463-70.

	This volume. Trophic state indices in reservoirs. In
  Lake and Reservoir Management. Proc. Int. Symp. N Am
  Lake Manage. Soc. Knoxville, Tenn. Oct. 18-20, 1983.
                                                     434

-------
TROPHIC STATE INDICES  IN  RESERVOIRS
WILLIAM W. WALKER,  JR.
Environmental Engineer
Concord, Massachusetts
             ABSTRACT

             Trophic state index systems provide a framework for data summary, interpretation, and communica-
             tion. Carlson's indices can be viewed as reexpressions of bivariate regression analyses derived from
             phosphorus-limited, northern, natural lakes Several studies have shown that lakes or reservoirs which
             are nitrogen limited and/or have relatively high concentrations of non-algal turbidity tend to deviate
             in various ways from these regressions. These deviations are "problems" only if misinterpreted but
             limit the use of the index system for comparisons or rankings within certain regions and/or types of
             impoundments. Analysis of data from 65 Corps of Engineer impoundments indicates that a classifica-
             tion or index system which incorporates effects of nitrogen and non-algal turbidity would be of more
             general use in reservoirs. A principal components analysis is used to summarize impoundment response
             data into two composite variables which explain 93 percent of the variance in the original measurements
             The first component is interpreted as a quantitative dimension which reflects the total amounts of
             nutrients and light extinction in the water column. The second is a qualitative dimension which reflects
             the partitioning of nutrients and light extinction between organic and inorganic forms.  Basically, this
             system takes advantage of the fact that some of the deviations from a Carlson-type index system are
             systematic and contain information on the partitioning of nutrients and light extinction If the objective
             is a concise summary of water quality data, information on both dimensions provides a more com-
             plete description of reservoir water quality than any single composite variable or index.
INTRODUCTION

Trophic state index systems provide a framework for
summarizing, interpreting, and communicating water
quality information. While the concept of a continuous
index is more realistic than that of a discrete classifi-
cation system, its use involves simplification and can-
not serve as a complete substitute for analysis and in-
terpretation of individual measurements. Of the prac-
tical and common water quality measurements chlor-
ophyll a is the most direct indicator of algal standing
crop. As discussed by Carlson (1983), given that there
are estimates of mean chlorophyll a derived from ade-
quate sampling programs, water bodies can be ranked
and classified based on these  measurements alone
and there would be no need for more elaborate "index
systems." This approach  is limited, however, by (1)
weaknesses in chlorophyll a as an indicator which can
be  attributed to variations in chlorophyll a/biomass
ratios, and (2) practical difficulties in obtaining statis-
tically reliable estimates of mean (or especially, max-
imum) chlorophyll a  concentrations  from  limited
monitoring data because of the relatively high tem-
poral or spatial variability  that  may occur within a
given water body and growing  season. These  pro-
blems suggest that interpretations should  also be
based upon other types of measurements related to
trophic  state,  including  transparency  and  nutrient
concentrations. Complications arise when more than
one variable is introduced because an  underlying
model must be assumed.
  Carlson's (1977) index system  and its  descendants
(Walker, 1979;  Kratzner and Brezonik, 1981; Osgood,
1982) can be  viewed as simple transformations of
bivariate regression  equations derived from popula-
tions  of  lakes.  While  the  index   system  is
"multivariate"  in the sense that more than one type of
measurement is considered, it is one-dimensional in
the sense that  it assumes  a unique relationship be-
tween each  measurement  and a common scale. The
underlying model is that phosphorus exclusively con-
trols chlorophyll and transparency. When applied to
data from a given lake, deviations among the various
versions of the index will arise from combinations of
(1) random sampling and analytical errors and (2) ef-
fects of deterministic  factors which are not  con-
sidered in the underlying  model. As discussed by
Walker (1979), and Kratzner and  Brezonik (1981), and
Osgood (1982), averaging the index  versions helps to
reduce the effects of random data errors and incor-
porates information from each type of measurement.
If model errors (attributed to nitrogen limitation or tur-
bidity,  for example)  are responsible for large devia-
tions among the index versions, averaging can be mis-
leading and  results will be in error if phosphorus
and/or transparency  indices are used in the absence
of chlorophyll a measurements. Osgood (1982) and
Carlson  (1983) present  frameworks for  interpreting
systematic deviations among the indices in relation to
various deterministic factors; this is one of the most
useful  applications for the index system because it
places lake conditions into perspective and when con-
sidered along with other lake characteristics can pro-
vide insights into controlling factors other than phos-
phorus.
  Significant deviations among Carlson-type  indices
have been shown when they are applied to data from
some reservoirs (Carlson, 1980; Walker, 1980). Some of
these deviations reflect systematic influences of  non-
algal turbidity on chlorophyll production and trans-
parency at a  given phosphorus level. Nitrogen limita-
tion also causes systematic deviations when  applied
to lake or reservoir data (Carlson, 1980; Kratzner and
Brezonik, 1981).
  This paper  describes a  classification  system for
reservoirs which explicitly considers effects of turbidi-
ty and  nitrogen, in addition to those considered  in a
Carlson-type  system. This involves a basic change in
model structure which goes beyond simple  recalibra-
tion of the existing framework. The analysis is based
                                                 435

-------
 LAKE AND RESERVOIR MANAGEMENT
 upon surface-layer, growing-season means of chloro-
 phyll a,  transparency, organic  nitrogen, and  com-
 posite nutrient concentration derived from 65 Corps of
 Engineer impoundments (Walker, 1983).
   Composite  nutrient  concentration  is  computed
from total phosphorus and total  nitrogen concentra-
tions and has  been designed as a measure of nutrient
supply that is independent of whether phosphorus or
nitrogen is limiting (Walker, 1983):

     Xpn  = [P-2 + ((N-150)/12)-2]-s

where,
     Xpn  = composite nutrient concentration (mg/m3)
     P =  total phosphorus (mg/m3)
     N =  total nitrogen (mg/m3)

At high N/P ratios, the expression is independent of
nitrogen and  approaches the total phosphorus  con-
centration.  At low N/P ratios,  it  is independent of
phosphorus and approaches (N-150)/12. The para-
meters used in computing the composite nutrient con-
centration are based upon partitioning models which
relate organic nitrogen and particulate  phosphorus
concentrations to chlorophyll a and nonalgal turbidity
(Walker,  1983).  The  nitrogen  intercept (150 mg/m3)
reflects  an average organic  nitrogen  component
which is uncorrelated with chlorophyll a or turbidity.
The intercept is needed to stabilize the ratio of organic:
N  to particulate phosphorus (12) over the range o:
observed  chlorophyll a  values.  Use of  composite
nutrient concentration as  a trophic state index is an
alternative to  the scheme proposed by Kratzner and
Brezonik (1981) involving the minimum of Carlson-type
phosphorus and nitrogen indices.
  Relationships among the previous measurements
are shown  in  Figure 1,  using  different  symbols to
distinguish  reservoirs with chlorophyll transparency
products  above and below 10  mg/m2. This value
divides the data set  roughly in  half. As discussed in
more detail later, this product is proportional to light-
limited productivity and to the fraction of light extinc
tion  attributed to chlorophyll and chlorophyll-related
substances. The symbol  distributions suggest  thai
each of the bivariate relationships  is  influenced to
some extent by the chlorophyll-transparency product.
The  chlorophyll  versus nutrient  plot indicates  thai
reservoirs in which chlorophyll  accounts for a majoi
portion of light extinction also show a higher response!
to nutrient concentrations. Some of  this  influence is,
spurious  because the chlorophyll value determines.
the vertical scale and partially determines the symbol
The organic nitrogen versus composite nutrient plot is.
free  of spurious correlation, however,  and shows z
similar  symbol  distribution,  particularly  at  higr
nutrient concentrations.
  One  measure  of  the  performance  of  an  inde>
system is  the extent to which it explains variance anc
covariance among the original measurements. This, ir
turn, reflects the generality of the underlying  mode
and the amount  of "lost" information when the inde>
summarizes water quality conditions (Reckhow, 1981)
Table 1  summarizes the performance of various one
dimensional and two-dimensional index systems  ap
plied to CE reservoir data.
  The one-dimensional systems are similar in concepl
to Carlson's (1977) and explain between 62.9 percenl
and  77.3  percent  of  the  total  variance  in all  foui
measurements. Principal components analyses have'
been found   useful  in  previous  developments  ol
 regional trophic state indices tor lakes (Shannon and
 Brezonik, 1972) (Boland, 1976) and  other types of
 classification problems  (Harris, 1975). The first prin-
 cipal component (PC-1)  defined in Table 1 captures
 82.2 percent of the source  variance. While PC-1 is
 analogous to an "average Carlson index", two reser-
 voirs can have similar PC-1 values but very different
 chlorophyll  a concentrations,  as  described  below.
 thus, it is risky to define PC-1  as  a "trophic state in-
 dex."
  Two-dimensional index systems explain significant-
 ly higher percentages  of  the source  variance. A
 system based upon the first two principal  components
 explains 95.5 percent (Fig. 2). The second component
 accounts for 13.3 percent (or  75  percent  of that re-
 maining after consideration of PC-1) and  is controlled
 largely by variations in the product of chlorophyll and
 transparency, since the signs and  magnitudes of
 these terms  are nearly identical. Because of the cor-
 relations  discussed later, the  classification system
 can  be  simplified by treating the composite nutrient
 concentration  as   the  first  dimension  and  the
 chlorophyll-transparency product as the  second. The
 revised  system  (Fig. 3) captures 91.6 percent of the
 variance in the individual measurements.
  Correlations between the principal components and
 impoundment characteristics are  listed   in Table 2,
 along with a series of multiple regression equations
 which help to provide physical  interpretations. While
 PC-1 is strongly correlated with each of the individual
 measurements,  PC-2 is strongly correlated with com-
 posite variables, such as  the chlorophyll-transparency
 product (r = .99) and the ratio of chlorophyll  a to
 limiting nutrient concentration (r = .87).
  The first principal component can be interpreted as
 a quantitative factor which  reflects total concentra-
tions, and, particularly, the total nutrient supply (Xpn).
The  second  component can  be  interpreted  as a
qualitative factor which  reflects the partitioning of
 light extinction and nutrients between algal and non-
algal components.  Based upon kinetic  theories of
algal growth, the chlorophyll-Secchi  product is also
proportional to  the areal  primary production rate in a
mixed, totally-absorbing  surface  layer  under  light-
limited conditions (Oskarn, 1973). The  revised classifi-
cation  system  permits  isolation  of  the composite
nutrient concentration along one  dimension,  which

     Table 1.—Peformance of various index systems.

 	Percent of Variance Explained	
 Index         Chl-a   Xpn   Org-N   Secchi    Total
               One-Dimensional  Indices
 Chl-a         100.0    59.8    71.5    31.4     67.0
 Xpn            59.8   100.0    77.1     72.8     77.3
 Org-N          71.5    77.1   100.0    45.0     70.0
 Secchi         31.4    72.8    45.0    100.0     62.9
 PC-1            783    93.7    84.5    72.6     82.2
               Two-Dimensional  Indices
 PC-1  & PC-2     97.9    95.5    87.2    96.5     95.5
 Xpn  & B'S      90.1   100.0    84.1     87.5     91.6
 all statistics on log scales
 PC-1,  PC-2 = principal components of covariance matrix
 PC-1 = 554 log(B) +  359 log(Norg) + 583 log (Xpn) - log(S)
 PC-2 = 689 logfB) +  162 log(Norg) - 205 log(Xpn) + 676 log(S)
 B = chlorophyll a (mg/m3)
 Norg  = organic nitrogen (mg/m3)
 Xpn - composite nutnen' concentration (mg/m3)
 S = Secchi depth (m)
                                                  436

-------
                                                                                          TROPHIC STATUS
 can be  interpreted as  a  causal  factor rather than a
 system  response (Carlson, 1983). Information on both
 dimensions provides a more complete picture of varia-
 tions  in  chlorophyll,  transparency,  and  organic
 nitrogen concentrations than can be derived from any
 of the one-dimensional systems.
   The vectors shown in Figures  2 and 3 depict direc-
 tions of increasing response  measurements,  based
 upon the multiple  regressions presented in  Table 2.
 These regressions should not be applied outside of
 the ranges of data shown in the figures.  The two-
 dimensional aspect of the classification system  is
 reflected by the divergence  of the vectors.  Projects
with the highest chlorophyll a concentrations tend to
be located in the upper right corner of the plots, where
nutrient supply and light-limited productivity are both
relatively high. Of the other three measurements, the
organic nitrogen vector is most similar to the chloro-
phyll  a vector. This  reflects the fact  that organic
nitrogen  concentrations  are only weakly  correlated
with nonalgal turbidity levels.
  Figure 4  verifies  the  chlorophyll distribution  by
using different symbols to depict variations in chloro-
phyll concentration. Observed chlorophyll a contours
are shown in relation to those predicted by the multi-
ple regression equation  in Table 2. The plot shows
 LOG  [ CHLOROPHYLL-A, MG/M3  ]
LOG  [  TRANSPARENCY, M  J
2.1-
1.8-


1.2H
0.9-
0.6-
0.3-
O.Oi
0.
o
00 °
o 0°°o
°o °°+
\;*+* { s
O ° O+ j.
°° ° ++
CD -tq.
+ +•*•*•*

6 0.9 1.2 1.5 1.8 2.1 2.4
0.8
0.6
0.4
0.2
0.0
-0.2-
-0.4-
-0.6-
-0.8n
0
\^-B*S » 10
\ 8
V O J- "fe^.
ODD^ ^L ^r
+ -I-0 0


LOG [
3.4-
3.2-

3.0-
2.8-
2.6-
2.4-
2.2-
2.0^
ORGANIC N, MG/M3 1

o
o
o
+ ° °
t 0
A T^ Ju9 O O
+ o °^^0
**fi*0 + +
+ +* °+°


    0.6   0.9    1.2   1.5    1.8    2.1   2.4

     LOG [ COMPOSITE NUTRIENT  CONC,  MG/M3  ]
   0.0  0.3   0.6  0.9  1.2  1.5 1.8  2.1

          LOG [ CHLOROBHYLL-A,  MG/M3 ]
Figure 1.—Relationships among reservoir trophic state indicatorsl.
1 (o) B"S > 10 mg/m2, "chlorophyll-dominated", (+) B'S f 10 mg/m", "turbidity-dominated"; B = chlorophyll a (mg/m1) S = Secchi (m)
                                                  437

-------
 LAKE AND RESERVOIR MANAGEMENT


£
i
§
u
|
M
Bu
g
§




1.4-

1.2
1.0

0.8
0.6

0.4

0.2


0.0
1
Secchi ;* Chl-a
A / Org- N
\ / o o -,-*1
•-• _ o \ ,0 0^0°
~- °0 ^0 / -^
° ° """^-.S 62''' it." ° "
° °^Q-^ ° ^
°° flD-P- OOV "J-. —
^-^ / \0°8^^Xpt
^ ^ ° ^
' 0°\
/ " \ 0
y \
' \
-

                   FIRST PRINCIPAL COMPOHIST
Figure 2.—CE  reservoirs distributed on PC-2 versus PC-1
Axes'.
' Arrows show directions of increasing chlorphyll a, transparency, organic
nitrogen, and composite nutrient concentrations, based upon multiple regres-
sion equations in Table 2

that, at a given nutrient level, chlorophyll a concentra-
tions can vary systematically by as much as eightfold
(.9 log units), depending upon the second component.
Since chlorophyll concentrations must be known  in
order to compute the second dimension, Figure 4 and
the regression equation are useful only for data inter-
pretation. A theoretically-based  model  for predicting
chlorophyll  as a function  of  phosphorus,  nitrogen,
nonalgal turbidity, depth, and predictive flushing rate
has  been developed  for use  in a  predictive mode
(Walker, 1983).
  Figure 5 compares the distributions of 65 CE reser-
voirs, 15 TVA  reservoirs (Higgins and Kim, 1981), and
73 natural lakes  sampled by the EPA National Eutro-
phication Survey (U.S. Environ.  Prot. Agency, 1978).
Because total nitrogen concentrations are required for
computation  of  composite  nutrient concentrations,
but were not  measured by the EPA/NES  in north-
central and northeastern States, the lakes  data are
primarily from middle and  southern latitudes of the
United States. While  a wide range  of nutrient con-
centrations  is apparent for each group, the distribu-
tion of chlorophyll-transparency products tends to be
higher for the lakes. A clear distinction between TVA
                   Table 2.—Impoundment characteristics versus two-dimensional index systems.
 Product-Moment Correlation Coefficients:
                                               Index System
Variable
Chlorophyll a
Organic nitrogen
Secchi depth
Composite nutrient
Nonalgal turbidity
Chl-a * Secchi
Chl-a/Xpn
Mean
Standard deviation
I
PC-1
.885
.919
-.852
.968
.610
.144
-.070
2.27
.58
PC-2
.443
.167
.490
-.137
-.756
.986
.870
.77
.24
II
Xpn
.774
.878
-.853
1.000
.670
.017
-.285
.89
.37
B*S
.564
.280
.368
.017
-.662
1.000
.827
1.47
.35
Mean
.89
2.63
.05
1.47
-.22
.94
-.58


Std.
Dev.
.37
.23
.32
.35
.38
.33
.24


Multiple Regression Equations:
                                                      Coefficients
System I
Chl-a
Org-N
Xpn
Secchi
Turbidity
Chl-a * Secchi
Chl-a / Xpn

System II
Chl-a
Org-N
Xpn
Secchi
Turbidity
Chl-a * Secchi
Chl-a / Xpn
Intercept
-.899
1.691
.304
.605
-.176
-.295
-1.203

Intercept
-.858
1.623
.000
.858
-.556
.000
-.859
PC-1
554
359
.583
-.474
.393
" .080
-.028

Xpn
.794
.567
1,000
-.794
.729
.000
-.206
PC-2
.689
.162
-.205
.676
-1.208
1.365
.894
Coefficients
B*S
.617
.185
.000
.382
-.777
1.000
.618
R2
.979
.872
.955
.965
.944
.993
.761

R2
.901
.841
1.000
.875
.903
1.000
.774
SE2
.0028
.0069
.0057
.0038
.0081
.0008
.0144

SEZ
.0136
.0085
.0000
.0136
.0141
.0000
.0136
all statistics computed on log scales
                                                   438

-------
                                                                                             TROPHIC STATUS
tributary and mainstem  impoundments is also ap-
parent. Turbidity and flushing rate are more important
as controlling factors  in the latter.
  Analysis of additional data from Vermont (Walker,
1982) and  Minnesota  (Osgood,  1982) indicates that
chlorophyll-transparency  products  in northern lakes
tend to exceed 10 mg/m2 and most are outside of the
range in which light-limitation effects are likely to be
important. At high  values, variations  in chlorophyll-
transparency products are probably related more to
effects of different  algal  species (Osgood, 1982) and
mixing regimes (metalimnetic algal populations)  on
the chlorophyll-Secchi relationship than to variations
in nonalgal  turbidity.  A  strong  correlation between
nutrient  partitioning (chlorophyll/nutrient) and light
partitioning (chlorophyll  x  transparency) would not
be expected  in this range.
  In  summary,   a  two-dimensional   classification
system that  describes nutrient supply  and potential
light-limited  productivity  provides  a  more  complete
summary of  reservoir water quality than  is possible
with a one-dimensional  system.   A  principal  com-
ponents  analysis  yields quantitative and qualitative
dimensions  which  account for  95.5  percent  of the
variance in four measurements. The  second compo-
nent (chlorophyll-transparency product) explains  75
percent of the variance remaining after consideration
- 1.6
2
gl.»
_- 1.2

M lt0
01
* 0.8
1 ••'
— o.*
J 0.2
0.0-
«5ecchi /Chl-a ^
/o -tf" Org-N
x-""
a ° °S>a°o''o>>g
8 o o ° o<>o 0
o ""

-------
LAKE AND RESERVOIR MANAGEMENT

REFERENCES

Boland, D.H.P.  1976. Trophic classification of lakes using
  LANDSAT-1  (ERTS-1)  multispectral  scanner  data.
  EPA-600/3-76-037. U.S.  Environ. Prot.  Agency, Corvallis,
  Ore.

Carlson,  R.E. 1977. A trophic state index for lakes. Limnol.
  Oceanogr. 22 (2):361-9.
	1980. Using  trophic  state  indices  to  examine
  the dynamics of eutrophication. In Restoration of Lakes
  and  Inland Waters. Proc.  Int.  Symp. Inland Waters and
  Lake Restoration. Portland, Maine. EPA-440/5-81-010. U 3
  Environ. Prot. Agency, Washington, D.C.
	1983. Discussion  of "Using  differences among
  Carlson's  trophic state index values in regional  water
  quality assessment," by R.A. Osgood. Water Resour. Bull.
  19 (2):307-8.

Harris,  R.J.  1975.  A  Primer of  Multivariate Statistics.
  Academic Press, New York.

Higgins,  J.M., and B.R. Kim. 1981. Phosphorus retention
  models for Tennessee Valley Authority reservoirs.  Water
  Resour, Res.  17(3):571-6.
Kratzner,  C.R.,  and  P.L  Brezonik.  1981.  A  Carlson-type
  trophic  state index  for  nitrogen in Florida  lakes.  Water
  Resour. Bull.  17 (4):713-15.
Osgood,  R.A.  1982.  Using  differences  among Carlson s
  trophic state index values in regional water quality assess-
  ment. Water Resour. Bull. 18 (1):67-74.
Oskam, G. 1973. A kinetic model  of phytoplankton growth
  and  its use in algal  control by reservoir mixing.  Pages
  629-31   in  Man-made  Lakes:  Their  Problems and  En-
  vironmental Effects. Geophys. Monogr. Ser. Vol. 17.
Reckhow, K.H. 1981. Lake data analysis and nutrient budget
   modeling. EPA-600/3-81-001. U.S.  Environ. Prot.  Agency,
   Corvallis, Ore.

Shannon,  E.E.,  and  P.L Brezonik. 1972. Eutrophication
   analysis: a multivariate approach. Environ. Eng. Div., Am.
   Soc. Civil Eng. 98 (SA1):37-57.

U.S. Environmental Protection Agency. 1978. National Eutro-
   phication Survey Compendium. Work. Pap. 474-7. Corvallis
   Environ. Res. Lab. and Las Vegas Environ. Monitor. Sup-
   port Lab.

Walker, W.W. 1979. Use  of hypolimnetic oxygen depletion
   rate as a trophic state index for lakes. Water Resour. Res.
   15(6):1463-70.

,	1980. Variability  of trophic  state indicators  in
   reservoirs. In Restoration of Lakes  and Inland  Waters.
   Proc. Int. Symp. Inland  Waters  and Lake Restoration.
   Portland, Maine. EPA-440/5-81-010. Environ. Prot.  Agency,
   Washington, D.C.

	1982. Calibration and testing of  a eutrophication
   analysis procedure  for Vermont lakes. Prepared  for Ver-
   mont Agency of Environ. Conserv.  Dep. Water Resour. En-
   viron. Eng., Lakes Program. Final Rep.
	1983. Empirical methods for predicting eutrophica-
   tion in  impoundments —  Phase  II extension: model
   refinements. Prepared  for  Off. Chief, Army  Corps Eng.
   Tech. Rep. E-81-9. Waterways Exp. Sta.,  Vicksburg, Miss.
   Draft.
                                                      440

-------
TROPHIC STATE INDICES: RATIONALE
FOR MULTIVARIATE APPROACHES
PATRICK L BREZONIK
Department of Civil & Mineral Engineering
University of Minnesota
Minneapolis, Minnesota
            ABSTRACT

            Trophic state indices (TSIs) have been widely used to rank and classify lakes. Applications and
            limitations of TSIs are discussed, and types of TSIs developed previously are reviewed. An index
            is a summary statistic, and most trophic indices have been multivariate metrics, reflecting the
            complexity of the concept of trophic state. A recently developed TSI scheme for Florida lakes is
            described. This index includes sub-indices for the major physical, chemical, and biological in-
            dicators of trophic conditions (Secchi disk transparency,  total P and N  concentrations, and
            chlorophyll a levels), which were developed from a data base of 313 lakes. Use of the index to
            rank Florida lakes and evaluate problem lakes is described.
INTRODUCTION

Trophic state is the  integrated expression of nutri-
tional status in surface waters. As such, no single in-
dicator is adequate to completely describe or quantify
the concept, and limnologists have used many phys-
ical, chemical, and biological indicators to describe
trophic state. The difficulty with this multivariable ap-
proach is  that quantification  becomes  difficult and
ambiguous. In the past, limnologists reviewed data on
trophic indicators and assigned a lake to a certain
trophic class by a simple nomenclatural system (oligo-
trophic, mesotrophic, eutrophic categories). This ap-
proach is  now  regarded as  too  subjective and im-
precise.
  One way to avoid  nomenclatural imprecision  is to
quantify trophic state by an  index (or by several in-
dices). Numerous trophic indices have been proposed
over the past 15 years. They vary widely in choice of in-
dicator variables, method of development, mathemati-
cal  complexity, precision,  quantitativeness, and ac-
ceptance by limnologists. Reckhow (1981) described
some recent indices for the U.S. EPA Clean Lakes Pro-
gram. The most important examples of these indices
are described briefly in this paper.
TROPHIC INDICES REPORTED IN
THE LITERATURE

Indices  are often  used to  simplify complicated
phenomena.  An  index  essentially  reduces  a
phenomenon to a single dimension or variable, which
usually is a function of several measured variables.
Commonly-cited indices in daily life are the gross na-
tional product (GNP), which  represents the entire
economic activity of a country in a single figure, the
cost of living index, and the Dow-Jones average in the
stock market. As Reckhow (1979) pointed out, an index
is a summary statistic used because its convenience
outweighs the disadvantage of  information lost  in
summarization. Of  course,  the original data can be
retrieved if the summary statistic provides insufficient
detail for the analysis.
  Indices have become popular in the field of environ-
mental quality. At least four reasons can be given for
developing  trophic state indices (Brezonik, 1976): A
numerical index is helpful in conveying information to
the nontechnical public, useful in comparing overall
trophic  conditions between  lakes,  as a  means  to
evaluate the direction and rate of trophic change, and
finally, in developing empirical models of trophic con-
ditions as functions of watershed "enrichment" fac-
tors.
  Aside from semiquantitative indices based on in-
dicator organisms, which  have been used for many
years, the first indices used to compare lake trophic
state were  simple ranking  schemes for closed data
sets. For example, Lueschow  et al. (1970)  ranked 12
Wisconsin lakes based on five trophic indicators. The
rankings according to each parameter were summed
to yield a composite ranking for each lake. A some-
what more sophisticated approach uses proportional
rankings for each parameter. These are obtained as
the difference between the value for a lake and the
minimum (or maximum) value for the data set, divided
by the  range for the parameter among all lakes. The
U.S. Environmental Protection Agency National Eutro-
phication Survey used a percentile system to rank the
lakes in each State based on six trophic indicators.
  Although relative  ranking schemes are useful  to
compare lakes in a closed data set, they have obvious
limitations for open data sets and even for very large
closed  sets. Consequently,  most recent trophic rank-
ings involve indices based on an absolute scale. Some
indices  involve  well-defined  and easily quantified
variables (like chlorophyll  a)  and Secchi  disk (SD)
transparency (Carlson, 1977), whereas others involve
ill-defined variables  like aesthetics  (Wis.  Dep. Nat.
Resour.,  1975) and use impairment (Uttormark  and
Wall, 1975).
  Parameters like use impairment are defined on ar-
bitrary,  dimensionless scales, and determination  of
the value for a given lake is subjective. For example,
Uttormark and Wall  (1975) developed a "Lake Condi-
tion Index" (LCI) based on  ratings  of  lakes in  four
categories:  dissolved oxygen (0-6 points), transparen-
cy (0-4  points), fishkills (0-4 points), and use impair-
ment (0-9 points). In each category, zero represented
the most desirable condition; penalty points were
                                               441

-------
LAKE AND RESERVOIR MANAGEMENT
assigned based on subjective evaluation of problem
severity. The LCI is the sum of the penalty points for
the four categories.
  The first quantitative trophic state index (TSI) in-
volving indicators like transparency, chlorophyll, and
nutrient  levels  was  developed  by  Shannon  arid
Brezonik (1972). The authors used data  from oligo-
trophic to hypereutrophic Florida lakes. Trophic state
was defined in terms of seven indicators: chlorophyll
a, Secchi disk transparency, total nitrogen (TN), total
phosphorus (TP), primary productivity, conductivity,
and a divalent/monovalent cation ratio. The index was
developed by a multivariate statistical method, prin-
cipal component analysis (PCA). Principal component
variables (PCVs) are obtained from the characteristic
roots of a correlation matrix of observed variables; the
first PCV explains the  maximum  variance  in the
original data.  In Shannon and Brezonik's case, the
first PCV accounted for 70  percent of the variance,
and  this variable was considered to represent the
underlying concept (trophic state) described in part by
each indicator. The Shannon-Brezonik TSI was a sim-
ple  transformation of  the  first  PCV,  scaled 1o
eliminate  negative  values.  This  index was used  n
Florida for 10 years but suffers from several problems,
including the limited geographic extent of the dala
base, the difficulty in measuring some indicators, and
lack of an intrinsic relationship between other in-
dicators and the central phenomenon of  eutrophica-
tion.
  Of more recent trophic indices, the most widely us-
ed is that of Carlson (1977). His approach is attractive
because of its theoretical basis and reliance on tre
quantified  indicators  (Secchi  transparency, chloro-
phyll a, and TP). Separate  indices were  derived  for
each  parameter  from  data  on temperate  lakes.
Carlson's goal was to have a 10 unit change in each in-
dex represent a doubling or halving of  algal biomass.
Each index was derived  for a single indicator; the in-
dices thus are  univariate.  The  three variables are
highly  intercorrelated  and  can be  considered  £.s
estimators of the same phenomenon (trophic state or
algal biomass).
  Although chlorophyll a is the most direct measure
of algal biomass, Carlson chose Secchi  disk trans-
parency as the primary indicator and derived an inde*
that  decreased 10 units for  every doubling of trans-
parency. The  index  was  scaled so that  TSI = 0
represents a transparency of 64 m (>the largest value
in literature), and TSI = 50 represents a transparency
of 2 m (the approximate demarcation between oligo-
trophic  and eutrophic  lakes. Carlson developed in-
dices for chlorophyll a and TP by substituting relation-
ships between these variables and Secchi disk into
TSI(SD). The  relationship between Secchi disk and
chlorophyll a was nonlinear and thus a  10 unit change
in TSI (chl a) does not represent a factor-of-two change
in chlorophyll  a. Instead, chlorophyll  a doubles for
each = 7 unit increase in TSI (chl a) (Carlson, 1980).
  Carlson (1977) recommended against using the
average of the three indices because this results in a
loss  of information. For a  "well-behaved" lake in
which  relationships among the variables follow the
regression  relationships, the  three indices should
have the same value. Carlson argued that where index
values are different, differences provide useful infor-
mation. For example, if TSI(TP) > TSI(chl a), phos-
phorus probably is not the limiting nutrient; TSI(SD) >
TSI(chl a) indicates the presence of nonalgal turbidity.
On the other hand,  these inferences could be made
from the untransformed  data without use of indices.
 As mentioned earlier, indices usually are integrative;
 loss of detail is often accepted as a trade-off for ad-
 vantages of integration. The desire to express trophic
 state on a single numerical scale has led some scien-
 tists  to combine Carlson's  three indices and report
 their  simple average.
   Several trophic state indices have been developed
 that rely on Carlson's approach. Kratzer and Brezonik
 (1981) developed a TSI based on TN  concentration
 from  data on  N-limited Florida  lakes. Kratzer and
 Brezonik proposed that the lesser of TSI(TN) and
 TSI(TP) indicates the limiting nutrient for a lake. They
 combined  the  smaller of these  with TSI(SD) and
 TSI(chl a) to produce an average TSI that integrates
 the major physical, chemical, and biological features
 of trophic state.
  Porcella et al. (1980) outlined a "Lake Evaluation In-
dex,"  LEI (as yet incomplete), based partially on  Carl-
son's  indices. The  index  is composed of up  to 6
variables (Secchi disk, Chl a, TP, TN,  dissolved O2,
macrophyte coverage) transformed into subindices
(X-values) that are averaged to produce the LEI. Walker
(1979) developed indices analogous to Carlson's that
offer some  advantages over the latter TSI's. Walker's
indices are based on chlorophyll a (probably the best
measure of the central problem of eutrophication —
an increase in  algal  biomass), and his TSI(SD)  in-
cludes a correction term for nonalgal turbidity.
  In  summary,  a large variety of trophic state and
"lake  condition" indices are available in the literature.
Many have similar features, and most are multivariate.
Most have been derived from data on temperate lakes,
but the approaches  and positive features of these in-
dices  are useful in  developing a trophic state index
suitable for subtropical lakes.


 BASIS FOR A TROPHIC STATE INDEX
 FOR FLORIDA LAKES

The approach used  by Carlson in deriving his indices
has many advantages, but direct application of his in-
dices to Florida lakes has some disadvantages which
can be summarized  as follows. Carlson's indices were
based on interrelationships among Secchi  disk, TP,
and chl a in north temperate lakes. Studies by Baker et
al. (1981) indicate that Florida lakes have somewhat
different relationships among these variables. Indices
for Florida  lakes should be based on regression rela-
tionships for Florida lakes. Carlson's index assumes P
is the limiting nutrient, but many Florida  lakes are
N-limited. Carlson's indices were based on  Secchi
disk.  An increase  in plant  biomass is the central
phenomenon of eutrophication, and it seems more ap-
propriate to base the index on a direct measure of
plant  biomass (like chl a). Carlson's indices do not ac-
count  for macrophyte problems. Such problems are
common in Florida lakes, and it would be desirable to
quantify and relate them to nutrient loading.
  Based on these considerations, a series of indices
was developed (Huber et al. 1982) to rank and classify
Florida lakes according lo trophic state. In each case,
the index represents the average of the main physical,
chemical, and biological expressions of the trophic
state   concept.  Secchi  disk  transparency  is  the
physical measure, chl a the biological  measure, and
concentrations  of TP  and/or TN are  the  chemical
measures. Subindices were developed for each  para-
meter. Macrophytes were considered separately and
not as a subindex (this aspect will not be discussed in
detail  here).   Different  nutrient   subindices   were
                                                442

-------
                                                                                         TROPHIC STATUS
developed depending on whether the lake is primarily
P-limited, primarily N-limited, or relatively balanced.
DEVELOPMENT OF SUBINDICES

An  increase in plant  biomass is the  central pheno-
menon  of eutrophication,  and  algal  biomass  as
measured by chl a was selected as the basis for the in-
dices. An index [TSI(chl a)] was developed such that a
doubling of chl a increases the index by 10 units. Sub-
indices  for Secchi disk transparency and TP and TN
concentrations were derived by substituting regres-
sion-derived relationships between these variables
and chl  a concentration into TSI(chl a).
  Relationships between chl a and the cited variables
were determined by regression analysis by least ab-
solute value (LAV)  regression, which  minimizes the
sum of the absolute deviations between the predicted
line and the data points (Barr et al. 1976). LAV regres-
sion is  preferred where data are widely scattered,
since it  is  less influenced by outliers. The LAV regres-
sion line is the median  line  of best fit; half the data
points are on  either side of it.  In contrast,  a least-
squares regression line may not be close to the me-
dian line, since the regression line will be skewed to
the side with the greatest magnitude outliers.
  Data  from 313  lakes  in  the  FLADAB data  base
(Huber et  al. 1982) were used. These  lakes represent
the best and most complete data at  the time of the
analysis. Later regressions were made using all data
from 573 lakes in FLADAB. Results differed insignifi-
cantly from the earlier runs.
  TSI (chl a). This subindex was developed based  on
two criteria: (1) doubling chl  a would increase the  in-
dex by  10 units, and  (2) an  index value of 50 would
represent chl a = 10 mg/m3 (the approximate dividing
line between  eutrophic  and noneutrophic  lakes).
TSI(chl a)  = 60 represents chl a = 20 mg/m3, which
corresponds to the "ad hoc" upper limit used by the
Florida  Department of Environmental Regulation  to
define problem lakes  (Hand, 1980). The  equation for
TSI(Chl  a)  is:
    TSIfChl a) =  16.8 + 14.4 1n chl a
(1)
The factor 16.8 is a scaling term so that chl a =  10
mg/m3 yields TSI = 50. Table 3-1 lists chl a concentra-
tions corresponding to TSI values from 0 to 100.
  TSI(SD). The subindex for 3D was obtained from
LAV regression between  Secchi  disk  transparency
and chlorophyll/chlorophyll a for the 313 lakes. Uncor-
rected (total) chlorophyll/chlorophyll a values were us-
ed since only about half of the studies in the data base
reported  phaeophytin  corrected  values.  The LAV
regression is:

    InSD = 1.46 - 0.484 1 n chl a (p<0.0001)      (2)

  There is no  widely accepted measure of goodness
of fit  for LAV  regression, and  analysts must rely  on
visual inspection and comparison of sums of absolute
values of the  residuals. Visual comparison of least
squares and LAV regression plots favored LAV regres-
sion (Huber et  al. 1982), which yielded a better descrip-
tion  of the overall  relationship  between  the two
variables. Substitution of the LAV expression for Sec-
chi disk versus chlorophyll a into (1), rearranging, and
simplifying leads to TSI(SD):
       Based on equation (3) a transparency of 1.0 m cor-
       responds to a TSI of 60, and a transparency of 2.0 m
       yields TSI s 40 (Table 1).
         TSI(TP). Based on a review of the literature, it was
       concluded  that lakes with  TN/TP > 30 (wt/wt) are
       P-limited. Lakes with TN/TP < 10 (wt/wt) may be con-
       sidered potentially nitrogen-limited, and lakes with in-
       termediate TN/TP ratios have  a relatively balanced
       nutritional  status.  Separate chlorophyll  a nutrient
       regressions were developed for these three subsets of
       conditions  to express the relationship between  algal
       biomass and limiting nutrient concentration.
         A subset of 95 P-limited lakes  with TN/TP ratios >30
       was obtained from the 313 lakes, and regression of un-
       corrected chlorophyll a versus TP by LAV yielded
           Inchla = 1.641nTP - 2.85
                                               (4)
      where both chlorophyll a and TP are expressed in ^g/l.
      Substitution into equation (1), rearranging, and simpli-
      fying leads to
           TSI(TP) = 10(2.36 1nTP - 2.38)
                                               (5)
      A TP of 20 ng/l corresponds to a TSI of 47; 50 ng/l
      (generally accepted as in the eutrophic range) yields a
      TSI of 69 (Table 1). Because lakes with TNfTP > 30
      reasonably  can be assumed to be wholly  P-limited,
      equation (4) should be the best predictive relationship
      between chlorophyll a and limiting nutrient concentra-
      tions for such lakes. The corresponding TSI thus is the
      best  nutrient-related  estimator of trophic  state for
      such  lakes.
         TSIfTN).  For potentially  N-limited lakes,  algal
      biomass should relate more closely to TN concentra-
      tions, and an index based  on TP would be a poor
      predictor of algal biomass (and trophic  state) in
      N-limited lakes. A subset of 50 lakes with TN/TP < 10
      (wt/wt) was obtained from the 313  lakes,  and  LAV
      regression of uncorrected chlorophyll a versus TN was
      obtained:
                                                         Inchla = 2.97 + 1.49 1nTN (p<0.001)
                                                     (6)
      where TN is in mg/l and chlorophyll a in /^g/l. Substitu-
      tion of  the relationship  into eq. (1) resulted  in the
      following TSI;
           TSI(TN) = 10(5.96 + 2.15 1nTN)
                                               (7)
      A TSI of 50 corresponds to TN  = 0.64 mg/l, and TSI of
      60  implies TN =  1.02  mg/l (Table 1). These values
      seem reasonable for slightly eutrophic Florida lakes.
        TSI (Nutrient Balanced Lakes). Lakes with TN/TP
      between 10 and 30 exhibit relatively well-balanced
      nutrition, and  it is not  possible to assign a single
      limiting nutrient  to such  lakes.  Recent  evidence
      (Smith,  1982)  suggests  that such  lakes respond to
      changes in loadings and concentrations of either N or
      P. Thus, it is appropriate to relate chlorophyll a levels
      to concentrations of both nutrients and to use both to
      define the chemical aspect  of trophic state. LAV
      regressions of chl a vs. TN and  TP were obtained from
      a subset of 169 lakes with TN/TP ratios between 10
      and 30:
          Inchla = 1.29 1nTP - 2.44
          Inchla = 1.37 1nTN + 2.7
                                               (8)
                                               (9)
    TSI(SD) = 10[6.0 - 3.01nSD]
(3)
Substitution of the regressions into eq. (1) leads to the
following subindices:
                                                 443

-------
LAKE AND RESERVOIR MANAGEMENT
     TSI(TN) = 10(5.6 + 1.981nTN)
     TSI(TP) = 10(1.861nTP - 1.84)
(10)
(11)
 The best  chemical measure  of  trophic  state  in
 nutrient-balanced lakes was considered to be the
 average of the TSI's for TN and TP appropriate to this
 subset (TSI(NUTR) =  0.5 [TSI(TN) +  TSIfTP)], where
 TSI(TN) is from equation (10) and TSI(TP) is from equa-
 tion (11).
   Macrophyte-Trophic State Relationships.  Develop-
 ment of a TSI to quantify macrophyte problems proved
 to be more difficult. Based on the following analysis
 and other considerations (Huber et al. 1982), it appears
 that such an index should be separate from any index
 (or indices) that quantify algal-related conditions  in
 lakes, such as the subindices developed above.
   There  were 167  Florida lakes that had both  a
 measured  chlorophyll  a  concentration  and  an
 estimate of percent macrophyte coverage. A  least
 squares regression of the two variables had an r2 of
 0.027, indicating that chlorophyll a and macrophyte
 coverage  appear to  be independent  variables  in
 Florida lakes. Moreover,  because rooted macrophytes
 obtain most  of their nutrients from the sediments,
 macrophyte problems may occur in lakes with relative-
 ly low nutrient concentrations (and thus relatively low
 levels of algal biomass). As a result, close correlations
 may not exist between in-lake nutrient concentrations
 and macrophyte  density.  Because  macrophyte
 coverage and  nuisance  problems cannot be closely
 correlated with other quantifiable indicators of trophic
 state, it may be more useful to consider macrophyte
 problems as a separate  issue with a separate indax
 (see Huber et al. 1982).
weighting factors for the parameters used here. The
parameters thus contribute approximately equally to
the variance in the composite variable (or the overall
TSI). Simple averages  also are easier to compute,
understand,  and interpret than are  results of  other
methods. If an index is to receive public acceptance, it
should be simple to understand. Finally, the "true"
weighting that should  be assigned to components
comprising the overall concept of  trophic state is
unknown and essentially unknowable; on this  basis
the simplest weighting procedure is justified.
  The algal-related average TSI for the three ranges of
TN/TP ratios are as follows:
       P-limited lakes

       TSI(AVE) = 1/3[TSI(chl a)
                         TSI(SD) + TSI(TP)]    (12)
      N-limited lakes

      TSI(AVE) = 1/3[TSI(chl a) f TSI(SD) + TSI(TN)]    (13)

      Nutrient-balanced lakes
           TSI(AVE =  1/3 [TSI(chl a) + TSI(SD) +
           0.5[TSI(TP) + TSI(TN)]                     (14)

      where TSI(TN) and (TP) for each average are determin-
      ed from the subindex  appropriate for the TN/TP ratio
      of the lake. For comparison,  a regression of the
      average TSI's based on Carlson's three TSI's versus
      the average TSI based on the Florida equations was
      done. The least square regression is


      TSI (Carlson) = 0.68TSI(Florida)  + 21.9r2 = 0.77  (15)
INTEGRATION  OF SUBINDICES INTO
AN OVERALL TSI

The overall trophic  state index was determined by
combining the appropriate  subindices to obtain an
average for  the  physical, chemical, and biological
features of trophic state. Although more sophisticated
techniques  (such  as principal component analysis)
are available to combine multivariate data into a com-
posite variable,  the  additional effort probably  is not
warranted. Previous multivariate TSI's  developed by
PCA (Shannon  and  Brezonik,  1972)  have similar
      APPLICATION OF THE FLORIDA TSI

      Using these equations, 573 Florida lakes were ranked
      by the new TSI's. The results range from a high of 141
      (Banana Lake in Seminole County) to  a low of -15
      (Lake Theresa  in  Volusia County).  Seventy lakes
      among  the 573  did  riot  contain  either N or  P
      measurements.  To evaluate their TSI's, these lakes
      were assigned to the nutrient-balanced category, and
      eq. (14) was  used  with  TSIfTP) or TSI(TN)  as ap-
      propriate (without: the factor 0.5). Similarly,  if chl a or
      Secchi  disk  measurements   were  not   available,
                      Table 1.—Trophic indicator ve lues associated with subindex values.


TSI(i)

0
10
20
30
40
50
60
70
80
90
100
• TN/TP > 30 (wt/wt)
" TN/TP < 10
"• 10< TN/TP < 30


Chl a
0
-------
                                                                                               TROPHIC STATUS
TSI(AVE) was computed as an average of the remain-
ing TSI values.
  A primary issue regarding use of a TSI for manage-
ment purposes is the selection of  a critical  value,
above which a lake  is considered to have trophic-
related problems. A summary of the criteria discussed
earlier is listed below:

     SD transparency < 1  m = (TSI = 60)
     Chi a >20 mg/m3 -» problem (TSI  = 60)
     TP >50 ^g/l -* problem (TSI = 69)
     TN > 1 mg/l - problem (TSI = 60)

Overall, it appears reasonable to use TSI = 60 as the
criterion. A frequency distribution of TSI values for the
573  lakes (Table 2)  shows  that 399  lakes  have TSI
values below 60;  these may be viewed as having no
problems or less urgent problems. Of the remaining
174 lakes, many have only limited data to support the
supposition  that  they have a problem.  The median
number of sample dates  for  the 573 lakes is about
four.  Deleting lakes with fewer than  five samples we
compiled a list of 90 "problem lakes" (with TSI >60).
The highest ranking  lake  on this list is Banana Lake
(Polk County) with TSI 95. Several familiar and well-
studied lakes are on the list:  Apopka, Griffin, Eustis,
Thonotasassa and Okeechobee.
Table 2.—Frequency distribution  of  TSI values  for 573
                     Florida lakes.
TSI Range
>100
90-100
80-89
70-79
60-69
50-59
40-49
30-39
20-29
10-19
0-9
Number of Lakes
4
8
19
46
92
139
129
85
32
13
6
          Table 3.—TSI by lake hydrologic type.
Lake Type
Unspecified
Inflow
Outflow
Inflow-Outflow
Landlocked
Number
of
Lakes
171
44
30
163
165
Mean
TSI
53.0
48.9
48.2
54.3
48.8
Standard
Deviation
18.6
13.8
16.3
15.8
19.8
          Table 4.—TSI by lake area (hectares).
Lake Area
(ha)
Unspecified
1-40
41-100
101-200
201-400
401-2000
2001-4000
>4000
Number
of Lakes
65
178
108
68
53
68
19
14
Mean
TSI
52.4
51.5
48.5
55.4
44.8
52.0
54.3
60.5
Standard
Deviation
22.9
17.9
17.7
17.8
14.8
15.2
15.1
13.0
   Hydrologic type is not a major factor influencing
TSI, although lakes with both surface inflows and out-
flows had the highest mean TSI (Table 3). The range in
mean TSI's for the other types was quite small (47.1 to
50.5). Similarly, there is little trend in mean TSI versus
lake surface area (Table 4). The lowest mean TSI (49.3)
occurred in relatively small lakes (40-100 ha) and  the
highest mean TSI (58.9) occurred in the largest lakes (>
4000 ha). A consistent trend of increasing  minimum
TSI values with increasing surface area is evident. The
smallest lakes (<40 ha) had TSIs as low as  4.6;  the
minimum TSI in the largest lakes (»4000 ha) is 40. Use
of the TSI to evaluate problem conditions in individual
lakes and  to  relate problem conditions to watershed
factors is under  investigation.

ACKNOWLEDGEMENTS: Assistance  of Bob  Dickinson in
compiling lakes for FLADAB and in conducting the computer
analyses is greatly appreciated.  Development  of  Florida
TSI's was supported by a U.S. EPA Clean Lakes grant to the
Florida Department of  Environmental  Regulation via a con-
tract with the University of Florida. Cooperation of Wayne
Huber, James Heaney, and Vernon Myers in the lake classifi-
cation contract is  greatly appreciated.


REFERENCES

Baker, L.A., P.L Brezonik, and C.R. Kratzer. 1981. Nutrient load-
  ing - trophic state relationships in Florida lakes. Publ. 56,
  Water Resour. Res. Center, Univ. Florida, Gainesville.
Barr, A.J., J.H.  Goodnight, J.P. Sail, and  J.T. Helwig. 1976. A
  user's guide to SAS. SAS Institute, Inc., Raleigh, N.C.
Brezonik, P.L 1976. Trophic  classifications and trophic state
  indices: rationale, progress,  prospects. Rep.  ENV-07-76-01.
  Dep. Environ. Eng. Sci., Univ. Florida,  Gainesville.
Carlson,  R.E. 1977. A trophic  state index for lakes.  Limnol.
  Oceanogr. 22:361-9.

	1980. More complications in the chlorophyll-Secchi
  disk relationship. Limnol. Oceanogr. 25:379-82.
Hand, J.  1980. Pers. comm. Fla. Dep. Environ. Reg., Tallahassee.

Huber, W.C., et al. 1982. A classification of Florida lakes.
  Comple. Rep. Fla. Dep. Environ. Reg. ENV-05-81-1. Dep. En-
  viron.  Eng. Sci.,  Univ. Florida, Gainesville.

Kratzer,  C.R., and  P.L. Brezonik. 1981.  A Carlson-type trophic
  state index for nitrogen in Florida lakes. Water Resour. Bull.
   17(4):713-15.
Lueschow, L.J., J.  Elm, D. Winter, and G. Karl. 1970. Trophic
   nature of selected Wisconsin lakes. Wis. Acad. Sci. Arts Let-
  ters 58:237-64.
Porcella, D.B., S.A.  Peterson, and D.P. Larsen. 1980. An index to
  evaluate lake restoration. Proc. Am. Soc. Civil Eng. J. Environ.
   Eng. Div.  106(EE6):1151-69.
Reckhow, K.H. 1979. Quantitative tools for trophic assessment.
  Spec.  Rep. U.S. Environ. Prot. Agency, Washington, D.C.

Shannon,  E.E., and  P.L.  Brezonik.  1972. Eutrophication
   analysis: a  multivariate  approach.  Proc. Am. Soc. Civil
   Eng. J. San. Eng. Div. 98(SA1):37-57.
Smith, V.H., Jr. 1982. The nutrient and  light dependence of
   phytoplankton productivity. Ph.D. Thesis. Univ. Minnesota,
   Minneapolis.
Uttormark,  P.O., and J.P. Wall. 1975.  Lake classification —
   A  trophic  characterization  of  Wisconsin   lakes.
   EPA-600/3-75-033, U.S. Environ. Prot. Agency, Washington,
   D.C.
Walker,  W.W., Jr. 1979. Use of  hypolimnetic oxygen depletion
   rate as a trophic state index for lakes. Water Resour. Res.
   15(6): 1463-70.
Wisconsin  Dep.  of Natural  Resources.  1975.  Lilly  Lake
   protection and rehabilitation project.  Prop, submitted U.S.
   Envrron.  Prot. Agency, Madison.
                                                     445

-------
 ASSESSING THE TROPHIC STATUS OF LAKES
 WITH  AQUATIC MACROPHYTES
DANIEL E. CANFIELD, JR.
Center for Aquatic Weeds
University of Florida
Gainesville, Florida

JOHN R.  JONES
School of Forestry,  Fisheries, and Wildlife
University of Missouri
Columbia, Missouri
            ABSTRACT

            We propose that as a first approach the trophic status of natural and artificial lakes having growths
            of aquatic macrophytes may be assessed by using the total nutrient concentration in the water col-
            umn (nutrients contained in the macrophytes plus those in the water) in conjunction with existing
            classification systems. We developed our approach because current approaches for assessing the
            trophic status of lakes do not adequately classify lakes dominated by aquatic macrophytes. This oc-
            curs because conventional sampling and trophic state assessment emphasize conditions in the water
            and do not consider the nutrients, plant biomass, or organic production associated with macrophytes.
            Relationships between aquatic macrophytes and other trophic indicators are discussed because
            changes in macrophyte abundance influence the structural and functional characteristics of lakes.
            These changes alter perceptions of water quality and overall lake quality.
INTRODUCTION

Aquatic macrophytes are found in nearly all of the
world's lakes.  In  many, these  plants contribute
significantly to nutrient cycling and primary produc-
tivity  (cf. Wetzel, 1975; Ewel and  Fontaine, 1983;
Shireman et al.  1982). Studies of macrophytes  have
also shown that these plants can influence the struc-
ture and function of other biotic communities within
a lake (Shireman et al. 1982). Yet with the exception
of the Lake Evaluation Index (Porcella et al. 1979), ex-
isting lake classification systems use only classical
indicators such as open-water nutrient concentra-
tions, algal biomass expressed as chlorophyll a, and
water transparency as measured by using a Secchi
disk to assess lake trophic status (e.g., Likens, 197i>;
Carlson, 1977; Forsberg and Ryding, 1980; Kratzer
and Brezonik, 1981).
  Even the Lake Evaluation Index, which includes a
term for macrophyte coverage, gives no considera-
tion to the  nutrients, plant biomass, or organic pro-
duction associated with macrophytes. Thus,  large er-
rors in trophic  state assessment  can occur when
classifying  macrophyte-dominated lakes. For exam-
ple, in Lake Baldwin (Fla.), Secchi disk transparen-
cies were greater than 5 m, total phosphorus concen-
trations averaged  11 mg/ms,  and chlorophyll a con-
centrations were less than 3  mg/m3 when extensive
growths (156 g  dry wt/m2) of hydrilla (Hydrilla ver-
ticillata) covered 80 percent of the lake bottom (Can-
field et al. 1983a).
  If only the open water nutrient and algal biomass
values were  considered, Lake  Baldwin would be
classified as oligotrophic (e.g., Forsberg and Rydinci,
1980) and given a low trophic state index (TSI) value.
Yet, the quantity of macrophytes clearly indicates
the  lake is  eutrophic. Current trophic classification
approaches classified Lake Baldwin as eutrophic on-
ly after submersed macrophytes were removed by
grass  carp  (Ctenopharyngodon  idella)  and  the
ecological structure  of  the  lake changed (macro-
phytes to phytoplankton); Secchi disk transparencies
decreased to less than 2 m, total phosphorus con-
centrations averaged 30 mg/m3,  and  chlorophyll  a
concentrations averaged 20  mg/m3 (Canfield et al.
1983a).
  The case of Lake Baldwin raised the question of
how we should assess the trophic status of lakes  if
the simple trophic standards of  total phosphorus,
total  nitrogen,  chlorophyll a,  and  Secchi  disk
transparency are unreliable trophic  indicators in
lakes  having extensive growths of aquatic  macro-
phytes. In this paper, we review our efforts to resolve
this problem. We discuss our recent proposal (Can-
field et al. 1983b) that as a preliminary approach the
trophic status of lakes having growths of aquatic
macrophytes may  be assessed  by adding  the
nutrients in the macrophytes to the nutrients in the
water  and then  using the potential water column
nutrient concentration in conjunction  with existing
classification  systems  (e.g., Carlson,  1977) to
classify the  lake.  We  also  discuss  an  empirical
multivariate  regression  equation (Canfield  et al.
1984)  describing  the  influence  of nutrient  (total
phosphorus and  total nitrogen) concentrations  and
macrophyte abundance (expressed as  a percent of  a
lake's total volume  infested) on planktonic
chlorophyll a concentrations. Finally, we discuss
problems that occur when the trophic  state concept
and trophic  state  index values are used to com-
municate lake quality.

TROPHIC STATE AND AQUATIC
MACROPHYTES
The concept  of trophic state has been reviewed  and
discussed many  times (Hutchinson,  1969;  Rodhe,
                                               446

-------
                                                                                        TROPHIC STATUS
1969; Carlson, 1979; Shapiro, 1979). However, as noted
by Carlson (1979) the meaning of the concept is still
not generally agreed upon  because of its two basic
aspects: Some limnologists define trophic state bas-
ed on the supply of nutrients entering a lake or the in-
lake nutrient  concentration whereas others prefer to
define trophic state based on the biology of the lake
(e.g., primary production  or chlorophyll a concentra-
tion).  Hutchinson (1969), however, suggested that we
should not think of oligotrophic or eutrophic  water
types, but of lakes and their drainage basins as form-
ing oligotrophic or eutrophic systems.  He further sug-
gested that trophic determinations should be based
on the total potential concentration of  nutrients since
at any given time a low concentration in the water may
result because part of the lake's nutrient supply is tied
up  elsewhere in the system  (e.g., sediments or the
bodies of organisms like  macrophytes).
  Although this approach would  be difficult to imple-
ment due to problems associated with measuring
nutrients in all  components of the system, we con-
cluded a modification of this approach might provide
a reasonable first approximation  of the trophic status
of  lakes   having  extensive  growths  of  aquatic
macrophytes. We hypothesized that as a preliminary
approach, trophic determinations could be based on a
potential water column nutrient  concentration which
would be  determined by  adding  the nutrients  in the
macrophytes to  the nutrients in the water. This ap-
proach is consistent with Hutchinson's (1969) sugges-
tion and it is consistent with methods that use in-lake
nutrient  concentrations   determined  by nutrient
loading, hydrology, and lake morphometry as a major
component of trophic state assessment (Dillion, 1975;
Vollenweider,  1968,  1975,   1976;  Canfield  and
Bachmann, 1981).
  To  test our hypothesis, we sampled  six  Florida
lakes  covering a range of limnological characteristics
(Table 1) during September and  October 1981 (the
period of peak macrophyte abundance) to determine
the nutrient content of the water and the biomass and
nutrient content of  the  submersed aquatic  macro-
phytes. We  used these  data  to  estimate the  total
potential phosphorus content of the  water column
(WCP values). Phosphorus was  emphasized  as the
criterion for trophic state assessment because phos-
phorus is often the limiting nutrient in lakes and these
Florida lakes had  nitrogen  to  phosphorus  ratios
greater than 10 (Table 1). Nitrogen, however, could be
used  for  nitrogen-limited  lakes  (see Kratzer and
Brezonik, 1981). Details of the  procedure are given in
Canfield et al. (1983b).
  Total submersed macrophyte biomass in the study
lakes  ranged from 18,100 kg dry  wt in Lake  Kerr to
2,170,000 kg dry wt in Lake Lochloosa (Table 2). The
WCP  values were 1.2 to 26 times the measured open-
water  concentrations with 20 to 96 percent of the
phosphorus associated with submersed macrophytes
(Table 2). We found the effect of macrophytes on WCP
values  depends  on the  quantity  of  macrophytes
relative to the total lake volume.  For example,  Lake
Fairview has extensive growths (49 g dry wt/m3) of
submersed macrophytes. Based on  our measured
open-water  total  phosphorus concentrations (10
mg/m3) and  conventional  criteria (Likens,  1975;
Forsberg and Ryding, 1980) we would classify  Lake
Fairview as  oligotrophic.  The  calculated  Carlson
(1977) TSI value would be 37.
 Table 1.—Average chemical conditions for the surface waters of six Florida lakes between September 1979 and August 1980
                                             (Canfield, 1981).
Lake
Down
Fairview
Kerr
Lochloosa
Okahumpka
Stella
PH
6.5
8.0
6.1
7.4
8.3
7.0
Total Alkalinity
(mg/l as CaCOj)
3
52
3
23
50
16
Specific Conductance
(^mhos/cm at 25°C)
208
173
105
77
177
239
Total P
(mg/m3)
8
15
13
36
14
13
Total N
(mg/m3)
310
450
220
1200
880
460
Chlorophyll a
(mg/m3)
1.0
2.5
1.5
32
5
3
Secchi
Depth (m)
6.2
4.8
3.3
0.7
1.2A
4.1
 ASecchi depth represents bottom readings

 Table 2.—Comparison of TSMB, total submersed macrophyte biomass; TSMP, total submersed macrophyte phosphorus; SA,
     surface area; V, volume; TP, measured total phosphorus concentration; WCP, potential water column phosphorus
      assuming 100 percent release of phosphorus from the macrophytes; and TSMB*V~1, macrophyte concentration
         measured in six Florida lakes during September-October 1981. Numbers in parentheses for TSMB are 95
          percent confidence interval. For other variables, numbers represent empirical 95 percent confidence
                 intervals calculated assuming all errors are associated with measurements of TSMB
                                        (from Canfield et al. 1983b).
Variable

TSMB (kg dry wt)
TSMP (kg)
SA (ha)
V(m3)
TP (mg»m-3)
WCP(mg»m-3)
TSMB»V-1
(g dry wt«m-3)

Kerr
18,100
(±6,100)
65
(±22)
1,130
42,000,000
8
9.6
(±0.5)
0.4
(±0.2)

Down
82,800
(±17,700)
132
(±28)
360
12,000,000
9
20
(±2)
7
(±D

Stella
139,000
( ± 30,000)
180
(±39)
123
4,300,000
12
54
(±9)
32
(±7)
Lake
Lochloosa
2,170,000
( ± 530,000)
5,640
( ± 1 ,380)
2,190
46,000,000
25
148
(±30)
47
(±12)

Fairview
211,000
( ± 48,500)
300
(±69)
114
4,300,000
10
80
(±16)
49
(±11)

Okahumpka
500,000
(±118,000)
1,050
( ± 250)
208
2,600,000
16
420
(±96)
192
(±46)
                                                 447

-------
  LAKE AND RESERVOIR MANAGEMENT
    Using the calculated WCP value (80 mg/m3), how-
  ever, the Carlson TSI value would be 67 and the lake
  would be classified  as eutrophic, which is similar to
  other lakes located in the same physiographic region
  (Canfield, 1981). In contrast, macrophyte abundance in
  Lake Kerr is negligible relative to the total lake volune
  (0.4 g dry wt/m3).  The WCP value was only 1.6 mg/rn3
  higher than the measured total phosphorus concen-
  tration of 8 mg/m3 (Table 3). Thus, the nutrients con-
  tained in the macrophytes would not affect the tropr ic
  classification of the  lake.
    To determine if our approach provides reasonable
  estimates  of open-water phosphorus concentrations
  in lakes where macrophyte abundance is low, we com-
  pared our  predicted WCP value with the measured
  open-water  phosphorus  concentration   in  Lake
  Baldwin, Fla., where submersed  macrophytes  were
  removed by grass carp (Ctenopharyngodon idella). In
  1978, Lake Baldwin supported approximately 100,000
  kg dry wt of hydrilla  which contained 140 kg  of phos-
  phorus (Shireman and Maceina, 1981; Canfield et  al.
  1983a). Open-water phosphorus values  averaged  11
  mg/m3. We calculated a WCP value for Lake  Baldwin
  of 52 mg/m3. After hydrilla was eliminated by the grass
  carp, the open-water phosphorus concentration in the
  lake averaged 30  mg/ms,  substantially lower than our
  initial calculated WCP value. However, we did not ac-
  count for the phosphorus (72 kg) retained by the grass
  carp (Canfield et al. 1983a). If the phosphorus consum-
  ed by the grass carp is subtracted from that contained
  in the macrophytes, Lake Baldwin's WCP value would
  be 31  mg/m3  which  agrees  with the measured
  phosphorus concentration. We also found that ojr
  calculated  WCP values for our six study lakes  were
  comparable to open-water  phosphorus concentra-
  tions measured in  phytoplankton dominated lakes
  located in the same physiographic region (Canfield et
  al. 1983b).
    Although using  WCP values for trophic assessment
  presents some potential problems, including  the fact
  that estimating WCP values is a relatively labor inten-
  sive process that is inconsistent with current ap-
  proaches to trophic  state classification  that  require
  minimal data (Carlson,  1977; Kratzer and Brezonik,
  1981; Osgood, 1982), we believe our approach  reduces
 the danger of incorrectly assessing trophic status for
                                macrophyte-dominated  lakes  (see  Canfield  et al.
                                1983b). This  is  especially  important because
                                regulatory and management decisions are often made
                                using open-water nutrient, chlorophyll a,  and Secchi
                                disk transparency data obtained from limnological
                                surveys. Values  for WCP  may also prove useful in
                                predicting the impact of changes in macrophyte abun-
                                dance  on  limnological  characteristics  when  the
                                nutrient supply to the lake remains unchanged. Cur-
                                rently, there are no accepted methods for evaluating
                                how open-water nutrient concentrations, chlorophyll a
                                values, and  Secchi transparencies will change with
                                partial  to complete removal of macrophyte biomass
                                by  natural factors  or  management  practices
                                (harvesting, herbicides, or herbivores).
                                  At this time, however,  we cannot provide definitive
                                criteria for when WCP values should be considered in
                                trophic state assessment. We suggest that the impor-
                                tance of using WCP values to evaluate the  trophic
                                status of lakes having aquatic macrophytes is directly
                                related to the macrophyte  abundance per volume of
                                lake  or epilimnion  (our lakes were not thermally
                                stratified).  Our analysis  indicates that macrophytes
                                have little effect on trophic state assessment when<
                                25 percent of the phosphorus in the water column is
                                associated with macrophytes  and the mean  macro-
                                phyte concentration in the lake is less than 1 g  dry
                                wt/m3 (Canfield et  al. 1983b). Until more lakes  are
                                sampled to provide such criteria, we believe that deci-
                                sions to use WCP values should be made on the basis
                                of the extent of macrophyte coverage (%  of surface
                                area) in relation to lake volume. For large deep lakes
                                with small littoral areas, the effect of macrophytes on
                                lake trophic state assessment will be negligible. Our
                                approach, however, is likely to be  most useful when
                                classifying shallow macrophyte-dominated lakes.


                                RELATIONS BETWEEN MACROPHYTE AND
                                OTHER TROPHIC INDICATORS

                                Studies of lakes having a wide range of limnological
                                conditions and located in different geographical areas
                                have  demonstrated a strong relation  between  total
                                phosphorus and nitrogen concentrations and chloro-
                               phyll a concentrations (Sakamoto, 1966; Smith, 1982;
 Table 3.—Chlorophyll a concentrations (mg/ms, predicted by use of Eq. 4) in five hypothetical cases which depict different
 c°UJb'natlons °'t0tal PhosPhoms fP)'total nitrogen (TN), and percent of the total lake volume infested with macrophytes
  (PVI). Values in parentheses are Secchi disk transparences (m) predicted using the Secchi-chlorophyll relationships of'
 	Jones and Bachmann (1978). Table is from Canfield et al. 1984.

 	Hypothetical Case
PVI
TP =
                         1
                         10
                    2
                    20
                     3
                    40
                    4
                    80
                    5
                   160
TN =

  0
  10
  20
  30
  40
  50
  60
  70
  80
  90
100
  200
3.5 (3.2)

3.1 (3.4)

2.8 (3.6)
2.5 (3.9)

2.2 (4.1)

2.0 (4.4)

1.8(4.6)

1.6(4.9)

1.4(5.3)

1.3(5.5)

1.1 (6.0)
  400
8.7 (2.0)

7.8 (2.1)

6.9 (2.2)

6.2 (2.4)

5.5 (2.5)
4.9 (2.7)

4.4 (2.8)

3.9 (3.0)

3.5 (3.2)

3.1 (3.4)

2.8 (3.6)
  800
21 (1.2)

19(1.3)
17(1.4)

15(1.5)

13(1.6)
12(1.6)

11(1.7)
10(1.8)

8.5 (2.0)

7.6 (2.1)

6.8 (2.2)
 1600
53 (.7)

47 (.8)

42 (.8)
37 (.9)

33 (.9)

30(1.0)

26(1.1)

23(1.2)

21 (1.2)

19(1.3)

17(1.4)
                                                                                               3200

                                                                                             129 (.4)

                                                                                             115 (.5)
                                                                                             102 (.5)

                                                                                              91 (.5)

                                                                                              81 (.6)
                                                                                              72 (.6)

                                                                                              65 (.7)
                                                                                              58 (.7)

                                                                                              51 (.7)

                                                                                              46 (.8)

                                                                                              41 (.8)
                                                 448

-------
                                                                                         TROPHIC STATUS
 Canfield,  1983). Other studies have shown a signifi-
 cant hyperbolic relation between water transparency
 and algal  biomass (Bachmann and Jones, 1974; Dillon
 and Rigler, 1975; Canfield and Hodgson,  1983).  From
 these  studies,  simple quantitative empirical  models
 have been  developed to  describe  these relations
 (Jones and Bachmann, 1976, 1978; Smith, 1982; Can-
 field, 1983; Canfield and Hodgson, 1983).
   Despite these research efforts, however, relations
 between  macrophytes and other trophic indicators
 such as planktonic chlorophyll  a concentrations re-
 main poorly defined. Although it  is generally recogniz-
 ed that macrophytes, especially submersed macro-
 phytes, can inhibit the development of phytoplankton
 (Hasler and Jones, 1949; Hogetsu et al. 1960; Goulder,
 1969),  existing  macrophyte studies have  provided no
 quantitative  information on  how different levels of
 macrophyte abundance influence planktonic chloro-
 phyll a concentrations in  lakes. This lack of quan-
 titative information has contributed to the problems
 associated  with classifying the trophic status of
 macrophyte-dominated lakes.
   Our initial efforts to quantify relationships between
 macrophytes and other trophic indicators  centered on
 two whole-lake manipulations (Lake Baldwin and Lake
 Pearl, Fla.) where herbicides and  grass carp were used
 to  reduce macrophyte abundance.  Details of these
 studies are given in Canfield et  al. (1983a, 1984) and
 Shireman et  al. (1983). In both studies, we found no
 relation  between macrophyte   coverage and  total
 phosphorus, total nitrogen, Secchi disk, or chlorophyll
 a values. This agrees with the findings of  Huber et al.
 (1982)  for chlorophyll  a  concentrations.  We found,
 however, that over time chlorophyll a  concentrations
 (CHLA) were inversely related to the percentage of the
 lake's  total  volume infested  with macrophytes (PVI)
 (Fig. 1). For Lake Pearl, the correlation between CHLA
 and PVI was -0.63 (P < 0.001).
   To test the hypothesis that variations in the percent
 of a lake's total volume infested with macrophytes
 could  be  a  component of  the  variance  in nutrient-
 chlorophyll regressions (Canfield, 1983), we sampled
 32  Florida lakes to determine total phosphorus (TP),
 total  nitrogen  (TN), chlorophyll  a (CHLA), and PVI
 levels. Details are given in Canfield et al. (1984). Using
 the TP, TN, CHLA, and PVI data, we developed regres-
 sion models to determine if the addition of a term for
 PVI could improve the predictive ability  of nutrient-
 chlorophyll models. Although  limited, our sampling in-
 cluded a wide range of limnological conditions. Con-
 centrations of TP ranged from 0.6 to 159 mg/m3 and
TN concentrations  ranged  from 65 to 6,020  mg/m3.
Values of CHLA ranged from 0.5 to 174 mg/m3 and PVI
 levels ranged from 0 to 95 percent. Regression models
for our data set were:
log CHLA = -0.40 + 1.09 log TP      R2 = 0.73   (1)
log CHLA = -2.24 + 1.16 log TN      R2 = 0.78   (2)
log CHLA =  -1.65 + 0.51 log TP + 0.73 log TN
             R2 = 0.82                          (3)
log CHLA =  -2.08 + 0.28 lop Tn + 1.02 log TN
             - 0.005 PVI             R2 = 0.86   (4)


By incorporating a PVI term (Eq. 4), we found that an
additional 4 percent of the variance in our chlorophyll
data was accounted for. Regression coefficients for
TP and TN in Eq. 4 were also similar to those reported
by Canfield (1983). Although small, the 4 percent in-
crease in R2 was significant and suggested, similar to
the findings at Lake Pearl and Lake Baldwin, that the
 percent of a lake's total volume infested with aquatic
 macrophytes  significantly  influences  planktonic
 chlorophyll concentrations in lakes.
   Our regression equation (Eq. 4) suggests the poten-
 tial impact of  macrophytes on chlorophyll yields  in
 lakes varies with trophic state. To assess the possible
 impact of different levels of macrophyte abundance in
 lakes ranging from oligotrophic to eutrophic, we used
 Eq. 4 to predict CHLA values for five different com-
 binations of TP and TN values (assuming phosphorus
 limitation and a TN:TP = 20:1) given PVI values rang-
 ing from 0 to 100 percent (Table 3). For a nutrient-poor
 lake such as Case 1 (Table 3), the expected  reduction
 in CHLA  with  increasing  macrophyte abundance  is
 small even if macrophytes could occupy 100 percent
 of the lake volume. Major changes in Secchi disk
 transparencies,  however,  could  occur  if  major
 changes in  PVI occurred. In Case 2  and 3 (Table 3), a
 large  increase in  macrophyte  abundance  could
 substantially reduce chlorophyll yields and change
 the trophic classification of a lake. It is also likely that
 improvements in water transparency would be noted
 by the public as PVI values increase and chlorophyll
 values decrease. For nutrient-rich lakes (Case 4 and 5),
 the change in CHLA would be large  but even at a PVI
 of 100 percent there would be sufficient chlorophyll to
 classify a lake  as eutrophic and to maintain reduced
 water transparency.
  We are not certain of the causative mechanisms of
 the inverse relationship between  chlorophyll a con-
 centrations and the percent of a lake's total volume in-
 fested with aquatic macrophytes.  Several  factors,
 however, are probably involved, including: (1) release
 and  uptake of nutrients by macrophytes and  their
 associated epiphyton; (2) reduction in nutrient cycling
 because  macrophytes reduce wind  mixing  and the
 resuspension of nutrients from the bottom sediments;
and (3)  increased sedimentation of plankton  algae
 resulting from  a  reduction in water turbulence by
 macrophytes. Whatever the mechanisms may be, our
analysis suggests that the percent of the lake's total
volume infested with aquatic  macrophytes may be  a
 useful empirical measure that can be used to assess
the impact of aquatic macrophytes on  lake chloro-
                                                    in
   70 —

   60
 b
 Z  50
 ot
 J  40


 I  3°

 o  eo

 5  .0

    0
                      	% VOLUME INFESTED «ITH HVDRILLA
                         ". HYOWLLA COVERA6E
       i  i   i  i   i  i  i  i  i  i  r i  i  i  i  n
       2  10  IB  2T  9  14  22 SO 10  l«  28 6  14  22 31  FO
      MM JUN SEP DEC APR JUL OCT JAN KAY AU6 MOV MAS JUN 8CP DEC APR
         1979     1880       IBM      1862    1988

Figure  1.—Changes  in chlorophyll a  concentrations,
macrophyte (hydrllla) coverage, and the percent of the lake's
total volume infested measured in Lake Pearl, Fla. H = a her-
bicide treatment and G = a grass carp stocking.
                                                 449

-------
 LAKE AND RESERVOIR MANAGEMENT
 phyll levels. Our sampling, however, has been limited
 and further testing is needed to test the applicability
 of our results  to other Florida  lakes and to lakes
 located in other geographical regions.
 CLASSIFICATION OF LAKES

 In the United States, Section 314 of the Water Pollu-
 tion Control Act requires all States to classify lakes
 according to trophic state. This classification process
 is intended to help prioritize lakes for possible restora-
 tion and protection. Consequently, lake classification
 has become an  integral  part of the overall  U.S.
 strategy for  developing lake management program!?.
 Various trophic state indices  have, therefore, been
 developed to evaluate trophic status and rank lakes
 according to their overall quality. As noted by Carlson
 (1980) these indices by one method or another attach a
 label or a number to the lake. However, two practical
 problems are associated with the use of  indices:  (1)
 the  various  trophic indices do not always classify
 lakes similarly, and (2) implicit with the use of the  in-
 dices is the assumption that eutrophic (high TSI) lakes
 are of poorer quality than oligotrophic (low TSI) lakes.
   The problem of different trophic state indices rank-
 ing lakes differently seems to  be  especially serious
 when classifying  macrophyte-dominated  lakes. For
 example, Huber et at.  (1982) classified 573 Florida
 lakes using  chlorophyll a,  Secchi  disk, total phosi-
 phorus, and total nitrogen data. In their classification,
 Lake Fairview (Table 2) ranked among the 50 least
 eutrophic lakes in  the State.  Using  our  calculated
 WCP value (80 mg/m3) which considers macrophytes;,
 Lake Fairview would be ranked among the 100 most
 eutrophic lakes. Recently, Myers and Edmiston (1983)
 of the Florida Department of Environmental Regula.-
 tion ranked  Lake Fairview among the top 50 lakes  in
 Florida  in need of  restoration.  These differences  in
 trophic  ranking, however, need not be  a  problem if
 considered in the proper context. Differences among
 the  different  trophic  indices  can  be  used   to
 demonstrate basic differences in the ecological struc-
 ture and function of lakes (Carlson,  1980).
  Over the last few decades, a management ethic has
 emerged that nutrient loadings to lakes should  not bo
 increased and that  lakes should  not  be  eutrophic.
 Eutrophication has often become  synonymous with
 anthropogenic pollution. Vollenweider (1968,1976) has
 used terms such as dangerous,  nonacceptable, or ex-
 cessive to describe nutrient loading rates that result
 in eutrophic  lakes.  Many  water  quality  experts,
 especially limnologists,  correlate the quality of a lake
 with characteristics that are typical of oligotrophic
 lakes (see Fusilier,  1982).  Thus eutrophic lakes or
 lakes with high TSI  values are commonly considered
 of poorer quality than are clear,  unproductive waters.
  Although  controlling  eutrophication  is a worthy
 goal, especially where water is used for multiple pur-
 poses, scientists and natural resource management
 agencies  must explicitly define their management
 criteria  when presenting various  nutrient control
 strategies. As noted by Bachmann (1980), the high
algal levels associated  with eutrophic lakes can  in-
crease the amount  of treatment needed if the lake
water is to be used for water supply. Some algae can
contribute to taste and odors in  the water. For recrea-
tional purposes, clear waters are generally more ap-
 pealing  for swimming and aesthetically pleasing, but
eutrophic waters are commonly  used.
    From  a fisheries standpoint  the choice between
 oligotrophic  and  eutrophic waters is  less  defined
 (Bachmann,  1980). In  northern  areas,  oligotrophic
 waters generally support the highly prized saimonid
 fisheries, but the overall productivity of these waters
 is low and sportfish harvest is low. Within limits, in-
 creases in nutrient inputs can increase total fish yield
 (Oglesby, 1977; Jones and Lee, 1982). For warmwater
 fisheries, the productivity of a lake can  be increased
 well above the level that causes a decline in coldwater
 fisheries. Recently, Jones and Hoyer (1982)  showed
 that warmwater sportfish harvest is directly correlated
 with  planktonic   chlorophyll  a  concentrations.
 Although there is  most likely  a level of productivity
 beyond which warmwater fishery yield is diminished
 (e.g., where oxygen depletions occur),  it is obvious
 that eutrophication can benefit  the  yield  of sport-
 fishes from lakes and reservoirs. For this reason, it is
 inappropriate to assume that a lake with a eutrophic
 classification or a high trophic state index value is a
 poor quality lake for all human uses.
   In the  future, we can  expect continued population
 growth, and with it, increased development within lake
 watersheds.  Thus, environmental  changes within
 lakes will be inevitable. Even remote lakes will be af-
 fected  by increased  anthropogenic  activities. We,
 however, have the  ability to either minimize changes
 or to exploit them to our benefit, but we must develop
 workable lake management plans that have definable
 criteria.
   We must recognize that lakes in various geographi-
 cal regions have different limnological potentials bas-
 ed on regional geology (Deevey,  1940;  Moyle, 1956;
 Jones and Bachmann, 1978; Canfield,  1981). Within a
 given  geographical region,  lake morphometry and
 hydrology will affect the trophic status  of individual
 lakes and reservoirs (Vollenweider,  1968, 1976; Can-
 field and Bachmann, 1981). Thus, our lake manage-
 ment goals must be realistic  in  their expectations.
 Many shallow lakes in fertile areas are naturally pro-
 ductive and  no reasonable amount of management
 will make them oligotrophic.
   We must also define aquatic environmental quality
 in terms of "for what" and "for whom"  (see Harvey,
 1976). The trophic state concept has proven  useful in
 limnological investigations, but continued reliance on
 the concept or trophic state indices as a management
 tool without defining management criteria will do little
 to improve our capabilities to manage lakes.

 ACKNOWLEDGEMENT: Journal Series  No.  5003,  Florida
 Agricultural Experiment Station.
REFERENCES

Bachmann, R.W. 1980. The role of agricultural sediments and
  chemicals in eutrophication. J. Water Pollut. Control Fed
  52:2425-32.

Bachmann, R.W., and J.R. Jones. 1974. Phosphorus inputs
  and algal blooms in lakes. Iowa State J. Res. 48:155-60.
Canfield, D.E., Jr. 1981. Chemical and trophic state char-
  acteristics of Florida lakes in relation to regional geology.
  Fla. Agric. Exp. Sta. J. Ser. No. 3513.

	1983. Prediction of chlorophyll a concentrations in
  Florida lakes: The importance of phosphorus and nitrogen.
  Water Resour. Bull. 19:255-62.

Canfield, D.E., Jr., and R.W. Bachmann. 1981. Predictions of
  total phosphorus concentrations, chlorophyll a and Secchi
  depth in natural and artificial lakes. Can. J. Fish. Aquat
  Sci. 38:414-23.
                                                 450

-------
Canfield, D.E., Jr., and LM. Hodgson. 1983. Prediction of Sec-
  chi disc depths in Florida lakes: Impact of algal biomass
  and organic color. Hydrobiologia 99:51-60.
Canfield, D.E. Jr., MJ. Maceina, and J.V. Shireman. 1983a.
  Effects of  hydrilla and grass carp on water quality in a
  Florida lake. Water Resour. Bull. 19:773-8.
Canfield, D.E. Jr., et al. 1983b. Trophic state classification of
  lakes with  aquatic macrophytes. Can. J. Fish. Aquat. Sci.
  40:1713-18.
	. 1984.  Predication of chlorophyll a concentrations
  in lakes: The importance of aquatic macrophytes. Can. J.
  Aquat. Sci. (In press).
Carlson, R.E. 1977.  A trophic state index for lakes. Limnol.
  Oceanogr.  22:361-9.
	1979. A review of the philosophy and construction
  of trophic  state indices. Pages 1-52 In T.E. Malone,  ed.
  Lake  Reservoir Classification System. EPA-600/3-79-074.
  Corvallis, Ore.
	1980.  Using trophic state indices to examine  the
  dynamics of eutrophication. Pages 218-21 in Restoration
  of Lakes and Inland Waters. Proc. Int. Symp. Inland Waters
  and Lake Restoration. Portland, Maine. EPA 440/5-81-010.
  U.S. Environ. Prot. Agency, Washington, D.C.

 Deevey, E.S., Jr. 1940. Limnological studies in Connecticut.
  V.  A  contribution  to regional limnology. Am.  J. Sci.
  238:717-41.
 Dillon, P.J. 1975. The phosphorus budget of Cameron Lake,
  Ontario: the importance  of flushing rate to the degree of
  eutrophy of lakes. Limnol. Oceanogr. 20:28-39.
 Dillon, P.J., and F.H.  Rigler. 1975.  A simple method for pre-
  dicting the capacity of a lake for development based on
  lake trophic status. J. Fish. Res. Board Can. 32:1519-31.
 Ewel, K.C., and  T.D. Fontaine. 1983. Structure and function
  of a warm monomictic lake. Ecol. Model. 19:139-61.
 Forsberg, C., and S. Ryding. 1980. Eutrophication parameters
  and trophic state indices in 30 Swedish waste-receiving
  lakes. Arch. Hydrobiol. 80:189-207.
 Fusilier, W.E.  1982.  The  LWQI:  An opinion derived  un-
  weighted multiplicative water quality index for lakes. Ph.D.
  Dissertation. Univ.  Michigan, Ann Arbor.
 Goulder, R. 1969. Interaction between the rates of production
  of a freshwater macrophyte and phytoplankton in a pond.
  Oikos 20:300-9.
 Harvey, H.H. 1976. Aquatic environmental quality: problems
  and proposals. J. Fish. Res. Board Can. 33:2634-70.
 Hasler, A.D.,  and E. Jones. 1949. Demonstration of the antag-
  onistic action of large aquatic plants on algae and rotifers.
  Ecology 30:359-64.
 Hogetsu, K.,  Y. Okanishi, and H. Sugawara. 1960. Studies on
  the antagonistic  relationship between phytoplankton and
  rooted aquatic plants. Japan J. Limnol. 21:124-230.
 Huber,  W.C., et al. 1982. A classification of Florida  lakes.
  Final  rep. ENV-05-82-1.  Fla.  Dep.  Environ. Reg.,  Talla-
  hassee.
 Hutchinson,  G.E. 1969. Eutrophication, past and  present. In
  Eutrophication: Causes, Consequences, Correctives. Natl.
  Acad. Sci., Washington,  D.C.
 Jones, J.R., and R.W. Bachmann. 1976. Predictions of phos-
  phorus and chlorophyll  levels in lakes. J. Water  Pollut.
  Control Fed. 48:2176-82.
        _. 1978. Trophic status of  Iowa  lakes in  relation to
   origin and glacial geology. Hydrobiologia 57:267-73.
 Jones,  J.R.,  and M.V. Hoyer. 1982. Sportfish harvest pre-
   dicted  by  summer chlorophyll  a concentration  in  mid-
   western lakes  and reservoirs.  Trans.  Am.  Fish.  Soc.
   111:176-9.

 Jones, R.A., and G.F. Lee. 1982. Recent advances in assess-
   ing impact of phosphorus loads on eutrophication-related
   water quality. Water Res. 16:503-15.
                                        TROPHIC STATUS


Kratzer, C.R., and P.L Brezonik. 1981. A Carlson-type trophic
  state index for nitrogen in Florida lakes. Water Res. Bull.
  17:713-15.
Likens, G.E. 1975. Primary production of inland aquatic eco-
  systems.  Pages 185-202 in H. Leith  and R.R.  Whittaker,
  eds. Primary Production of the Biosphere. Springer-Verlag,
  New York.
Moyle, J.B.  1956.  Relationship between the chemistry of
  Minnesota surface water and wildlife management. J.
  Wildl. Manage. 20:303-20.
Myers, V.B., and H.L Edmiston. 1983. Florida lake classifica-
  tion and prioritization project No. S004388. Final Rep. for
  Clean Lakes Program, U.S. Environ. Prot. Agency. Fla. Dep.
  Environ. Reg., Tallahassee.
Oglesby, R.T. 1977. Relationships of fish yield to lake phyto-
  plankton standing crop, production,  and morphoedaphic
  factors. J. Fish. Res. Board Can. 34:2271-9.
Osgood,  R.A.   1982.  Using  differences  among  Carlson's
  trophic state  index values in regional water quality assess-
  ment. Water Res. Bull. 18:67-74.
Porcella,  D.B.,  S.A. Peterson,  and D.P. Larsen.  1979.  Pro-
  posed method for evaluating the effects of restoring lakes.
  Pages 265-310 in S.A. Peterson,  ed.  Limnological  and
  Socioeconomic Evaluation of Lake Restoration Projects:
  Approaches  and Preliminary Results. EPA-600-3-79-005.
  U.S. Environ.  Prot. Agency, Corvallis,  Ore.
Rodhe, W. 1969. Crystallization of eutrophication concepts
   in  northern   Europe.   Pages 50-64 in  Eutrophication:
   Causes,  Consequences,  Correctives. Natl.  Acad.  Sci.,
   Washington, D.C.

Sakamoto, M. 1966. Primary production by phytoplankton
  community in some Japanese lakes  and its dependence
  on lake depth. Arch. Hydrobiol. 62:1-28.
Shapiro, J. 1979. The current status of lake trophic  indices -
  A review. Pages 53-99 in T.E. Maloney, ed. Lake  and Reser-
  voir Classification  Systems. EPA 600/3-79-074.  U.S. En-
  viron. Prot. Agency, Washington, D.C.

Shireman, J.V., and  M.J.  Maceina. 1981. The utilization of
  grass carp, Ctenopharyngodon idella  Val., for hydrilla con-
  trol in Lake Baldwin. Florida. J. Fish. Biol. 19:629-36.

Shireman, J.W., W.T. Haller, D.E. Canfield, Jr. and V.T. Van-
  diver.  1982.  The  impact of aquatic plants  and  their
  management techniques on the aquatic resources of the
  United  States: An  Overview. EPA-600/4-81-007.  U.S. En-
  viron. Prot. Agency, Washington, D.C.
Shireman, J.V., et al. 1983.  The ecological  impact of inte-
  grated chemical and biological aquatic weed control. Final
  Rep.  Submitted to U.S. Environ. Prot. Agency,  Environ.
  Res. Lab. Sabine Island, Gulf Breeze, Fla.
Smith, V. 1982. The nitrogen and phosphorus dependence of
  algal biomass in  lakes:  An empirical and  theoretical
  analysis. Limnol. Oceanogr. 27:1101-12.
Vollenweider,  R.A.  1968. Scientific  fundamentals of  the
  eutrophication of lakes and flowing waters, with particular
  reference to  nitrogen  and  phosphorus  as  factors in
  eutrophication.  Organ.  Econ.  Coop.  Dev. Tech.  Rep.
  DAS/CS1/68.27. Paris, France.

	1975. Input-output models with special reference to
  the phosphorus loading concept in limnology. Schweizer-
  ische Zeitschrift fur Hydrol. 37:53-84.

	1976.  Advances in defining critical  loading levels
  for  phosphorus  in  lake  management.  Mem.  1st.  Ital.
  Idrobiol. 33:53-83.
Wetzel. R.G. 1975. Limnology. Saunders, New York.
                                                        451

-------
                                            Macrophyte  Control
AN OVERVIEW OF CHEMICALS FOR AQUATIC PLANT  CONTROL
JAMES C. SCHMIDT

Technical Director
Applied Biochemists, Inc.
Mequon, Wisconsin


             ABSTRACT

             Scientists' role is to separate fact from fiction to assure that decisions be based upon valid evidence
             rather than emotions and misinformation. The use of chemicals for aquatic vegetation control is one
             issue commonly surrounded with fear and uncertainty by the general public, environmental groups,
             and politicians. This has led to restrictive laws in some States which severely limit or prohibit the ap-
             plication of registered aquatic pesticides. In-lake rehabilitation methods and watershed protection
             measures have achieved mixed results in controlling lake nutrient concentrations to suitably limit
             macrophyte or algae growth. For many bodies of water, high quality water is not presently economically
             or technologically feasible  However, symptomatic treatment of nuisance aquatic plant growth with
             chemicals or through integrated pest management methods can provide waterways acceptable for
             recreational and functional use.  Historically, the chemicals used for aquatic macrophyte and algae
             control were sodium arsenite and copper sulfate, respectively. These did pose a threat to the environ-
             ment from overdose and abuse in attempts to eradicate rather than manage aquatic plant problems.
             Today's chemicals and application techniques are more selective and sophisticated. Toxicity limits,
             breakdown times, and tolerances are established prior to EPA registration. Organic herbicides that
             characteristically degrade in the environment or become biologically inactive have replaced persis-
             tent chemicals such as sodium arsenite. Formulated, chelated copper compounds are replacing cop-
             per sulfate treatments as a safer, more effective approach. Improved application equipment, techni-
             ques using spray adjuvants, and  granular formulations have helped overcome problems with drift and
             uneven distribution. The judicious use of chemicals for aquatic vegetation control is imperative. Loss
             of any of our limited number of tools for maintaining suitable recreational and functional waterways
             through regulations imposed by the uninformed could severely limit our ability to achieve future water
             quality objectives.
It is necessary for scientists involved in lake manage-
ment consulting and research to evaluate their way of
thinking  about the use  of aquatic herbicides and
algaecides. The idea of using pesticide chemicals in a
body of  water often creates  immediate controversy
within  the  scientific  community.  This  issue  may
become even more emotional when brought to the at-
tention of the public and politicians. A crisis is reach-
ed when  emotions get in the way of facts, generating
misinformation that leads  to poor  decisions. Often
federally registered chemicals are banned locally or
treatment permit procurements severely restricted.
  It is not within the scope of this paper to detail the
technical data supporting the contention that present-
ly approved aquatic pesticides pose minimal risk to
man and environment when used according to label
instructions. The fact is that these products already
have been  tried and  have been  proven acceptable
through the  U.S. Environmental Protection Agency's
registration process. It is the intent of this discussion
to encourage scientists to familiarize themselves with
the chemical tools available and accept these pro-
ducts as a viable management alternative for main-
taining suitable recreational, functional, and aesthetic
waterways.
  One of  the common  arguments used  against
chemical control  of  aquatic plants is that this  ap-
proach is cosmetic  in nature,  treating  a symptom
                                                 453

-------
 LAKE AND RESERVOIR MANAGEMENT
  rather than the source of a  problem.  As a result,
  chemical treatment has not been considered a fun-
  dable management alternative under the Clean Lakes
  Program. Granted, the ultimate goal is to develop ir-
  lake rehabilitation  and watershed protection tech-
  niques   to  solve  eutrophication.  Being  realistic,
  however, attainment of high water quality to the point
  of  limiting  nuisance  aquatic vegetative  growth is
  presently not  always  economically, technologically,
  nor culturally feasible. Chemical control offers an im-
  mediate  management alternative  that can be imple-
  mented  while  other rehabilitative  or  protective
  measures are  being studied and  developed. It is  in-
  teresting to note that  a large number of the most ac-
  tive  lake community  organizations were originally
  formed  to combat nuisance  vegetative growth.  In
  many cases, these annual weed and algae control pro-
  grams have encouraged communities to rally around
  the lake  improvement cause. This  involvement  is a
  necessary prerequisite for future lake managemen:
  activities on many lakes.
   The question  of environmental impact  is always;
  raised  when a  chemical  control  program  is con-
 sidered. Certainly, this is a valid concern. Eliminatino
 vegetative growth within areas of a lake may affect the
 movement of  fish  and localized  populations of in
 vertebrates.  However,  chemical treatment allows the
 flexibility to control plants on a selective basis as tc
 species  and area.  Using chemicals to  kill aquatic
 vegetation can realistically be viewed as a means  of
 accelerating their death since plant senesence will oc-
 cur naturally later  in  the season. Therefore, claims
 that this approach contributes nutrients and sediment
 to the lake bottom  are unfounded. Actually, controll-
 ing plants in their younger stages of growth results in
 less accumulation and decay than under natural con-
 ditions. Furthermore,  preventing flowers and  seeds
 from forming may reduce the reproductive potential of
 these plant populations.
   The localized  impact of controlling  portions  of
 aquatic plant populations  with chemicals  does not
 compare  with some of the major impacts resulting
 from dredging, drawdown, nutrient deactivation, diver-
 sion, and other total lake  management  techniques.
 Although  these latter approaches are intended to im-
 prove water quality, what is the effect of total habitat
 change upon the fishery in these  less productive
 waters? Any manipulation of the delicately balanced
 aquatic ecosystem  will have some impact. It should
 not be assumed that the use of chemicals is a drastic
 technique.
  Certainly, past uses and abuses of chemicals in our
 waterways have contributed to some of these negative
 attitudes. In  addition,  pesticide scares and ground-
 water contamination have  made the general public
 quite wary when plans are proposed to put chemicals
 into water. It is the scientist's obligation to educate
 the public and  pacify unfounded concerns. Unfor-
 tunately,  urban universities,  in particular, tend  to
 stress the environmental impacts of pesticides with
 little mention of their economic necessity. They cite
 historical  examples such as the widespread  use  of
 sodium arsenite in the early and mid-1900's. Similarly,
 long-term  use of high doses of copper sulfate have led
 to residue problems from copper precipitates within
 the hydrosoil. The old approach to aquatic vegetative
 control was geared more  towards eradication than
 management and control.
  Fortunately, the herbicides used today have much
 more  environmentally  acceptable  properties.  They
characteristically  biodegrade or become biologically
  inactive.  These products  include  endothall  com-
  pounds (Aquathol K, Hydrothol 191),  Diquat, fluridone
  (Sonar), glyphosate (Rodeo),  and 2,4-D esters and
  amines. Similarly, formulated chelated copper com-
  pounds (e.g.,  Cutrine-Plus)  are  replacing copper
  sulfate treatments. These products can be used in
  lower dosages and less frequently for more effective
  control.
    EPA  registration of aquatic pesticides requires data
  on nontarget organisms, toxicity, environmental per-
  sistence,  and breakdown products.  These  products
  are viewed very similarly to  a pesticide used on food
  crops in that acceptable residue levels must be estab-
  lished.  Under cross-referenced requirements of the
  Federal Food, Drug, and Cosmetic Act and the Federal
  Insecticide, Fungicide  and Rodenticide  Act,  it  is
  necessary to determine safe concentrations of these
  pesticides in water used for  human consumption and
  raw agricultural  commodities which might be directly
  or indirectly contacted by the treated water (e.g., fish
  shellfish,  irrigated crops, etc.) These tolerances and
  exemptions from tolerance  as they are called are
  listed, by active ingredient, under chapters 20 and 40
  respectively,  of  the Code of  Federal  Regulations.'
  These clearances dictate  labeled product use sites
  (e.g.,  irrigation water, fish hatcheries, potable water
  reservoirs,  etc.), maximum allowable dosage rates,
  and water use restrictions.
   Obviously, pesticide manufacturers should promote
  the safe, proper use of their products. They do not
  want  product misuse and the consequent negative
  publicity to jeopardize the millions of dollars spent on
  product  research and development.  Product labels
  provide key information needed for proper application.
 Supplemental technical and  promotional information
  is also available for those who desire  a more in-depth
 understanding of these materials.
   Nationally,  professional interest in aquatic plant
 control has resulted in the formation of the Aquatic
 Plant  Management Society.  Within the past several
 years, regional chapters of this group have organized
 in Florida, South Carolina, the Midwest, the Midsouth,
 and the  West. In addition, the U.S.  Army Corps  of
 Engineers, which is responsible for numerous water
 management projects, generates and disseminates in-
 formation on aquatic plant control technology.
   All of these organizations hold annual conferences
 and publish proceedings, newsletters, and magazines.
 Although a significant portion of these meetings in-
 volves chemical  control  technology, sessions  on
 biological,  mechanical,  and  habitat manipulation
 techniques  are   included.  Professional  aquatic
 pesticide  applicators  and  chemical  company
 representatives who make up a significant portion of
 the audience are exposed to a wide  range  of tech-
 niques and disciplines. There is a willingness to ex-
 change ideas and develop more scientific approaches
 to aquatic plant control including integrated manage-
 ment  approaches. Technical  developments such  as
 applicating equipment, spray adjuvants, tank mixes,
 and granular formulations  have improved chemical
 distribution and drift control, and lowered application
 rates.  Professional  applicators who  are trained  as
 scientists should  become an integral part of the total
 lake management scheme.
  No  single method  will  answer all the problems
 associated with maintaining suitable water resources
 for recreational, functional, and aesthetic purposes.
Just as the study of an  aquatic system requires  a
multi-disciplinary  approach, its  management  will re-
quire integrating available technology.  Therefore, it is
                                                454

-------
                                                                                   MACROPHYTE CONTROL
imperative  that  lake  managers,  scientists,  con-    tions imposed upon these products by the uninformed
sultants, politicians, and the concerned public con-    could severely limit our ability to achieve future water
tinue to'support the judicious  use of chemicals for    use objectives.
aquatic vegetation  control.  Overly-restrictive regula-
                                                   455

-------
  EFFECTS OF MECHANICAL CONTROL OF AQUATIC VEGETATION  ON
  BIOMASS, REGROWTH RATES,  AND JUVENILE FISH POPULATIONS
  AT SARATOGA LAKE, NEW YORK
 GERALD F. MIKOL
 Bureau of Water Research
 New York State Department of Environmental Conservation
 Albany, New York
             ABSTRACT

             Experimental mechanical barge/conveyor harvesting conducted in June and August of 1981 and
             1982 significantly reduced total plant biomass initially, relative to control areas Regrowth of the
             predominant species, Eurasian watermilfoil [Myriophyllum spicatum), reached pre-harvest levels
             within 30 days after June and August harvestings. Total plant and M. spicatum biomass (q«m -2
             dry) in harvested areas peaked later than control area vegetation both years, and was slightly
             higher. The ratio of the average annual biomass of the harvested area to the control area was the
             same both years (0.73), despite significant decreases in overall average total biomass in 1982
             Similar harvesting resulted in the removal of primarily juvenile fish. Harvesting was conducted in
             dense littoral zone stands  of milfoil (M. apicatum)  and curly-leaf pondweed  (Potamogeton
             crispus). The harvesting operation removed primarily bluegill sunfish (Lepomis macrochirus) and
             to a lesser degree, largemouth bass (Micropterus salmoides), yellow perch (Perca flavescens)
             and pumpkmseed sunfish (Lepomis gibbosus). Harvesting effectively removed  approximately
             2-3 percent of the total standing crop of juvenile fish  in both June 1981 and 1982 and approx-
             imately 2-8 percent in August of both years                             '
 INTRODUCTION

 The short-term  effectiveness of aquatic vegetation
 removal by  harvesting  has been well  documentec
 (Dunst et al. 1974; Neel  et al. 1973; Nichols and Cot
 tarn, 1972;  Garrison, 1979; Wile and Hitchin, 1977;
 Bedrosian,  1982; Storch and Winter, 1983).  The im-
 mediate effect of harvesting is to improve recreational
 use by creating (1) access  lanes to areas where
 vegetation is less dense, (2) open areas that expand
 anglers' access to sport fish, and (3) open shoreline
 areas for swimming. Indirectly, harvesting can reduce
 the  potential  for  odor  and  aesthetic  problems
 associated with naturally decaying vegetation. Addi-
 tionally, when  all   other  controllable  sources of
 nutrient inputs  are reduced, removing vegetation
 could have a positive effect on the total nutrient pool
 available for vegetative  production. The results of
 several studies on the latter theory have been mixed
 (Neel et al.  1973; Bedrosian, 1982). While the total
 reduction in  nutrient availability may be  insignificant
 in many cases, the plant biomass reductions observed
 during  subsequent  growth  periods  were significant,
 and may be  related to sediment/plant nutrient inter-
 actions (Aiken and Picard, 1980; Bole and Allen, 1978;
 Kimbel and Carpenter, 1981; Carpenter  and  Adams,
 1977; Barko and Smart, 1983; Barko, 1983).


STUDY AREA

Saratoga Lake is a publicly owned body  of water ap-
proximately 48 km  north of Albany, N.Y, in central
Saratoga County (Fig. 1). It is a dimictic lake 1,630 ha
in size with a mean depth of 8 m (Table 1). Over 25 per-
cent of the surface area is less than 3 m deep. High
developmental pressure  and recreational use in the
1960's to 1970's have resulted in degraded water quali-
 ty and impairment of all recreational uses. Municipal
 wastewater treatment plant effluents were diverted
 from the watershed in 1977. Since then, water quality,
 measured  by  nutrient  levels,  Secchi   depth  and
 chlorophyll  a concentrations, has improved  signif-
 icantly.  Concurrently,  two  species   of  rooted
 macrophytes,  Myriophyllum   spicatum   and
 Potamogeton crispus, have dramatically increased in
 areal and depth distribution. In 1982, the entire lake
 surface area to the 4 m contour was impacted. While
 sport fishing for walleye (Stizostedion vitreum), nor-
                           5 km
Figure 1.—Saratoga Lake watershed boundary.
                                              456

-------
them pike (Esox luclus), and black bass (Micropterus
salmoides  and  Micropterus  dolomieui), is still  ex-
cellent  (Nashett,  1982),  all  recreational  uses  at
Saratoga Lake have been severely impaired (Mikol,
1982).  Public  opinion  concerning the  lake is  now
centered on the vegetation problem and controlling its
nuisance densities (Saratoga County,  1981).
  The objectives of this portion of the study were: (1)
to investigate the effectiveness of mechanical tech-
niques  commonly available to  control   nuisance
aquatic vegetation, and (2) to document the effects of
the major technique, harvesting, on the juvenile fish
populations associated with the nuisance vegetation.
Mechanical harvesting was the major physical control
method evaluated. The rate of reinfestation  of the
nuisance species after control measures had been in-
stituted was also determined.
METHODS

An area on the northwest side of the Lake, Manning's
Cove, was chosen as the study area (Fig. 2). Two tran-
sects were harvested in early June and early August of
1981 (Table 2). The areas were 48.0 m  x 3.0 m and 52.0
m x 3.0 m, respectfully. The transects were cut in the
1-2 m depths. Each was harvested by a barge-type
mechanical harvester  that  had a cutting bar  1.2 m
wide and 1.5 m long and a maximum cutting depth of
1.5 m. The wet weight of the vegetation harvested in
each  area was estimated  from the weights of sub-
samples of the plants cut in each transect. In August,
half of the north site was reharvested as  well  as
another transect of similar dimensions in an adjacent
area that had not been harvested in June. In 1982, ex-
perimental harvesting  was again conducted in  Mann-
ing's Cove in early June and in early August. The areas
harvested, however, were  significantly larger than

   Table 1.—Saratoga Lake morphometric characteristics.
 Surface area
 Total watershed area
 Effective watershed area
 Length
 Length of shoreline
 Maximum width
 Maximum depth
 Mean depth
 Volume
 Hydr. retention time
 Mean elevation
16.3 km*
632 km2
544 km*
7.2 km
37km
2.6km
29.2 m
7.7m
1.3 x 108 m3
0.4 yr
62m
                                             MACROPHYTE CONTROL


               those in 1981 (Table 2.) In August 1982, half of the June
               plot was reharvested and another adjacent plot was
               harvested for  the first  time.  Depths of the experi-
               mental plots were the same in 1982 as 1981.
                 The effect of harvesting on juvenile fish was deter-
               mined by counting, identifying, and determining the
               year class of fish found in the plants harvested by the
               barge in June and in August of both 1981 and 1982. By
               enclosing a control plot of similar plant populations
               and depth, an estimate of the standing  crop of fish
               was determined.  The enclosed area  was seined con-
               secutively, and the population was  estimated by an
               appropriate removal method (Everhart et al. 1973). The
               enclosure was set up just prior to each collection. A
               30.3m x 1.2m x 6.3mm mesh seine was used as the
               enclosure wall to enclose an area 7.6 m x  7.6 m. The
               seine used was 7.6 m x 1.2 m x  6.3 mm mesh. The
               enclosure was  seined until no fish were collected in
               two consecutive seinings.  Four  fish standing  crop
               estimates were made in June, July, and August 1981
               while three were made  in May, June, and  August of
               1982.
                 The change in species composition of the vegeta-
               tion during the growing season and biomass  levels
               were  determined at the harvested areas by randomly
               harvesting three  to four 0.25 m2  quadrat samples in
                        North Stain
                    ,  ,,~-,vTransect/1981
                    \c\
                      I  Mech Control

                 Sn  /studyArea
            *-= = =ZJj  lit  /         Statistics
            ^ff=^?  .*".' /      Cove Surface Area - 59.6 ha
            <'.*  /.  *?,' I /       Mean Depth   » 2 m
        Saratoga  *•=•   //
        Sailing Club   " ~
Figure 2.—Experimental vegetation control areas in Mann-
ing's Cove.
          Table 2.—Aquatic vegetation harvesting comparisons for 1981 and 1982 experimental control areas.

                                             Collection Dates
6/13/81

Area harvested (ha)
Harvesting time (hrs.)
Total wet weight removed (kg)
Total dry1 weight removed (kg)
Wet wt. (kg) removed/day @ 8 hrs.
Harvesting efficiency2 (ha/hr)
(ha/day) @ 8 hrs.
North
Site
0.016
—
115
20
7188
0.08
0.64
South
Site
0.015
—
86
15
5333
0.08
0.64
8/21/81
North
Site
0.008
—
58
10
7250
0.08
0.64
South
Site
0.015
—
47
8
3133
0.08
0.64
6/7/82

0.065-0.100
~2
273
46
2730-4200
0.041
0.33
6/8/82

0.100
~3
955
162
9550
0.033
0.26
8/13/82

0.175
~2.5
1600
272
9143
0.070
0.56
 1 Approximate; based on average of 17% dry weight/wet weight of oven dried samples.
 ' Approximate, includes unloading time
                                                  457

-------
 LAKE AND RESERVOIR MANAGEMENT
 adjacent, unharvested  areas  at approximately the
 same depths. The data from these  plots were ex-
 trapolated to 1.0  m2 samples.  Quadrat collections
 were made approximately every 19 days from the time
 of harvesting by scuba or  raking within a quadrat
 sampler. The dry weight biomass for individual  plant
 species collected was then determined by oven drying
 the samples for at  least 48 hours at 70°C.
   The reinfestation rate of vegetation after harvesting
 was determined by randomly sampling  four 0.25 m2
 plots in a control area. Reinfestation was determined
 by comparing total biomass (dry weight) and species
 composition  in the  harvested areas with  the
 unharvested areas.
 RESULTS

 In 1981, the entire surface area of the study area was
 affected by dense macrophyte growth. M. spicatum
 predominated; however, a band of P. crispus reached
 the surface just outside the milfoil-affected area in ap-
 proximately 4 meters of water. This band comprised
 the majority of the species harvested in the south stci-
 tion and was chosen for that reason. Approximate!/
 half the surface area of the north site was harvested
 again  in August 1981, despite the fact that the entire
 transect  had  returned  to  preharvesting  levels  of
 macrophyte densities. However,  milfoil at the north
 site did not regrow to the water surface in 1981 in ap-
 proximately half of the transect (1.5-2.0 meters deep.i.
   Harvesting  in 1982 was accomplished  on  similar
 dates as in 1981. Wind conditions on June 7 precluded
 completion of work until the following day, but data
 were recorded both days and evaluated separately.
 Areas  harvested  in 1982 were  approximately 6-10
 times larger than in 1981. These larger areas were cut
 because of difficulties in locating the 1981 transects
 under the ice and to better evaluate harvester efficien-
 cy and reinfestation. The south site of predominately
 pondweed had to be eliminated because of plant bio-
 mass  sampling difficulties.  Harvesting efficiencies;
 were lower in 1982. This could be due to the additional
 time required to more effectively remove plants within
 the plots or may be related to the actual reduced planl
 biomass levels noted lakewide in  1982.
  The  direct  effects  of  harvesting on juvenile fish
 populations in the study area were different in 1981
 and 1982. Only bluegill sunfish (Lepomis macrochirus}
juveniles (young-of-the-year) were removed by the har-
vester  in 1981. On  June 13,  1981, the harvester  re-
moved only 16 juvenile bluegills from the  north site
and 28 juvenile bluegills from the south site during
normal operation (Table  3). Similar low numbers of
young-of-the-year   bluegills  were  removed  during
harvesting on  Aug. 21, 1981, from the north (9) and
               south (17) sites. The 1981 harvesting operations would
               have removed  approximately 1,000 to 1,800 juvenile
               bluegills per hectare from either site. The fish stand-
               ing crop estimates from June and August 1981 were
               approximately  42,000 to 44,000 juvenile fish per hec-
               tare. The estimates are similar to those determined by
               Haller et al. (1980) for bluegills utilizing Hydrilla beds
               in Orange Lake, Fla., and for total fish standing crop
               estimates made by Wile and  Hitchin (1977) for an  On-
               tario Lake infested with Myriophyllum. Based on these
               numbers, 1981  harvesting removed approximately 2.4
               to 2.6 percent of the total fish standing  crop in this
               area. Bluegills comprised approximately 94 percent of
               the total standing crop in June and approximately 90
               percent in  August. Two distinct size ranges of this
               species were evident in June 1981 and three distinct
               size ranges were noted in the August 1981 collections.
               This is probably due to multiple spawnings in the area.
               Largemouth bass (M. salmoides) were the next most
               prevalent species collected.  Other  fish species col-
               lected during the  juvenile standing crop estimates in-
               cluded Esox lucius, L  gibbosus,  Perca  flavescens,
               and Pomoxis nigromaculatus.
                Fish removed by harvesting vegetation in June 1982
              were similar to those of June 1981, despite the fact
              that the area harvested  in 1982 was greater than the
              surface area harvested in 1981. Bluegill juveniles were
              again the  predominant species  collected  during
              standing crop estimates (95-99 percent) and removed
              by the harvester (93-95 percent). The number of fish
              collected extrapolates to a standing crop  estimate of
              approximately 39,000-46,000 juvenile fish  per hectare
              and  37,000-  44,000 bluegill  juveniles per hectare
              (Table 4). Other species collected in 1982 included L.
              gibbosus, P. flavescens, E. niger, and Umbra limi.  No
              largemouth  bass  were  collected  or removed  by
              harvesting in May  or June 1982 while they comprised 4
              percent of similar  collections in 1981. The  percentage
              of the standing crop of juvenile fish that were removed
              by harvesting in June 1982 was at most 2.8 percent.
                In August of  1982, significantly different numbers
              and species of juvenile fish  were collected during
              standing crop estimates as well as during harvesting
              operations. The  numbers of juvenile bluegills dropped
              to 77  percent of the standing crop estimates while
              yellow  perch   increased  to  almost  20  percent.
              Largemouth bass  were first collected  in 1982 during
              this period  and comprised 2 percent of the total
              estimate. Other species  collected included  L gib-
              bosus(<-\ percent), Ictalurus nebulosus (< 1 percent),
              and unidentified Notropis (
-------
to the area. This is supported by the fact that the
harvester  did not remove similarly high  numbers  of
yellow perch 13 days earlier.
  Significantly higher numbers of bluegills were col-
lected and harvested in August than in June 1982. The
August data relate to a  4.7 percent standing crop
removal rate by harvesting for bluegills and a 7.8 per-
cent removal rate of total juvenile fish. These data are
for a site  that had been previously harvested in June
1982. Similar evaluations of largemouth bass data  in-
dicate approximately 5 percent of that population was
removed by harvesting in  August (Table 5). Investiga-
tions by Haller et al. (1980) determined that as high  as
32 percent of the standing crop estimates were  af-
fected by harvesting operations in a HydnV/a-infested
Florida lake.  However, when  considering  just the
bluegill population examined  in that study, 14 percent
had  been  removed by harvesting and 1  percent of the
largemouth  bass population  was affected. Standing
crop estimates  made in  that  study were also  very
similar to those  determined here.
  Yearly comparisons of total mean biomass (g»m-2,
dry) for the major plant species collected are made in
Table 6. The data exhibit  relatively large ranges and
standard deviations of biomass for most species,  re-
flecting seasonal changes. However, average annual
biomass was lower in 1982 for all the  major species
collected  except coontail  (Ceratophyllum demersum).
These figures are comparable  to those reported for
other eutrophic waters (Wile and Hitchin, 1977; Kimbel
and  Carpenter, 1981). The total annual average bio-
mass estimates were 105.1 g»m-2 for 1981 and 41.8
g»m-2 for 1982 in the study area.
  Seasonal growth  patterns  and regrowth  after
harvesting of M.  spicatum and total plant biomass are
presented for  1981 and 1982 in  Figures 3 and 4.  In
1981, peak total  and milfoil biomass occurred in July
in the control transect and declined steadily through
September before leveling off in late October and
November. Significant differences  between the late
June control  transect  samples  and  the  harvested
                                    MACROPHYTE CONTROL


      samples were apparent. Average total and milfoil bio-
      mass estimates  in the  harvested  area  increased
      steadily through late August and peaked at that time.
      Regrowth  and preharvest  levels had occurred in less
      than 30 days. Milfoil and  total plant biomass in the
      area harvested in August declined to similar levels as
      the first harvest had produced. Biomass levels in the
      area harvested twice in 1981  remained significantly
      lower than the August harvest date until October 1981
      when regrowth  had brought this area up to levels not
      significantly different from the control transect or the
      area harvested once in June.
        Much less  variability in  the data was observed dur-
      ing the 1982 collections. Biomass levels were lower in
      May than  October of the  previous year, indicating a
      slight decrease in biomass over the winter.  Total
      levels remained fairly low the entire season with peak
      growth in  July, as in 1981. The mean  biomass levels
      were near zero 1 week after June harvesting and were
      still very  low 3 weeks later. In  1981, total biomass
      levels were approximately 100 g»m-2 2 weeks after
      harvesting in June. Regrowth  after harvesting occur-
      red similarly  to 1981. Biomass in the  harvested tran-
      sect was not  significantly  different than in the control
      area after approximately  30 days. By early August,
      milfoil biomass in the harvested  transect had sur-
      passed  the control transect  levels. In fact, they re-
      mained  higher than in the control area through Oc-
      tober.
        These differences were not always significant on
      every samoling date; however, the differences in the
      two areas  could be indicative of a trend. In both years
      regrowth of milfoil in the harvested areas  was to the
      water surface in 0.75-1.2 meter depths after approxi-
      mately 30  days. However,  in the 1.2-1.8 meter depths,
      milfoil reached the surface only after 45-48 days. Re-
      growth  in  the area harvested twice also  followed a
      similar pattern in 1982 as  in 1981. Sample sizes were
      smaller; however, a general increase in biomass (total
      and milfoil) was observed  from August (harvest date)
      to October 1982. Biomass levels again dropped to near
 Table 4.—Summary of direct effects of mechanical harvesting on juvenile fish populations in the experimental control area
                                                for 1982.
                                                                   Collection Dates
 Area harvested (ha)
 Total number fish removed
 Total number bluegills removed
 Number fish removed/ha
 Number bluegills removed/ha
 Fish standing crop estimate (no./ha)
 Bluegill standing crop estimate (no./ha)
 Percent standing crop (total) removed
 Percent standing crop (bluegills) removed
  6/7/82


0.065-0.100
    65
    61
 650-1,000
 610-938

       38,750-45,811

       36,339-43,745
         1.4-2.8

         1.4-2.8
 6/8/82


   0.100
 129

 123
1290

1230
 8/13/82

 Sitel

   0.050
 371
 172
7420

3440
 Site 2

   0.124
 276
 175
2226
1411
                       95,411
                       73,883
                  7.8

                  4.7
                  2.3

                  1.9
 Table 5.—Summary of direct effects of mechanical harvesting on juvenile largemouth bass (Micropterus salmoides), 8/13/82.


Area harvested (ha)
Total number fish removed
Number fish removed/ha
Fish standing crop estimate (no./ha)
Percent standing crop removed

Sitel
0.050
11
220
1,894
11.6
Collection Date
8/13/82'
Site 2
0.124
7
56
1,894
3.0

Combined
0.174
18
103
1,894
5.4
 ' Site 1 was harvested 6/8/82 and reharvested 8/13/82. Site 2 was harvested only on 8/13/82 No bass were harvested on June 7-8, 1982.
                                                  459

-------
 LAKE AND RESERVOIR MANAGEMENT
 zero during  the harvesting but increased to control
 levels in approximately 30 days after harvest.
   Annual total biomass statistics for the control and
 harvested  areas are shown  in Table 7 for 1981  and
 1982.The decrease in overall total biomass from 1981
 to 1982 is  apparent from these data. Average control
 area biomass levels in 1981 were 2.5 times higher than
 the  levels observed in 1982. The figures for the areas
 harvested  once show similar  differences, while tie
 area harvested  twice in  1981  was over 3 times trie
 levels for similar areas in 1982. A small sample size in
 1981 may  be the reason for this discrepancy. Of in-
 terest in the current study is the  ratio of the average
 annual biomass of the area harvested once, to the bio-
 mass of the control area. This ratio is the same for
 both years (0.73). These data seem to indicate that at
 least superficially, on  an annual basis,  the experi-
 mental harvesting  effected a similar plant biomass
 reduction in  1981 and 1982.


 DISCUSSION

The  effect of harvesting on  fish populations is  dif-
ficult  to assess.  The  direct effect of  removal by
mechanical  means  has been  studied only recently
(Wile and  Hitchin,  1977;  Haller et  al.  1980;  Swales,
1982). The difficulty is not in  assessing numbers and
species removed, but the short- and long-term effects
on the populations.  Indirect effects  of  harvesting
vegetation  on fish forage,  protective  cover,  and
predator-prey interactions are somewhat speculative,
but could be more important to the lake and pond eco-
system than is  immediately apparent, especially in
localized nursery areas. An attempt was made in th s
study to better define the direct effects of harvesting
on the overall juvenile fish populations in Saratoga
Lake.
   Haller et al. (1980) determined that over 30 percent
of the standing  crop of fish in a Florida lake weie
directly affected by mechanical harvesting. Juvenile
sportfish and small species of the 20 different types,
they   noted,  were  the most  adversely  affected.  At
Saratoga Lake, approximately 2-5 percent of the total
 juvenile population was affected in the experimental
 harvesting conducted for this study. The two studies
 are not comparable  in many ways (numbers of fish
 species,  plant  species, areas  harvested,  growing
 season, etc.); however, the numbers of bluegills and
 largemouth  bass estimated  in  standing crops by
 Haller et al. (1980) were very similar to those found
 here, as was the removal of mostly juvenile fish.
   Previous   studies  on   bass/bluegill  interactions
 (Carlander,  1969), the effects of vegetation densities
 on bass predation (Savino and Stein, 1982), and the in-
                                  t Harvest Date
                                 D—a Harvest I
                                   Harvest 2
                                 °--o Control
                      AUG     SEP

                         I98!
Figure 3.—1981  total  and milfoil biomass  (g»m-2, dry)
estimates for harvested and control transects (bracket =  1
std. dev.).
 Table 6.—Mean biomass (g»m -2, dry) comparisons of tho major plant species collected in 1981 and 1982 (N = sample size).
Species
Myriophyllum spicatum
Potamogeton crispus
Vallisneria americana
Ceratophyllum demersum
Nymphaea sp.
Eloda canadensis
Heteranthera dubia
Epiphytic algae
Unidentified
Total Ave. Biomass =
SD =
N =
Avg. Biomass
(S.D.)
271.8(205.8)
20.8(34.0)
38.6(61.8)
2.6(3.2)
35.4(45.8)
—
2.7(4.0)
3.8(6.2)
21.9(40.6)
105.1
(±167.4)
238
1981
Range
33.7-1075.5
0.2- 80.1
0.1- 313.3
0.1- 8.0
0.2- 159.0
	 	
0.1- 12.8
0.1- 23.3
D.2-147.9



N
72
5
57
5
18
	
18
21
27



Avg. Biomass
(S.D.)
103.8(82.5)
4.9(6.9)
15.2(51.0)
31.7(109.6)
2.3(1.1)
176(-)
2.3(4.5)
0.8(0.6)
3.3(3.2)
41.8
(±79.6)
280
1982
Range
0.8-386.8
0.2- 29.2
0.4-258.6
0.4-758.8
1.6- 4.0

0.2- 24.0
0.2- 1.6
0.2- 11.2



N
83
36
25
50
4

28
4
26



 Table 7.—Total mean annual biomass (g»m~2, dry) comparisons of harvested and onharvested transects in 1981 and 1982.
                      AIT"             1981                    ~                      	
                      Ave. Biomass                                 Avg. Biomass
 Transect                 (S.D.)            Range          N            '«n \
Control
Harvested once
Harvested twice
122.6 (192.4)
89.0 (146.6)
102.6 (87.5)
0.1-1075.5
0.1-675.9
0.2-247.2
109
117
12
49.5 (90.0)
36.1 (72.0)
31.9(58.6)
0.2-758.8
0.2-386.8
0.2-258.6
131
114
35
                                                 460

-------
                                                                                 MACROPHYTE CONTROL
formation collected for this study seem to indicate
that  at the removal rates observed, the largemouth
bass and bluegill populations would not be adversely
affected  by harvesting at Saratoga  Lake.  In  fact,
harvesting  could  possibly improve  Centrarchid
populations. This assumes that nursery areas are not
totally harvested and that other sportfish interactions
are not affected. This latter condition, especially for
species interations between yellow perch and walleye,
cannot be predicted from this information.  The tem-
porary effects of harvesting, including increased tur-
bidity, may also have a direct effect on fish feeding
and   behavior  (Gardner,  1981),  additionally   com-
plicating the impacts.
   Historically,  similar numbers of  aquatic   plant
species were observed in Saratoga Lake in the late
sixties (Dean,  1969). In fact, except for the exotic, M
spicatum, the species present were very similar to cur-
rent  conditions. However, densities  have increased
markedly since that time. Except for  the heavy den-
sities of M.  spicatum and P.  crispus,  a  relatively
diverse population of aquatic plants still exists in the
lake. On an annual basis, milfoil poses a more severe
problem since pondweed experiences a naturally oc-
curring die-off during July each year. Heavy milfoil
growth  persisted  through September  during  both
years studied.
   The 1981 average  biomass determinations  were
similar to data collected on other North American
lakes with severe milfoil infestations (Carpenter, 1980;
Adams and McCracken, 1974; Wile and Hitchin, 1977).
In 1982, the biomass levels observed in Saratoga Lake
were  much lower than  in  1981 and less variable
throughout the season.  Decreased  biomass  levels
were  also observed  for V. americana and  epiphytic
algae, while C. demersum increased somewhat. These
shifts were small compared to their relative percent
(10-20 percent) of the total biomass. Differences could
be due to several factors. Sampling efficiency  is one
possibility. However, climatic factors or the  influence
of the epiphytic algae are more likely explanations.
   Similar differences in P. crispus and M. spicatum
were observed in Lake Bomoseen in 1975 and 1977-79
(Garrison, 1979) and were attributed  to climatic fac-
tors.  Harvesting  reduced the total   plant  biomass
significantly for a relatively short period of time when
performed in June. A slightly larger reduction was ef-
fected when harvesting was performed in August. This
is consistent  with reported effects of  later season
harvesting by Wile and Hitchin (1977).  The effect may
involve a reduction of plant nutrients at a time when
levels are highest (Carpenter and Adams, 1977).
   Regrowth after harvesting approached preharvest
levels both years after approximately 30 days. In 1981,
biomass levels seemed  to  peak later  in harvested
areas compared to control areas in  Saratoga Lake.
This slight difference was almost entirely due to M.
spicatum  levels.  These  differences were  not
statistically significant, however.  Harvesting does not,
therefore, appear to have  increased  plant biomass
levels after regrowth following control.
   Additionally, multiple harvesting shows some pro-
mise for increasing the possibility of longer-term con-
trol.  Nichols  and Cottam (1972)  found that a  single
harvest  reduced growth by 50 percent, two harvests
reduced it 75 percent, and three almost eliminated the
vegetation  for 1 year.  Wile and Hitchin (1977) also
determined single harvests to be  least  effective, while
three harvests were most effective.
   In the current study, harvesting areas twice was
done on  a small  scale and did  not produce the ex-
                      JUL    AUG

                         I982
  I too
               t  Harvest Date
              D	a Harvest I
              •	• Harvest 2
              o- —o Control
   o
   5
         MAY     JUN     JUL    AUG     SEP     OCT
                         1982
Figure 4.—1982  total  and  milfoil  biomass (g»m-2,  dry)
estimates for harvested and control areas (bracket = 1 std.
dev.).

pected reductions; however, milfoil never grew back to
the water surface, despite normal densities in an area
harvested twice in  1981. This  could  be  due to the
depth above the sediment the plants were cut (Wile
and Hitchin, 1977). The more effective control observ-
ed in this study correlates with this hypothesis in that
the depths where better control  was observed were in
waters 1.5-2.0 meters deep. The  harvester should have
effectively removed  most of the leafy portions of the
plants at those  depths.  Light intensity has been
shown to be an important factor influencing the depth
and extent of macrophytes (Sheldon and Boylen, 1977;
Barko et al. 1982; Tobiessen and Snow, 1983). It is not
felt to be the limiting factor in this case, however,
because of the relatively shallow nature of the study
area.
SUMMARY AND CONCLUSIONS

Harvesting  aquatic  vegetation  at  Saratoga  Lake
removed juvenile fish during normal  operation. The
numbers removed relate to approximately 2-5 percent
of the total  juvenile fish standing crop estimates for
the study area. The predominant species affected was
bluegill sunfish (Lepomis macrochirus) and to a lesser
degree, largemouth bass (Micropterus salmoides) and
yellow perch (Perca flavescens). The dense vegetation
is an important nursery area for these and other sport
and forage fish  species.  Late  summer harvesting
seems to remove a slightly  higher percentage of the
juvenile  fish standing crop than spring harvesting.
August  1981  harvesting removed  only  0.2 percent
more fish than June 1981 harvesting. However, August
1982 harvesting removed 2.9 percent more juveniles
than were affected in June 1982.
   Very few  adult fish were affected. The open areas
produced by harvesting could have a beneficial effect
on sport fish and fishing as well  as a cropping effect
on the dense juvenile bluegill population. Indirect ef-
fects of harvesting on these  factors and predator-prey
interactions are difficult to  assess  and  require long-
term monitoring.
   Harvesting  is an  effective  means of temporarily
reducing plant biomass in dense  areas. The control of
milfoil had  a short-term effect when done  in June.
                                                 461

-------
 LAKE AND RESERVOIR MANAGEMENT
 Regrowth to preharvest levels took approximately 30
 days. Multiple harvests and later season harvests may
 have a more long-term effect on plant biomass reduc-
 tions.

 ACKNOWLEDGEMENTS: The  author  would like to  thank
 Sharon Hotalling  for preparation  of  the manuscript,  Jay
 Bloomfield for data evaluation, and  Denise Polsinelli for
 data collection, summarization, and graphics.
 REFERENCES

 Adams, M.S., and M.D. McCracken. 1974. Seasonal produc-
   tion of the Myriophyllum component of the littoral of Lake
   Wingra, Wis. J. Ecol. 62:457-65.
 Aiken, S.G., and R.R. Picard.  1980. The influence of su>
   strate on the growth and morphology of Myriophyllum ex-
   albescens  and Myriophyllum  spicatum.  Can   J  Bet
   58:1111-18.
 Barko, J.W. 1983. The growth of Myriophyllum spicatum L. in
   relation to selected characteristics of sediment and solu-
   tion. Aquat. Bot. 15:91-103.
 Barko, J.W. and R.M. Smart. 1983. Effects of organic matter
   additions to sediment on the growth of aquatic plants  J
   Ecol. 71:161-75.
 Barko, J.W., D.G. Hardin, and M.S. Matthews. 1982. Growth
   and morphology of submersed freshwater macrophytes in
   relation to light and temperature. Can. J. Bot. 60(6):877-87.
 Bedrosian, A.J.  1982. The evaluation  of macrophyte har-
   vesting at Lake Noquebay, Wis., for the period of May thru
   December 1981. Wis. Dep. Natl.  Resour.
 Bole, J.B., and J.R. Allen. 1978. Uptake of phosphorus from
   sediment by aquatic plants, Myriophyllum  spicatum and
   Hydrilla verticulata. Water Res. 12:353-8.
Carlander,  K.D.  1969.  Handbook of  Freshwater  Fishery
   Biology. Vol. 1. Iowa State Univ. Press, Ames.
Carpenter, S.R. 1980. The decline of Myriophyllum spicatum
   in a eutrophic Wisconsin lake. Can J. Bot. 58:527-35.
Carpenter, S.R.,  and M.S. Adams. 1977. The macrophyte
  tissue nutrient pool of a hardwater eutrophic lake: implica-
  tions for macrophyte harvesting. Aquat. Bot. 3:239-55.
Dean, H. 1969. Aquatic vegetation survey of Saratoga Lake
  Bur. Pestic.  Rep., N.Y. State  Dep.  Environ. Conserv,
  Albany.
Dunst, R.C., et al. 1974.  Survey of lake rehabilitation tech-
  niques and experiences. Tech. Bull. 75. Dep. Nat. Resour.,
  Madison, Wis.
Everhart, W.H.,  A.W. Eipper, and  W.D. Youngs.  1973.
  Principles  of  Fishery  Science. Comstock Publ. Ass
  Ithaca, N.Y.
 Gardner, M.B. 1981. Effects of turbidity on feeding rates and
   selectivity of bluegills. Trans. Am. Fish. Soc. 110:446-50.
 Garrison, V. 1979. Lake Bomoseen  water quality improve-
   ment project. Vermont Dep. Water Resour. Environ Enq
   Montpelier.

 Haller, W.T.,  J.V. Shireman,  and D.F.  Durant.  1980. Fish
   harvest resulting from  mechanical control of  Hydrilla
   Trans. Am.  Fish. Soc. 109:517-20.

 Kimbel, J.C., and S.R. Carpenter. 1981. Effects of mechanical
   harvesting on Myriophyllum spicatum L. regrowth and car-
   bohydrate allocation to roots  and shoots. Aquat. Bot.

 Mikol,  G.F. 1982.  New York  State lake  classification and
   inventory annual report.  Bur. Water Res., N.Y.  State Dep
   Environ. Conserv., Albany.

 Nashett, L 1982. Saratoga Lake fishery study. 1975-80 Com-
   preh. rep. (draft). Bur. Fish., Region 5. N.Y. State Dep En-
   viron. Conserv.,  Flay  Brook.

 Neel, J.K., S.A. Peterson, and W.L Smith. 1973. Weed harvest
   and lake nutrient dynamics. EPA-600/3-73-001. Off. Res.
   Monitor. U.S. Environ. Prot. Agency, Washington, D.C.
 Nichols, S., and G. Cottam.  1972. Harvesting as a  control for
   aquatic plants. Water Resour. Bull. 8:1205-10.

 Saratoga County.  1981. Saratoga  lake user survey, public
   participation workplan. Clean Lakes Progr. Phase I Study
   1980. U.S. Environ. Prot. Agency, Washington, D.C.
 Savino, J.F., and R.A. Stein. 1982. Predator-prey interaction
   between largemouth bass and bluegills as influenced by
   simulated, submersed vegetation.  Trans. Am.  Fish  Soc
   111(3):255-66.

 Sheldon, R.B., and C.W. Boylen. 1977. Maximum depth in-
   habited  by  aquatic  vascular  plants.  Am. Midi   Nat
   97(1):248-54.

 Storch, T.A., and  J.D.  Winter. 1983.  Investigation of the
   relationship  between aquatic weed growth,  fish  com-
   munities and weed management practices in Chautauqua
   Lake. Interim Rep. Environ.  Res.  Center, State  Univ. N Y
   College at Fredonia.

 Swales, S.  1982. Impacts of weed-cutting on fisheries: an
   experimental study in a small lowland river. Fish  Manaae
   13(4):125-35.                                       y  '

Tobiessen, P., and P. Snow. 1983. Temperature, light and sub-
   strate effects on the growth of Potamogeton crispus L. in
   Collins Lake. Can. J.  Bot.

Wile, I., and G.  Hitchin. 1977. An assessment of the practical
   and environmental implications of mechanical harvesting
   of aquatic vegetation in southern Chemung Lake. Ontario
   Ministry Environ.
                                                     462

-------
RESTRUCTURING LITTORAL  ZONES: A DIFFERENT APPROACH
TO AN OLD PROBLEM
SANDY ENGEL
Bureau of Research-Nevin Hatchery
Wisconsin Department of Natural Resources
Madison, Wisconsin


            ABSTRACT
            Dense carpets of submersed macrophytes in lakes restrict boating and swimming, limit move-
            ment of predator fishes, trap fish fry, and contribute to poor fishing. Eliminating plant beds,
            although good for boating and swimming, removes the plant cover, habitat diversity, and inverte-
            brates needed to support a sport fishery. Fiberglass screens and selective plant harvesting have
            proven useful in breaking up continuous stands of plants, reducing summer biomass and stored
            nutrients, and forming boating lanes. The screens kept areas free of vegetation all summer when
            placed on the lake bed in spring and rapidly removed  plants when spread over them in summer.
            Selective harvesting created islands of vegetation, gave predator fishes access to young fishes
            hiding among the remaining  plants, and opened bottom areas for the  spread of new plant
            species. Although intended to destroy vegetation, these and other methods can be used selec-
            tively and economically to rebuild littoral zones to benefit both people and lake biota.
 INTRODUCTION

 Littoral zones are complex heterogeneous environ-
 ments. The submergent macrophytes growing  near-
 shore provide high surface area and diverse habitats
 for colonizing algae and invertebrates. Many fishes
 congregate alongshore for food, shelter, and spawning
 (Keast and Marker, 1977). The structural heterogeneity
 partly created by  inshore plants and substrate ex-
 pands niche space, partitions food, and spatially se-
 gregates benthos and fishes (Mrachek, 1966; Werner
 et al. 1977). Many species respond to this habitat com-
 plexity by food or habitat specialization (Keast, 1978).
 Predatory  and  competitive  relationships  develop
 among littoral species. Littoral  zones are ultimately
 essential for growth and survival of many species in
 lakes.
   Some lake restoration techniques  destroy littoral
 zones. Herbicides, mechanical harvesting, dredging,
 and drawdown unselectively remove habitat, prey, and
 shelter. Structurally  complex habitats become rela-
 tively simple ones (Crowder and Cooper, 1979).  Com-
 munity development and established species interac-
 tions are  eliminated. Herbicides leave plants to de-
 compose and fuel new plant growth, whereas dredging
 and drawdown alter sediment texture and destroy pro-
 pagules of desirable plants on the lake bed (Dunst et
 al. 1974). New plant species can spread after  treat-
 ment and eventually monotypic stands replace a more
 varied macrophyte community.
   Bottom screening  and selective harvesting can be
 used to improve, rather than destroy littoral zones.
 Monotypic vegetation provides  fewer microhabitats
 and less  opportunity for species interactions  than
 stands of  mixed  species.  Breaking up  monotypic
 stands to allow  growth of other plants can foster a
 greater diversity of colonizing prey. Screens and har-
 vesting can create openings in  plant beds to permit
 predatory fishes and anglers greater access to the
 macrophytes. This  can  reduce overabundance  of
 young panfish trapped in macrophyte beds. By confin-
 ing  treatment to selected areas, a more open and
 varied littoral environment can be developed.
   Fiberglass screens can be used without a covering
 of sand or gravel, since they are  nearly three times as
dense as water and readily sink. They are meant to be
removed each year for cleaning and relocating. This
saves on material and prevents sediment accumula-
tion from encouraging attachment  of plants. Other
materials have been used to cover  lake bottoms for
macrophyte control, but cannot be easily removed and
soon become covered with silt and vegetation (Engel,
1984). The screens were first used on beds of Eurasian
watermilfoil (Myriophyllum spicatum L.) in Chautau-
qua Lake, N.Y. (Mayer, 1978). Screens were tested on
different plant species in Cox Hollow Lake, Wis., and
survival of benthos under them assessed.
  Many harvester operators remove as much vegeta-
tion as possible. This denudes large areas of shoreline
and requires disposing of a considerable mass of
vegetation.  A more selective  harvesting effort was
studied  on  Halverson Lake, Wis. Changes in  plant
growth,  macrophyte species composition, and fish
feeding  activity  before and  after  harvesting  were
evaluated. The work was part of a  broader study of
aquatic community  interactions of  submersed
macrophytes (Engel, in prep.).

STUDY AREAS

Cox Hollow and Halverson Lakes are located within 3
km of each other in Governor Dodge State Park, Iowa
County, Wis. They are situated in the steep unglaci-
ated terrain of Wisconsin's Driftless Area. The lakes
were built in 1958-59 by damming separate branches
of Mill Creek, a northerly flowing tributary of the Wis-
consin River. Both lakes receive flow  from  several
streams and empty separately into Twin Valley Lake.
Cox Hollow Lake is deeper and much larger than Hal-
verson Lake (Table 1). Motor trolling and gasoline en-
gines are prohibited on  both  lakes. Boating, swim-
ming, and fishing are popular on Cox Hollow  Lake.
Halverson Lake is more secluded, has an unimproved
boat access, and  is less used by park visitors.
  The lakes have moderately hard, alkaline waters.
They thermally stratify offshore in  summer and be-
come anoxic in  deep water. Blue-green algae bloom
each summer and lower the Secchi  disk visibility
                                                463

-------
 LAKE AND RESERVOIR MANAGEMENT


 from about 4 m in June to 0.6 m in August. Several
 shallow coves and adjoining areas on each lake be-
 come choked with submersed macrophytes.


 METHODS

 Polyvinyl-coated fiberglass screens, tested on a shal-
 low cove of Cox Hollow Lake, had a uniform mesh of
 1 mm2 (64 meshes/cm2). A screen, 2 by 18 m, was set
 on the lake bed, in water about 0.5 m deep, on May 9,
 1979, and May 9,1980. Another screen, 4 by 30 m, was
 set at a depth of 1.5  m on May 9,1980. In June, July,
 and August  1979, a 4.5-m  section of the shallcw
 screen was removed,  cleaned,  and  spread  over
 macrophytes in nearby areas. The screens were an-
 chored to the lake bed with bricks. All screens were
 removed by the fall of  1981, after 1 to 3 summers in
 the lake. Macrophyte  growth was measured on
 replicate 0.2-m2 samples collected by divers on,
 under, and  around the screens.  Areas around the
 screens  served as  untreated control  plots. The
   v
    O)

   to
       500-
       400-
       300-
   o
   5  200-
   i-
   z
   <  100-
              SHALLOW(0.5m)
 samples were cleaned, sorted to species, and dried
 in an oven at 105 C for 48 to 72 hours.
   A mechanical harvester was operated on  Halver-
 son Lake in June and July of 1980 and 1981. The cut
 plants were weighed and  removed to a park dump
 site away from the  lake's drainage basin.  Macro-
 phytes were sampled in the lake in June, July, and
 August of 1977 through  1982. About  75 samples of
 0.2-m2 were collected by divers  along 15  line  in-
 tercepts  randomly  located  around  the  shoreline.
 Sampled   plants  were  processed  similar to  Cox
 Hollow Lake samples. Plant cover and distribution
 were  mapped on  each sampling date by diving and
 taking aerial photographs.

 RESULTS  AND DISCUSSION

 Few macrophytes grew on or under screens  placed
on the lake bed in Spring (Fig. 1). The screens suppor-
ted about 19 percent of the mean  plant biomass  of
control plots. Many screens were free of vegetation;
two screens had a dry-weight biomass of 26 and 44
200n
                                                    100-
s
) 25

50
I
75
	 v-^
100
125
                                                                      50     75

                             1979                                       i960

                       DAYS AFTER SETTING SCREENS  IN  LAKE ON 9 MAY
                                 100    125
Figure 1.—Dry-weight biomass (mean ± 1 SE) of macrophytes on shallow and deep screens (dashed lines and open circles)
and surrounding control plots (solid lines and closed circles) in Cox Hollow Lake.

	Table 1.—Morphometry ;md water quality of study areas.
 Watershed area (ha)
 Lake area (ha)
 Shoreline length (km)
 Shoreline development
 Maximum depth (m)
 Mean depth (m)


 Chlorophyll a (^g/l)
 PH
 Secchi disk (m)
 Specific conductance (^mhos/cm)
 Total alkalinity (mg
                                                  Cox Hollow Lake
   1,600
     39
      5.3
      2.4
      8.8
      3.7


     60
      7.6
      0.7
    375
    200
                                                                    Morphometry
Water quality1
                    Halverson
                      Lake


                      250
                       4
                       1.4
                       2.0
                       7.1
                       2.6
                      50
                       8.1
                       0.6
                     275
                     115
 'Several samples were analyzed from the epilimnion in August 1978; chlorophylls were volume-weighted means unoorrected for pheophytm
                                               464

-------
                                                                                  MACROPHYTE CONTROL
 g/m2.  The vegetation on these  screens grew  on
 pockets of sediment that accumulated on the uneven
 surface. Filamentous green algae (Spirogyra) spread
 for a few weeks on some screens and then died. Few
 plants grew  back  after  removing screens;  plant
 growth reached 15 g/m2 on sites exposed for 1 to 3
 months. Screens  left in the lake  beyond  a  summer
 accumulated  sediment and became overgrown (126
 g/m2) the next year.
   In contrast, submersed macrophytes on shallow
 (0.5 m deep) control plots grew rapidly in May and
 June,  peaked in  July at  about  200-300 g/m2 (dry
 weight), and declined in August. Plants on deeper (1.5
 m) control plots reached about 100 g/m2  from June
 through August, partly because of light  restriction
 from blue-green algae. Leafy pondweed (Potamoge-
 ton foliosus)  comprised  about  80-87 percent  of
 submersed  macrophytes  on  control plots and
 dominated on the screens; coontail (Ceratophyllum
 demersum) and  curly-leaf pondweed  (P.  crispus)
 made up most of the remaining plants (Table  2).
 Coontail, a plant  without roots, did not gain an  ad-
 vantage  over rooted  species  in colonizing the
 screens.
   Macrophytes and benthos rapidly  deteriorated
 under screens placed on the water surface in sum-
 mer. The weight of the screens forced the  vegetation
 below the water surface; adding bricks compressed
 them against the  lake bed. After a few weeks green
 plants were scarce under the screens; a maximum
 dry-weight biomass of 18 g/m2 was found  under a
 shallow screen. Benthic macroinvertebrates,  sam-
 pled under and around the screens with  an Ekman
 dredge,  decreased  from June through  August  to
 about one third of the number found on control plots.
 Almost no benthos were found under screens after 1
 year. Poor water circulation and low dissolved oxy-
 gen accounted for the loss of macroinvertebrates.
   400n
   300
   200
   100-
z
                     WATER  DEPTH (m)

Figure 2.—Depth distribution of macrophyte biomass (dry-
weight) along a line transect in Halverson Lake during July
1979 (pre-harvest) and a week after the July harvests of 1980
and 1981.
Harvesting

Harvesting in  Halverson Lake  removed  about 50
percent of the macrophytes in 1980 and about 70 per-
cent in 1981 (Fig. 2). Vegetation was cut to a depth of
about 1.5 m. Macrophyte biomass, however, declined
beyond this  cutting limit, because of  low light
penetration. About 16,000 kg (wet weight) of vegeta-
tion were removed  after 32 hours of harvesting in
1980. Harvesting  for 42 hours in 1981 doubled  the
harvest size and left a much smaller standing crop of
plants in the lake. Plants grew rapidly after the June
harvests and nearly reached the preharvest densities
of 150-200 g/m2 in July. A short growing season con-
tributed to a slower recovery of vegetation after  the
July harvests.
  Harvesting left a  continuous carpet of vegetation
on the lake bottom, but sharply reduced the foliage in
mid-water and on the water surface (Fig. 3). The har-
vester had  difficulty operating in water less than 0.5
m deep, because of the depth of the barge and motor
prop. This left a rim of plants along shore, intersec-
ted  occasionally when  the  harvester  nosed  into
shore. Numerous offshore islands of vegetation  ap-
peared after  harvesting.  These consisted of uncut
plants and those that floated to the  water surface
shortly  after harvesting. The  vegetation  partially
                          8-10 JULY 1980
Figure 3.—Spread of submersed macrophytes on the water
surface and lake bed of Halverson Lake in July of 1979 and
1980. Depth contours and scale are in meters.
                  Table 2.—Dominant vascular plants smapled in Cox Hollow and Halverson LakesJ
       Text name
                                                                               Scientific Name
  Berchtold's pondweed
  Bushy pondweed
  Coontail
  Curly-leaf pondweed
  Elodea
  Leafy pondweed
  Sago pondweed
  Water milfoil
  Water stargrass
                 Potamogeton berchtoldii Fieber*
                 Najas flexilis (Willd.) Rost. & Schmt.
                 Ceratophyllum demersum L.
                 Potamogeton crispus L.
                 Elodea canadensis Michaux
                 Potamogeton foliosus Rafinesque
                 Potamogeton pectinatus L.
                 Myriophyllum exalbescens Fernald
                 Heteranthera dubia (Jacq.) MacMillan
  'Nomenclature and identification followed Voss (1972) for Potamogeton spp. and Fassett (1966) for the other plants.
  'Considered synonymous with P. pusillus L, following Voss (1972).
                                                 465

-------
 LAKE AND RESERVOIR MANAGEMENT
                                                                    VBERCHTOLD'S&SAGO PONDWEEDS

                                                                    flu
                                                                          .CURLY-LEAF PONDWEED
                                                                                 COONTAIL



                                                                        BUSHY PONDWEED



                                                                    WATER STARGRASS
                A
              1977
J J A
 1978
J J A  JUN  JUL A   JJN JUL A   J J A
 1979      1980        1981       1982
 Figure 4.-Relative frequency of the dominant macrophyte species of Halverson Lake in June, July, and August.
 grew together within a few weeks of harvesting, but
 remained  riddled with narrow channels left by the
 harvester.
   Macrophyte species composition changed drama-
 tically during the 6 years of study. Berchtold's pond-
 weed (P. berchtoldii) and sago pondweed (P. pectin a-
 tus) dominated on most sampling dates (Fig. 4). They
 comprised at least  90 percent of the total biomaas
 during the three summers before harvesting, but then
 gradually decreased to about 20 percent by August
 1982. Curly-leaf pondweed, coontail and bushy pond-
 weed (Najas flexills) increased in relative frequency
 and occurred in more sampling plots during  the 2
 harvest years. Some  other  species (Table 2) also
 spread, but contributed little to the total biomasis.
 The most  striking  change, however, occurred with
 water  stargrass (Heteranthera  dubia).  It  was  not
 found  before harvesting, yet spread into  nearly all
 sampling  plots and comprised  70 percent of  the
 August 1982  biomass. The  codominance  of  water
 stargrass and pondweeds continued through  1983,
 although samples were not collected.
   Bluegills (Lepomis macrochirus) continued to feed
 avidly  on aquatic insects in the macrophyte beds
 after harvesting, but  also ate  macroinvertebrates
 that were dislodged by the harvester and settled to
 the  bottom  (Engel, in  prep.).  Largemouth  bass
 (Micropterus salmoides), some over 400 mm  long,
 used channels left  in  the macrophyte beds by the
 harvester to search for prey. Their diet of fish  in-
 creased after harvesting.


 CONCLUSIONS

Screening  and harvesting produced a  more  open
macrophyte community. The  screens worked well in
spring and summer,  but needed to be cleaned annu-
ally  for continued macrophyte  control.  Harvesting
left channels that became cruising lanes for fishes,
spread plant fragments of some species,  and ex-
posed other plants growing underneath taller pond-
                                   weeds. These openings  were partly filled  by new
                                   species, creating a more varied composition. Screen-
                                   ing and selective harvesting, consequently,  helped
                                   create a more diverse and desirable littoral environ-
                                   ment. These and other techniques need to be further
                                   explored for restructuring littoral zones.


                                   REFERENCES

                                   Crowder, L.B., and W.E. Cooper. 1979. The effects of macro-
                                    phyte  control on the feeding efficiency and growth of
                                    sunfishes: evidence from pond studies. Pages 251-268 in
                                    J.E. Breck, R.T. Prentki and O.L Loucks, eds.  Aquatic
                                    Plants, Lake Management and Ecosystem Consequen-
                                    ces of Lake Harvesting. Inst. Environ. Stud. Univ. Wiscon-
                                    sin, Madison.

                                  Dunst, R.C., et al.  1974. Survey of lake rehabilitation tech-
                                    niques and experiences. Tech. Bull. 75. Wis. Dep. Nat.
                                    Resour.

                                  Engel, S. 1984. Evaluating stationary blankets and remov-
                                    able screens for macrophyte control in lakes. J. Aquat.
                                    Plant Manage. 22(1): in press.

                                  	In prep. Aquatic community interactions of sub-
                                    mersed macrophytes. Tech. Bull. Wis.  Dep. Nat. Resour.

                                  Fassett, N.C. 1966. A Manual of Aquatic  Plants. Univ. Wis-
                                    consin Press, Madison.

                                  Keast, A. 1978. Trophic and spatial interrelationships in the
                                    fish species of an Ontario  temperate lake. Environ.  Biol.
                                    Fish. 3:7-31.

                                  Keast, A., and J. Harker. 1977. Fish distribution and benthic
                                    invertebrate biomass relative to depth in an Ontario lake .
                                    Environ. Bio. Fish. 2:235-40.

                                  Mayer, J.R. 1978. Aquatic weed management  by  benthic
                                    semi-barriers. J. Aquat. Plant Manage. 16:31-3.

                                  Mrachek, R.J. 1966. Macroscopic invertebrates on the higher
                                    plants at Clear Lake, Iowa.  Iowa Acad. Sci. 73:168-77.

                                  Voss, E.G. 1972. Michigan  flora. Part  I. Gymnosperms and
                                    monocots. Cranbrook Inst. Sci. Bull. 55.

                                  Werner, E.E. et al. 1977. Habitat partitioning in a freshwater
                                    fish  community. J. Fish. Res. Board Can. 34:360-70.
                                                 466

-------
AN EVALUATION OF  PIGMENTED  NYLON  FILM  FOR USE IN
AQUATIC PLANT MANAGEMENT
MICHAEL A. PERKINS
Department of Civil  Engineering
University of Washington
Seattle, Washington
            ABSTRACT

            Experimental applications of pigmented nylon film were conducted in the Seattle metropolitan
            area in order to evaluate gas venting characteristics and instalation procedures. The material
            was highly effective in controlling aquatic plant growth, causing death and decomposition of
            covered plant materials in 30 to 35 days. Gas entrappment, ballooning, and lifting was observed
            with unvented material but this was overcome by placing venting slits in the material. A vent slit
            design of 12/meter2 gave the best result. Installation of the film is straightforward and simple.
            Difficulties encountered with dense plant growth and soft organic sediments are described.
INTRODUCTION

The growth of aquatic vascular plants in freshwater
lakes often reaches proportions that impair recrea-
tional and aesthetic quality. The nuisance characteris-
tics of such growth are well documented and quite
familiar to most lakeshore residents and  recreational
users.
   Many different attempts have been made to control
or manage aquatic vegetation, ranging from the use of
chemical  herbicides to  physically  manipulating the
aquatic habitat. The choice of management technique
for use in a given lake system  would depend upon a
variety of site-specific factors.  Included among such
factors would be considerations relating to: (1) the
type of problem plant and its areal extent; (2) the phys-
ical-chemical  characteristics of the water body  and
impacted area; (3) proposed and existing uses of the
water body; (4) the attitudes of user groups toward the
various options;  and (5)  the desired efficacy and re-
source base available.
   It is becoming  increasingly evident that no single
aquatic plant  management technique is generally ap-
plicable to all situations. All management techniques
have their drawbacks, expressed either  in terms of
secondary effects upon nontarget components of the
system being treated, limited duration of benefit, or
excessive costs of materials application. Clearly, the
choice of technique for a given situation would be that
which  maximizes both short- and long-term  benefits
and minimizes costs and secondary effects.
   The  use of benthic barriers to blanket or shade
aquatic plants has been reviewed at some length by
Nichols (1974), Armour et al. (1979), and Cooke (1980).
The concept is quite simple. Depending upon the ma-
terial  used, plant stress, death, and subsequent de-
composition are induced either  by reducing photosyn-
thetically active  radiation or by physically compres-
sing plant biomass into  contact with  the sediments,
hence  restricting  growth and enhancing  decomposi-
tion (Boston and Perkins, 1982).
   A variety of materials have been used as  benthic
barriers to prevent or retard  aquatic  plant  growth.
These  include sand-gravel, polyethylene, polypropy-
lene, synthetic rubber, fiberglass screen, and burlap
(Cooke, 1980;  Nichols and Shaw, 1983). The degree of
success associated with the use of these materials
has been varied and directly related to the nature of
the material being used. The major problems that have
detracted  from using benthic barriers as a viable
management  option  have  been  cost  of  materials,
labor intensive application process, lifting and tearing
of installed barriers by trapped decomposition gases,
current movements or wave action, and regrowth of
aquatic plants on or through installed barriers either
because of sedimentation or the nature of the barrier
material itself.
  Costs of materials and application frequently limit
use.  Materials such as fiberglass  screen, although
highly effective and relatively easy to install,  become
cost  prohibitive for large scale application (Perkins,
1980).  Bouyant  materials  such   as  polyethylene
sheeting  are  exceedingly  difficult to install  and
necessitate substantial anchoring.  Polyethylene has
also been shown to lack durability, easily tearing dur-
ing the installation  process,  which increases its
susceptibility to  lifting and  further  tearing, reducing
its effectiveness (Armour et  al. 1979).
   Provision for gas venting  is an obvious requirement
for benthic barrier materials.  However,  the size and
spacing of the vents is of  significance. Ideally, the
vent design should allow for  gas  transfer  and yet
retard plant growth through the barrier. Loosely woven
materials such as fiberglass screen and burlap offer
excellent gas venting properties but  the aperture sizes
are such that finer leaved macrophytes may penetrate
the  material  and  establish  substantial  biomass
(Perkins, unpubl.; Cooke,  pers. comm.). This feature
may quickly offset initial biomass reductions and may
promote succession to other undesirable plant types
(e.g., Potamogeton pectinatus).
   Woven mesh materials and sedimentation  on  other
types of barrier  material provide substrate for  plant
rooting on top of the barrier. Resultant plant growth
may be  particularly pronounced   under conditions
where the problem plant may be one  that depends
upon vegetative fragmentation as a  primary mode for
propagation   and dispersal  (e.g.,  Myriophyllum
spicatum). Plant fragments  readily settle on and root
through fiberglass screen  and comparable observa-
                                                467

-------
 LAKE AND RESERVOIR MANAGEMENT
 tions have  been reported for burlap  (Perkins et  al.
 1980; Cooke, pers. comm.).
   A general overview of past experiences with benthic
 barriers would suggest that the ideal benthic covering
 for  aquatic plant management  would possess the
 following attributes:
   1. The material should have an opacity sufficient to
 block photosynthetically active radiation;
   2. The material  should be durable  to physical-
 chemical-biological degradation during and  subse-
 quent to installation;
   3. The material should be negatively bouyant so as
 to facilitate the installation  process and impede lift-
 ing;
   4. The material should be provided with a gas vent-
 ing  system  that would allow for gas transfer  and  in-
 hibit plant growth through the vents;
   5. The material should offer a smooth upper surface
 in order to inhibit fragment rooting;
   6. The material should be competitively priced.
   Recent  examination and experimentation  with
 black pigmented nylon film (DARTEK, DuPont Canada)
 would suggest that it possesses these characteristics
 and that many of the difficulties associated with other
 barrier materials would be circumvented. The material
 is composed of nylon 6,6 containing 2 percent  carbon
 black and is commercially available in 2 mil thickness
 on rolls  of  2.5x30.5  meters. Experimental applica-
 tions of the film were conducted in the metropolitan
 Seattle area (Green Lake  and Lake Washington)  to
 evaluate installation procedures and  to assess the
 gas venting characteristics of different slit designs.
 The results of these applications are  the subject  of
 this report.
METHODS AND MATERIALS

The gas venting characteristics of three different slit
designs were evaluated on small test panels of film
that had been secured to 2 x 2 meter PVC frames.
Four panels were constructed: an unvented control a
diagonal slit pattern  (12 vents/m2), and  two  cross-
hatch slit patterns (12 and 5 vents/m2). We hypothe-
sized that  gases produced as a  result  of bentNc
respiration and plant  decomposition (primarily CO2)
would  accumulate beneath the  installed pane s.
Those  panels which  ventilated most poorly  would
show higher concentrations  of dissolved inorganic
carbon (DIG) and higher rates of DIG accumulation.
The null hypothesis was one of no difference in DIG
concentration and  rate  of increase between the
various test panels.
  The panels were installed in Portage Bay along the
Lake Washington ship canal between Union Bay (Lake
Washington outlet) and  Lake Union. The site  was
located along the northern shoreline of the bay adja-
cent to the  University of Washington (Fig. 1).  Water
depth within the site ranged from  1.5 to 2.0 meters.
The plant  community consisted   predominantly of
Elodea canadensis and  Myriophyllum spicatum L.
Plant distribution was relatively uniform over the test
area and quite dense. Biomass samples, taken shortly
after panel  installation, averaged 166 grams dry
weight/meter2 with E.  canadensis accounting for ap-
proximately 88  percent  of the  biomass. Bottom
substrates consisted of approximately 30 cm of floc-
culant organic muck over consolidated sand-gravel.
The panels were installed by divers in July and were
secured to the bottom using 1 meter lengths of 0.5 cm
cold roll steel rod which had been bent to form a hook.
One rod was placed at each corner of the panels.
   Observation  and sampling  were  conducted  on
 seven dates over a 35-day period following installa-
 tion. Chemical sampling consisted of analysis for DIG
 and pH beneath the panels and under ambient condi-
 tions. Samples for DIG analysis were taken using an
 evacuated blood sampling tube (Vacutainer) and syr-
 inge which was inserted through the film. Vacutainers
 were used to  avoid sample  contact  with the atmo-
 sphere. DIG  concentration was  determined by injec-
 tion  of sample into 6N  H2SO4 contained within a
 nitrogen gas sparging system with measurement of
 evolved  CO2  by  infrared  gas  analysis. Ambient
 samples were  taken at the sediment surface in an
 area adjacent  to  the  panels. Three replicates were
 taken for each determination. All  samples were col-
 lected by diver.
   Samples for pH determination were taken with a 50
 ml syringe equipped with a 10 cm cannula. Samples
 were transferred to 60 ml polyethylene bottles  and
 taken to the  laboratory for measurement.  Water
 temperature  at the time of sampling was also deter-
 mined.
   Evaluation of field application procedures entailed
 a description of problems encountered and recom-
 mendations  for facilitating the  installation process.
 The evaluation  was based  upon two field applications,
 Green Lake and  Lake Washington, which were con-
 sidered to approximate typical  commercial applica-
 tions.
   The Green Lake installation was conducted in May.
 A continuous roll of film (2.5x30.5 m) was run adja-
 cent to a public dock in waters ranging in depth for 1
 to 3  meters. Coverage was over a very dense plant
 community  consisting  of Elodea canadensis  and
 Myriophyllum spicatum, Elodea  dominating. Bottom
 substrate over the majority of the 30.5 m strip con-
 sisted of unconsolidated  organic  muck of undeter-
 mined depth  but in excess of  1.5 meters.
   The Lake Washington installation was conducted at
 a private residence along the western shoreline of the
Figure 1.—Location of Lake Washington and Portage Bay
test applications of pigmented nylon film.
                                                 468

-------
                                                                                 MACROPHYTE CONTROL
lake. Panels were placed according to the desires of
the owner who used the area for diving and swimming.
The installation consisted of four 7.6x2.5 m panels
placed in overlapping pairs (30 cm overlap) at the end
of a 30 m dock (Fig. 1). Water depth was approximately
3 meters with a bottom substrate consisting of well
consolidated sand-gravel with very little organic ac-
cumulation. The  plant community was  moderately
dense  and  consisted  of  Potamogeton  richardsoni,
Myriophyllum  spicatum, Potamogeton pectinatus  L,
and Najas flexilis. The pondweeds and milfoil extend-
ed to the surface within the site. The application was
conducted  in August.
RESULTS AND DISCUSSION

Gas Accumulation and Venting

Sampling for DIG analysis commenced 4 days after in-
stallation of the panels and continued up to 25 days
post-installation. The results of the DIG  analysis are
shown in Table 1. With the exception of the control
panel,  DIG concentrations were significantly greater
than the ambient samples on all sampling dates. Con-
centrations showed a consistent increase under each
panel up to 12 or 14 days post-installation and then
declined through days 14 to 19. Twenty-five days after
installation, DIG concentrations again began  to  in-
crease.
  While all of the panels demonstrated some degree
of ballooning resulting from presence of covered plant
materials, ballooning  and subsequent lifting of  the
control  and  diagonal  slit panels  was particularly
severe. Seven days after installation of the  panels the
control and diagonal had lifted by 20 to 30 cm and had
to be resecured. Gas entrappment by day 12 was such
that the control and diagonal panels had lifted by 45
cm. On day 14, the  lift on the control panel had been
sufficient to completely remove one of the corner
stakes and the panel was approximately 60 cm off of
the bottom. The diagonal panel had again lifted by 45
cm. Despite efforts to resecure the panels, the same
conditions were present on days 19 and 25. Thirty-five
days post-installation,  lift on the  control  panel had
been sufficient to break loose the  film from its PVC
frame. The diagonal panel was again 45 to 50 cm off
the bottom. Ballooning was also evident on the cross-
hatch  design panels but did not lift them. The maxi-
mum observed ballooning was  approximately 30 cm
for the 5 vent/m2  panel and 20 cm  for the  12 vent/m2
panel.
   Because  of the probability  of water   exchange
through the sides of the panel, lifting obviously con-
stituted a problem for the  evaluation of DIG increase.
The observed lifting, however, did  reflect the relative
lack of venting by the diagonal slit design.
  We assumed that the initial increase in DIG concen-
tration  resulted from the entrappment and dissolution
of CO2 gas produced by benthic respiration. Provided
that  the only avenue for  exchange was through the
slits of the panels, the rate of increase in DIG should
reflect the extent to which the various slit designs ex-
changed accumulated CO2 with the overlaying water
column. The greater the  rate of  accumulation, the
slower the rate of exchange or venting.
  To evaluate exchange rate,  the change in DIG con-
centration from day 4 to  each subsequent sampling
date was  calculated. Change in  DIG concentration
against time (days  after installation)  is  shown  in
Figure  2. The least squares "line of best fit" was then
calculated over the period from day 4 to the maximum
change in DIG and the computed slope  was taken as
the rate of DIG accumulation  in mg C/liter/day. Com-
puted slopes are also shown  in Figure 2. The rate of
venting would be inversely related to the rate of ac-
cumulation.
  For  reasons outlined previously, the control and
diagonal panels must be excluded from this analysis;
                         Slope = 034
        DIAGONAL
                      10       15      20      25
                        Slope- 0 SO
   o
   o>
   E
   o
   o
         HAT C H- 12
                            Slope -
         HATCH-5
                         Slope - 0
                days
                       10
                       after
 15      20

i nst allot ion
 Figure 2.—Temporal variation in change in DIG concentra-
 tion beneath nylon film test panels. Slopes represent the rate
 of DIG increase from Day 4 after installation to the period of
 maximum change in DIG.
   Table 1.—Summary of temporal variation in DIG concentration between nylon film test panels and ambient conditions.
                         Values are means ±  one standard deviation for three replicates.
Date
7/26
7/29
8/03
8/05
8/10
8/16
Days
4
7
12
14
19
25

Ambient
7.8 ±0.15
8.4 ± 0.22
9.3 ±0.14
9.3 + 0.16
8.8 ±0.13
9.4 ±0.11

Control
8.1+0.25
9.9 ±0.14
10.9 ±0.33
9.8 + 0.13
9.0 ± 0.09
9.8 ± 0.35
mg carbon/liter
Diag.
9.4 + 0.14
10.4 + 0.77
11.0 + 0.20
10.0 + 0.15
10.5 + 0.23
12.0 ±0.35

Hach 12
8.6 ±0.33
9.6 ± 0.48
11.5 + 0.10
11.8±0.10
10.3 + 0.18
11.1 ±0.58

Hach 5
8.8 + 0.42
11.1 ±0.51
12.3 + 0.53
9.8 + 0.09
10.4 ±0.65
12.8 + 0.78
                                                 469

-------
 LAKE AND RESERVOIR MANAGEMENT
 the only valid comparison would be between the two
 cross-hatch designs. The 5 slit/m2 panel showed a rate
 of DIG accumulation significantly greater than that oc-
 curring with the 12  slit/m2 panel. While neither panel
 had lifted because of gas entrappment, the 12 slit/m2
 design exhibited greater venting characteristics.
    Plant growth through the panels was observed 7
 days  post-installation. After 12  days, the growth
 through consisted of both elodea and milfoil and ex-
 tended approximately 15 cm above the diagonal panel
 and 30 cm above the cross-hatch panels. Nineteen
 days  post-installation, all plant growth through Ihe
 slits had disappeared with the exception of a single 30
 cm shoot of milfoil extending through the cross-hatch
 12 panel.  Presumably, those plants that had grown
 through the panels had  been  grazed by herbivores
 (most likely crayfish which are common in  the area
 and were observed on top of the panels on at least one
 occasion). While the emergence of plants through the
 slits must be considered a negative aspect, the actual
 number of slits affected was only two to four per panel
 (approximately  6 percent). In general,  plant growth
 through the slits was more prevalent with the cross-
 hatch design and greater with  the 12-slit panel than
 with the 5-slit panel.
   Removal of the panels 35 days  post-installation
 revealed  almost complete  decomposition  of  the
 covered plant material. Remaining plant tissue con-
 sisted of  only a few yellowing,  leafless  stems. There
 was very little difference between the four plots.
 Installation Procedures

 DuPont Canada packages the film wound on AEJS
 cores equipped with caps on either end. For installa-
 tion, the core is filled with sand or gravel and the roll is
 allowed to soak for 12 to 24 hours. Presoaking the film
 makes it more pliable,  more resistant to tearing, and
 easier to handle.
   The Green  Lake installation proved  to  be infor-
 mative  relative  to applications  over  dense plant
 growth on unconsolidated organic sediments. Caution
 must be observed when adding weighting materials to
 the core. If the core is too heavy, the roll may sink into
 the sediment and retrieval can be difficult. The dense
 growth  of elodea tended to lift the panel as it was
 unrolled, resulting in considerable panel shifting and
 pulling loose of the stakes used to secure the leading
 edge. The problem was somewhat overcome by unroll-
 ing shorter lengths of film (1.5 m) and staking at these
 intervals.
   The staking material used for the Green Lake in-
 stallation was 1.3 cm concrete reinforcing bar (0.5 inch
 rebar) which had been cut on the diagonal into 30 to 90
 cm lengths and bent at one end to form a hook. The
 rebar proved to be too  heavy for the film. Because af
 the lifting and shifting of the panel as it was unrolled,
 the leading edge of the  film constartly tore loose from
 the  stakes  even though  they  had been inserted
 through double folds of material. The material is quite
 durable but it will tear at the stake locations  if sub-
 jected to stress. Using thinner staking material (0.5 cm
 steel rod) lessened the  extent of tearing.
  The manufacturing process leaves a thicker bead of
 nylon along the longer edges of the roll. This bead w&s
much more resistant to tearing and formed a  conve-
 nient means for securing the side stakes. Because of
the lifting described previously and the relatively short
 lengths of some stakes, they consistently pulled loose
from the soft bottom. This  problem was overcome by
 again unrolling  shorter lengths of film  and  using
 longer and thinner stakes.
   The net result of the Green Lake application was a
 very uneven and folded installation that required a
 return visit  for  straightening and  restaking. Four
 weeks after the installation, the film was lying flat on
 the bottom although some folding was still evident.
   The installation was done by one scuba diver and
 one  surface  swimmer.  The  flocculant  sediments
 limited visibility and  the  process required  approx-
 imately 1.5 hours of diving time on each visit.
   The Lake Washington installation was done with
 comparable ease. The film was first unrolled and cut
 into the four 7.6  m lengths. These were then rerolled
 onto the core which was then placed into the lake for a
 presoak time of 2 hours. The core was filled with sand
 to the point that  it sank just  below the surface.
   The roll was then transferred to the site and the
 leading edge was secured with three 1 m lengths of
 steel rod. The panel was unrolled in line with the edge
 of the dock, staking at 2 m intervals along the edges.
 The roll was then reversed and the unrolling process
 repeated  back along the edge of the first panel, over-
 lapping by 30 cm. Stakes  from the first panel were
 removed and reinserted through the overlap. The roll
 was then  moved to the site for the next pair of panels,
 repeating the unrolling staking process.
   The installation was again  done by one scuba diver
 and one surface swimmer. Even though the plant stem
 lengths were longer, biomass was considerably less
 than that  at Green Lake and  the panels lay relatively
 smooth and flat with very little ballooning. Tearing of
 the film at the stake locations was again evident but
 not as severe as with the Green Lake application. The
 total diving time for installation was 35 to 40 minutes.
   Two weeks after installation, the panels were lying
 flat on the bottom. There was very little plant growth
 through the venting slits but some plant shoots had
 emerged in the panel overlap areas.
CONCLUSIONS

Experimental applications  of  pigmented nylon film
proved to be highly effective  in removing  nuisance
aquatic plant growth from  the water column and in
causing  the death  and decomposition of treated
plants after 30 to 35 days of coverage.
  The requirement  for  venting  was amply demon-
strated by our experience with the control test panel
which had lifted by 30 cm in as little as 7 days follow-
ing installation.  Comparing the  three  slit  designs
tested, the  diagonal  design  with  12 vents/m2 was
clearly inferior to either of the cross-hatch designs.
Both cross-hatch designs remained secure to the bot-
tom although the 5  vent/m2 panel did exhibit slightly
greater ballooning than the 12 vent/m2 panel. On the
basis of DIG accumulation,  the 12 vent/m2 design did
exhibit a significantly  greater gas exchange rate
which would imply superior venting characteristics. It
is significant that neither of the cross-hatch panels
had lifted. The increased number of vents appeared to
be more susceptible to plant growth through the slits
and it may be that  an intermediate number of vents
between  5  and  12  /m2  would be a more  effective
design.
  Despite the problems encountered in  Green Lake,
the installation process was quite easy. Staking of the
film is a definite requirement but the choice of staking
materials may be critical. Stake length depends upon
the substrate condition,  longer stakes being required
                                                 470

-------
                                                                                      MACROPHYTE CONTROL
for softer bottom sediments. The  0.5 cm steel rod
reduced  the extent of  film  tearing. Increasing the
width of the bead along the edges of the film and stak-
ing through the bead may help  prevent tearing but
would not appear to be absolutely  necessary.  Place-
ment of  the film over dense plant  growth can be a
problem  but this can be reduced  by unrolling shorter
lengths of film and staking the sides at shorter inter-
vals. A local commercial applicator cuts extant plant
growth prior to film application but we feel that this is
unnecessary. The film compresses to the bottom in a
few  weeks and can be  easily manipulated if neces-
sary.
   Pigmented nylon film appears to be an ideal bottom
covering  material.   It  is opaque,  fairly  durable,
negatively bouyant, and  easy to  manipulate  under
water. The  film is currently supplied with the cross-
hatch 12 slit/m2 design which is quite adequate for
gas venting. Costs for the film are  currently  60 cents
per meter2 (5.5 cents/ft2) which make it quite attractive
compared  to  other barrier materials.  Approximately
5,600 square meters of film have been commercially
installed within the Seattle area.

ACKNOWLEDGEMENTS: The author wishes to thank Roland
Sanford and  Don Althauser for their assistance in the con-
duct of this study. The work was supported under contract to
DuPont Canada and we wish to thank Ted Pattenden for his
assistance.
REFERENCES

Armour, G.D.,  D. Brown, and K. Marsden. 1979. Studies on
  aquatic macrophytes, Part XV, An evaluation of bottom
  barriers for control of Eurasian watermilfoil in British Col-
  umbia. Water Invest. Br., Ministry Environ., Province of Br.
  Columbia.
Boston, H.L., and M.A. Perkins. 1982. Water column impacts
  of macrophyte decomposition beneath fiberglass screens.
  Aquat. Bot.  14:15-28.
Cooke, G.D. 1980. Covering bottom sediments as  a lake
  restoration technique. Water Res. Bull. 16: 921-6.
	. Pers. comm. Dep. Biolog. Sci.,  Kent State Univ.,
  Kent, Ohio.
Nichols, S.A. 1974. Mechanical and habitat manipulation for
  aquatic plant mangement. A review of techniques. Tech.
  Bull. 77. Dep. Nat. Resour. Madison, Wis.
Nichols, S.A.,  and B.H. Shaw. 1983. Review of management
  tactics for integrated aquatic weed management of Eura-
  sian watermilfoil (Myriophyllum spicatum), curlyleaf pond-
  weed (Potamogeton crispus) and elodea (Elodea canaden-
  sis). Pages  181-192 In Lake Restoration,  Protection and
  Management. Proc. 2nd Ann. Conf. N. Am. Lake Manage.
  Soc.  EPA  440/5-83-001.  U.S.  Environ.  Prot. Agency,
  Washington, D.C.
Perkins, M.A. 1980. Managing aquatic plants with fiberglass
  screen. Pages 245-248 in Restoration of Lakes and Inland
  Waters. Proc. Int. Symp.  EPA 440/5-81-010. U.S. Environ.
  Prot. Agency, Washington, D.C.
	Unpubl. Univ. Washington, Seattle.
Perkins, M.A.,  H.L Boston, and E.F. Curren. 1980. The use of
  fiberglass screen for the control of Eurasian watermilfoil.
  J. Aquat.  Plant Manage. 18: 13-19.
                                                    471

-------
                                           Role  of   Local   Lake
    Organizations  &  Public  Education
VOLUNTEER LAKE MONITORING: CITIZEN ACTION TO IMPROVE LAKES
DONNA F. SEFTON
JOHN  R. LITTLE
JILL A. HARDIN
J. WILLIAM  HAMMEL
Illinois  Environmental Protection Agency
Springfield, Illinois
           ABSTRACT

           Citizen activists participate year after year in the Illinois Volunteer Lake Monitoring Program—providing
           their own boating equipment and collecting data at least twice monthly from May through October.
           Over two thirds of the volunteers who started in 1981 continue to be active. The program was initiated
           by the Illinois EPA in 1981 to help citizens make more informed decisions about lakes' use, protec-
           tion, and management. Citizens are trained to measure Secchi disk transparency and total depth and
           record field observations in a systematic manner at designated sites. Secchi disks, special data reporting
           forms, and postage paid envelopes are provided by the Agency. Morphological data and assessment
           information are also collected for the lake and watershed. The sampling data are computerized and
           a statewide summary report is prepared. As resources permit, individual lake reports are also prepared
           which incorporate physiochemical data obtained under the Agency's Ambient Lake Monitoring Pro-
           gram and include general recommendations for lake protection and management. The program has
           been very successful: 141 volunteers  participated in monitoring 87 lakes in 1981; in 1983 approx-
           imately 240 volunteers are scheduled to monitor 160 lakes. The program provides the volunteers with
           current data on their lake and how its transparency compares to other lakes in the State. It also pro-
           vides a historic data base for determining seasonal and long-term trends in lake quality. The volunteer
           monitoring program has resulted in implementation of lake protection/restoration measures for several
           lakes. Federal, State, and local agencies have used the data collected to help assess the severity
           of water quality impacts from agricultural runoff and target resources for water quality benefits. Volunteer
           data have helped document water quality problems, point out critical areas most responsible for water
           quality degradation, guide the implementation of lake protection/management techniques, and evaluate
           their effectiveness.
INTRODUCTION

The Volunteer Lake Monitoring Program (VLMP) is an
outstanding success of the Illinois Environmental Pro-
tection Agency. This cooperative lake monitoring ef-
fort involves two divisions within IEPA, the Division of
Water Pollution Control and the Office of Public and
Intergovernmental Programs, working with  citizen
volunteers. The  Program was initiated in  1981  to
gather additional  information about the lakes of Il-
linois and to respond to public interest and concerns
about lake quality and lake management. The program
encourages local volunteers to solve local problems.
The main reason for the success of the VLMP is the
enthusiasm and participation of the volunteers them-
selves. The volunteers demonstrate a solid base of
support  for environmental programs that include
citizen involvement.
                                             473

-------
 LAKE AND RESERVOIR MANAGEMENT
 CURRENT PROGRAM

 Approximately 255 volunteers monitored 160 lakes in
 1983. This represents an 84 percent increase in lakes
 and a 71  percent  increase in volunteers since 1G81
 (Fig. 1). Public water supply operators, Soil and Water
 Conservation District personnel, and Illinois  Depart-
 ment of Conservation State park site personnel are
 well represented among the volunteers, as are lake
 association members, lake residents, sportspersons,
 and interested citizens.

 Citizen's Role

 Over two thirds of the volunteers who started  in 1981
 continue to be active in the program. The volunteers'
 commitment includes attending a mandatory training
 session, providing  their own boating equipment, and
 collecting Secchi disk and field observation data con-
 sistently throughout the monitoring season at desig-
 nated sites in their lake.
   Secchi readings and field observations are taken
 twice a month  (at approximately 2-week  intervals)
 from May through October. More frequent sampling is
 suggested for those wishing to define watershed/lake
 quality relationships or assess the effectiveness of
 lake and watershed management practices. Samples
 taken after each major rainfall help document the ef-
 fect of  agricultural  runoff on  lake  quality (Fig.  2).
 Samples taken immediately before and 2 to  5 days
 after implementation of lake  management  practices
 (such as chemical treatment for algae or weeds) also
 help assess these practices (Fig. 3).
   Citizens select the lake they wish to monitor from
 among Illinois' 2,900 public/private lakes that  are six
 acres or more in surface area. New  volunteers  are
  trained at the lake of their choice by Agency Public
  Participation Coordinators. The volunteers are loaned
  a Secchi disk and calibrated braided nylon rope and
  given a fact sheet describing the program. Detailed in-
  structions  on measuring  Secchi  transparency and
  total  depth, making field observations, and com-
  pleting data forms are included. A lake map showing
  the locations of three or more sampling sites desig-
  nated by the Illinois EPA is also provided.
   Volunteers are asked to sign a waiver of liability and
  an equipment loan form. The volunteers also complete
  a three-page lake assessment survey that provides in-
  formation on  lake  morphology, uses, water quality
  conditions,  shoreline   and  watershed  conditions,
  potential pollution sources, and current lake manage-
  ment practices.
   During the training  session, the coordinator and
  volunteer use the volunteer's boat to visit every desig-
  nated site on the lake, whereupon the volunteer is in-
  structed in the proper procedures for using the Secchi
  disk, recording field observations, and completing the
  required data forms for  each  site. Continuing quality
  control is assured because the coordinator actually
  takes part in collecting  data on the volunteer's lake.
                                                                           Jiu ,ll  L  Ji I,  ni  ,  I
                                                                           I I I I I ! I I I I ! I  | I  | | | I | |  | | M !  I
                                                      Figure 2.—The effects of rainfall events and resultant water-
                                                      shed runoff on the Secchi disk transparency of Lake Kincaid,
                                                      Jackson County, Illinois. (IEPA Volunteer Lake Monitoring
                                                      Program, 1982).
Figure 1.—Location of lakes in the 1983 Illinois Environmen-
tal Protection Agency's Volunteer Lake Monitoring Program.
Figure 3.—Impacts of chemical treatments of macrophyte
and algae on Secchi disk transparency in Lake Barrington,
Lake County, Illinois. (IEPA Volunteer Lake Monitoring Pro-
gram).
                                                  474

-------
                                                     ROLE OF LOCAL LAKE ORGANIZATIONS & PUBLIC EDUCATION
All subsequent data forms received from the volunteer
can be compared to these first readings for consis-
tency.
Agency Role

These individual training sessions on the volunteers'
lakes are essential  for  program  success and in-
valuable  in addressing the citizen's concerns.  Per-
sonal visits  to the  lake  are also  important  for
establishing quality control, evaluating lake problems,
determining sources of the problems, and involving
the volunteers in lake management. Notes made on
lake conditions, problems, and volunteers' concerns
are useful when analyzing data and preparing reports.
  The Agency  provides postage-paid envelopes for
the volunteers to mail in the  Secchi  monitoring and
field observations data forms  after each reading. The
forms are logged  in as received  and an up-to-date
checklist  kept  of  the  volunteers'• returns.  The
computer-coded portion of the data forms are com-
pleted, the forms keypunched, and the data submitted
to STORET.  A Tektronix  microcomputer terminal  is
used to prepare tables and graphs for reports.
  An annual  report by the Agency summarizes the
methods  used and  the results of the program (Sefton
and  Little, 1982; Little and Sefton, 1983). The report
also contains a primer to provide the nontechnical per-
son with  a basic understanding of Illinois lakes, fac-
tors affecting their  water quality, and actions that can
be taken to protect  and enhance them. Data sum-
maries are prepared  for each lake and individual
reports written  as  resources permit. Eighty-seven in-
dividual  lake  reports, which present and analyze the
data for each lake  and provide general recommenda-
tions regarding  lake  monitoring  and management,
were written in 1982 (IEPA, 1982). The reports provide
direct feedback and a sense of accomplishment to the
volunteers.
  Monthly newsletters are mailed to the volunteers
from June through September. They contain important
reminders regarding  the  program, transmit publica-
tions or other informative materials, and contain a
question  and  answer section  concerning lake condi-
tions, lake monitoring, and lake management tech-
niques.
  At the end of each monitoring season, all volunteers
are sent thank you  letters for their participation in the
program.  The volunteers who provided data for six or
more sampling periods (there are 12 sampling periods
in a season) are sent certificates of appreciation sign-
ed by the agency director.
  Pollution complaint forms are sent to all volunteers
who indicate on their field observation form that pollu-
tion problems are occurring in their lake. These com-
plaints are then investigated by IEPA Field Operations
Section staff  and the site visited if necessary.
  Lake watershed tours are arranged and conducted
for  volunteers  in various regions of the State by
Association of  Illinois Soil  and Water Conservation
Districts staff, in cooperation  with Soil Conservation
Service,  the local Soil and Water Conservation  Dis-
trict, and IEPA staff. The purpose of these tours is to
acquaint volunteers with the impact that land use in a
lake watershed may have on lake quality. The tours in-
clude a discussion of the erosion and sedimentation
process,  watershed sources of sediment and nutri-
ents, land uses  favorable and unfavorable to water
quality, and land treatment to control watershed run-
off.
  Volunteers are actively recruited for the following
year's program  in late winter. Letters and  sign-up
forms are sent to all former volunteers to encourage
them to continue  to monitor their lake.  Consistent
data gathered over a period of years  is necessary to
document water  quality trends, identify problems, and
evaluate lake and watershed management strategies.
Public water supply operators, State  park personnel,
and Soil and Water Conservation District personnel
are also recruited for the program and receive special
letters.  Recruitment  articles also appear in  various
publications, newsletters, and  newspapers. A poster
display about the VLMP  is used  at  various con-
ferences and speaking  engagements.


AGENCY  RESOURCE COMMITMENT

There is a misconception that the use of volunteers
will lessen the work load of Agency staff. This is not
true. Working with volunteers is a time-consuming ef-
fort. While it is true that volunteers save resources in
actual data collection, the overall VLMP coordination,
recruitment and training of volunteers, data handling
and subsequent analysis, newsletter preparation, and
final reports require a  substantial  resource commit-
ment. The success of any volunteer program is direct-
ly related to the amount of time allocated by the agen-
cy  staff to support the  program.
  An estimated  4  person-years of Agency staff time
was needed for  the 255 volunteers and 160 lakes in
lEPA's 1983 program. Staff  is needed throughout  the
year for program  management,  data handling, and
followup, but additional staff time is  required during
critical  periods (such as training sessions and report
writing for individual lakes).
  Staff is needed to complete  on-lake training  state-
wide for all new volunteers during a 6-week period
(April 1-May 15). Cold, wet,  windy weather can  make
training a real challenge! On the average, volunteers
on  two  lakes can  be trained per day. Staff normally
work 7 days a week during  this training period since
many volunteers are available only on weekends.
  Staff  also must be available from May through Oc-
tober to meet the needs of the volunteer. This program
cannot  be  put on a shelf  but must be  constantly
monitored to insure quality control and continued par-
ticipation. Data forms must  be logged in and checked
daily. If  no reports  are coming in from a lake or forms
are filled out incorrectly, the volunteer must  be con-
tacted. A delay in contacting a volunteer can result in
missed  or unusable data.
  Address  lists  and phone numbers of  volunteers
must be kept accurately. It is important that the volun-
teers' names are spelled correctly. These lists should
be set up in a format to permit multiple uses and must
be updated monthly, particularly if third class mail is
used to send newsletters to the volunteers.
  The role of the volunteer should never be underesti-
mated or belittled. The volunteers have a vested in-
terest in protecting a lake of their choice. It would be
wrong to squander such a valuable resource. Volun-
teer programs should be encouraged only if there is
committed agency support (staff and funding) to start
up the program,  to maintain the program, and to in-
crease the agency commitment as the successful pro-
gram grows.

PROGRAM RESULTS

The IEPA is pleased with the success of the Volunteer
Lake Monitoring  Program, both in terms of the useful
                                                 475

-------
 LAKE AND RESERVOIR MANAGEMENT
 data collected and the important service provided to
 the citizens of Illinois.

 Useful Data

 Spatial, seasonal, and long-term trends in Secchi disk
 transparency,  together with field observations and
 lake assessment information  are used to  identify
 water quality problems and possible causes and
 evaluate  and  implement  alternative  protection/im-
 provement strategies.
   Volunteer data are used  several ways:
   1. Assess the basic lake character and  possible
 pollutant sources;
   2. Identify prevailing conditions in different parts Df
 the lake so as to pinpoint inlake problems and possi-
 ble solutions;
   3. Estimate the dissolved oxygen resources of the
 lake, which affect the ability of the lake to support a
 sport fishery,  public water  supply, or recreational ac-
 tivities;
   4. Document water quality  impacts of nonpoht
 source pollution, in order to support applications for
 U.S. Department of Agriculture assistance programs
 in  the watershed, guide the implementation of agri-
 cultural  resource management systems  to critical
 areas, and evaluate the subsequent effectiveness (see
 Sefton and Little, 1982b, c);
   5. Guide lake management decisionmaking  (such
 as determining proper timing and application rates of
 copper sulfate for algal control or determining public
 water supply  withdrawal depths  for improved  water
 quality);
   6. Establish an historical  data base for the lake,
 which includes morphological  data, information on
 water quality  conditions and problems, lake, water-
 shed, and shoreline uses, potential pollution  sources,
 and lake management undertaken in addition to trans-
 parency, field observations, and total depth data col-
 lected under the VLMP. Without this historic record, it
 is almost impossible to document changes that have
 occurred  or predict the effects  of lake restoration or
 potential  pollutant sources. In many cases, the volun-
 teer  data  is the only monitoring data. It is almost
 always the most current data available.
  7. Compare data among other lakes in the State, in
order to target public and private resources  for Iak9
protection and management. In Figure 4, for example,
the  lakes are  categorized  by average Secchi  disk
transparency and plotted on a State  map. This mao
shows the area of the State with greatest problems
with sediment  or nutrients, and  has been used  by
State and  Federal agencies to target sediment control
programs  into  these areas.  The  data have also  been
used by Soil and Water Conservation Districts to  iden-
tify the lake watersheds most in need of soil conservsi-
tion measures, and by the IEPA to prioritize projects
for  Clean  Lakes funding and other assistance pro-
grams. Cooperative efforts fostered by the VLMP havs
also helped in  implementing lake protect ion/restora-
tion projects.
Service Function

The Volunteer Lake Monitoring  Program provides an
excellent opportunity to work with citizens concerned
with lakes and to foster cooperation and develop local
support for environmental programs.
  This was vividly demonstrated at  Lake Kinkaicl,
where  a  cooperative monitoring effort involving the
Jackson County Soil and Water Conservation District,
  the Kinkaid-Reed's Creek Conservancy District, the
  lake's public water supply operator, the Soil Conserva-
  tion Service, and the IEPA helped establish the Lake
  Kinkaid watershed as the number one Soil and Water
  Conservation priority in the State. The volunteer data
  helped  document the  effects of agricultural  runoff
  from the watershed on lake quality (Fig. 2) and guide
  the implementation of  soil conservation measures to
  critical  areas  of the watershed  (Sefton and  Little
  1982b).
   As a  result, about $35,000 of 1983 Jobs Bill funds
  were used for recreational development of Lake Kin-
  kaid, and planning authorization for a P.L. 566 water-
  shed project was received in August 1982. The Lake
  Kinkaid watershed is the next P.L. 566 land treatment
  project  scheduled for  funding  in Illinois, partly be-
  cause of the data, interest, and cooperation generated
  by the VLMP (Hendrickson and Fitzgerald, 1983).
   Similarly, at  Lake Sara, the  volunteer monitoring
  program  helped  establish  a working  relationship
  among  the  Effingham  Soil and Water Conservation
  District, the lake owner (Effingham Water Authority),
  and the City of Effingham (which uses the lake as an
  alternate public water supply). The Effingham SWCD
  has implemented a special watershed project based
  on the interest in protecting Lake Sara. The volunteer
  monitoring will also help evaluate the effectiveness of
  this project (Hendrickson and Fitzgerald, 1983).


  FUTURE PROGRAM

  IEPA has made a commitment to volunteers to con-
 tinue the program so that conclusions may be drawn
 from the data from the volunteer's  lake. The future
   SECCHI TRAMSPARCNC
    LEGEND

 * >79 INCHES
 • >48<7g  INCHES
 A >24<48  INCHES
 • <2t INCHES
          ""


           A.
                     >U.T -rirfi -" -T—- ->""  'f" -
            vl|^=.---*--iif •'••__ -A i
-•«
Figure  4.—Secchi  disk  transparency can  help  target
resources to areas in the State with the greatest sediment
and/or nutrient problems.
                                                 476

-------
success and expansion  of the program depends on
the Agency resources devoted  to  it. Currently, pro-
gram growth is limited by Agency resources.  Limita-
tions of staff time for aggressive followup is evidenc-
ed by a reduction  in the  number of data returns from
the volunteers  in  1983. New developments in the pro-
gram are also  necessary to maintain volunteers' in-
terest and participation.
  Suggestions for future program  development in-
clude preparation of informational  and educational
materials on lakes and lake watershed management,
an  annual  conference  (or  regional workshops) for
volunteers, preparation of individual lake reports that
incorporate chemical data collected under the Agen-
cy's Ambient Lake Monitoring Program, collection of
water samples by volunteers for  chemical analysis
(suspended solids, nutrients, and chlorophyll),  expan-
sion of  the  technical assistance aspects of the pro-
gram, and involvement of the regional planning agen-
cies  in  the administration  of the  program in their
areas. Implementing these suggestions  will require  a
substantial  increase in resources.

CONCLUSIONS

1. The Volunteer  Lake  Monitoring  Program  enlists
and  develops local grass roots  support for environ-
mental  programs  and  fosters  cooperation   among
citizens, agencies, and various units of government.
  2. The  VLMP  increase citizens'  knowledge and
awareness of the factors that affect lake quality and
promotes ecologically sound lake protection/mange-
ment techniques.
  3. The VLMP is a self-help program that  promotes
local  self reliance and implementation through local
resources.
  4. The VLMP targets public and private resources
for lake  protection and improvement.
  ROLE OF LOCAL LAKE ORGANIZATIONS & PUBLIC EDUCATION

   5. VLMP data documents water quality impacts of
 point and nonpoint source pollution.
   6. The VLMP provides a historic data baseline for
 documenting future changes and evaluating pollution
 control programs.
   7. VLMP data supports lake management decision-
 making.
   8. The VLMP provides the framework for a technical
 assistance program for lakes.
   9. The   VLMP  requires  a  substantial  agency
 resource commitment.
REFERENCES

Hendrickson, H. and W. Fitzgerald. 1983. Report to Associa-
  tion of Illinois Soil and Water Conservation Districts on the
  Volunteer Lake Monitoring Program. Springfield, III.
Illinois Environmental Protection Agency. 1982. 1981 Volun-
  teer Lake  Monitoring  Program Report. A  Cooperative
  Citizen—Illinois Environmental Protection Agency Project.
  Div. Water  Pollut. Control,  III. Environ. Prot. Agency,
  Springfield.
Little, J.R. and D.F. Sefton. 1983. Volunteer Lake Monitoring.
  1982. Monitor. Unit, Div. Water Pollut. Control, III. Environ.
  Prot. Agency, Springfield.
Sefton, D.F. and J.R.  Little. 1982a. Volunteer Lake Monitor-
  ing. 1981. Monitor. Unit, Div. Water Pollut. Control, III. En-
  viron. Prot. Agency, Springfield.
	1982b. Water Quality Assessment of Lake Kinkaid,
  Jackson County, Illinois, 1977-1981. Monitor.  Unit, Div.
  Water Pollut. Control, III. Environ. Prot. Agency, Spring-
  field.
	1982c. Water Quality Assessment of Lake Sara,
  Effingham County,  Illinois, 1977-1981. Monitor.  Unit. Div.
  Water Pollut. Control, III. Environ. Prot. Agency, Spring-
  field.
                                                  477

-------
 SMALL LAKES SYMPOSIA  PROGRAMS
 VIRGINIA M.  BALSAMO
 Barrington, Illinois

             ABSTRACT
             Lake conservation begins with grass roots efforts recognizing the needs of the area. The following
             help foster awareness and effective lake conservation programs: (1)stream cleanups create public
             awareness and provide case studies to help ethers; (2) lake conservation symposia at local colleges
             allow speakers to address the lake conservation needs of local communities; (3) collections of lake
             resource materials for a local library provide easy citizen access to information; (4) a core contact
             list of local people most active in lake conservation can be circulated to encourage participation at
             lake conservation activities; (5) a file of names and addresses of all people interested in the lake com-
             munity aids their notification of all events; and |6) educational projects for youth groups broaden public
             awareness of lake problems. Citizens should be encouraged to contact their local Extension Office,
             Soil & Water District, Health Department, Planr ing Office, Soil Conservation Service, and other citizens
             for assistance and advice in lake conservation projects. These efforts which have been implemented
             in northeastern Illinois are applicable wherever lake conservation is a matter of concern and interest.
INTRODUCTION

My interest in lake management began when the fami-
ly took a tent camping trip around Lake Michigan. Th«
clarity of Lake Superior told me something was ter-
ribly wrong with our small and less clear lake at home.
But what could an ordinary person do to help her com-
munity address the large task of improving lake water
quality? My first lesson in lake  management was to
realize you do not  need a specialized background to
be a driving force in your community.
  That fall I enrolled in a lakes class at the local com-
munity college for a 2-day seminar, and I found ou:
most of the other lakes in the county were also having
serious problems.
  We began a public awareness program  in Illinois;
step  by step—first  with  our  local  homeowners;
association—reading  articles,  magazines,  and
newspapers; attending local meetings, developing a
keen interest in lake ecology. Most work involves mak-
ing others aware of our lake water supply.
  Later, it seemed a good idea to coordinate a clean
stream activity as a 4-H project. That served two pur-
poses, giving the young people a chance at  conserva-
tion awards and credits and a clean stream.
  We were  so  impressed with the results  of  this
cleanup and community response that we were en
couraged in continuing to  help others become more*
aware of lake management through action and knowl
edge.
  In  1976 our community college decided to drop its
lake management workshop after 3 years. I was deter
mined not to let this happen, which eventually lead to
the first Management of Small Lakes Symposium ir
1977. Over the next 6 years additional seminars have1
emphasized watershed and lake management.
  During this same  period I began to collect lake1
management resource  information available  frorr
local, State, Federal  and private sources. Thanks to
our local Barrington Area Library, resource  materials.
have been made available to citizens for reference.
  Other spinoffs from our  lake symposium includec
setting up a core contact list of concerned citizens
collecting  an  extensive file  of lake management
materials, working with youth groups on educational
programs, making people aware of local agencies anc
their functions, and encouraging people to become in-
volved in State and local organizations.
STREAM PRESERVATION VOLUNTEER
PROGRAM: FLINT CREEK

Flint Creek is a tributary to the Fox River in southwest
Lake County, III. The communities of Barrington, Bar-
rington Hills,  Lake and North Barrington are in its
watershed. Several small  lakes  are located  on  the
creek  as are several  man-made  lakes  which  are
environmentally and financially important to the quali-
ty of life in the area.
   In the mid-1970's the lakes and Flint Creek began to
fill with sediment and debris and became a focal point
of concern to area citizens. Fecal conform counts in
the creek (a measure of the presence of human waste)
were as high as 800,000/ml. Water with counts higher
than 1.0/ml is considered unfit to drink. The people of
the area took their pollution problem to  the  Illinois
Pollution Control Board and Environmental Protection
Agency and  through these  agencies  corrected  the
fecal coliform problem.
   The  problem of  accumulated debris in  the stream
was also resolved by means of a coordinated and well
planned effort.  Volunteers  and  public agencies
organized a "Cleanup Junket Flint Creek." The key to
its success was the tapping of an unlimited source of
energy—high  school students—through  local  4-H
clubs,  Boy Scout, and Girl  Scout organizations.
  The  cleanup was assisted  by the Lake County Soil
and Water Conservation District, Public Works Depart-
ment,  Highway   Department,  Health  Department,
Cooperative Extension Service, Citizens  for Conserva-
tion, Flint Lake Interested  Property Owners Associa-
tion, and Defenders of the Fox River, Inc. These Agen-
cies and groups notified landowners of the cleanup;
arranged for  trucks to haul the debris  away to pre-
determined disposal sites; arranged for paramedics to
be available if injuries occurred (none did); obtained in-
surance; and  advertised the event  with  posters, in
local papers, and on local TV talk shows and radio.
                                                 478

-------
                                                       ROLE OF LOCAL LAKE ORGANIZATIONS & PUBLIC EDUCATION
 Free  lunches  were  furnished  to  the  volunteers,
 courtesy of McDonald Corp.
   The  creek was divided  into roughly half-mile seg-
 ments  (the total length of  the creek is approximately
 11 water miles) with each segment beginning and en-
 ding at a road crossing. Teams of eight students with
 an adult supervisor  were assigned  a segment and
 equipped with donated garbage bags. Each student
 was  also required to submit a short form indicating
 parental  consent. Warned on the dangers of poison
 ivy, barbed wire, broken glass, and deep water and ad-
 vised to be courteous to landowners and respectful of
 private property, they were  then sent out to remove the
 debris.
   A word on organization: start out with a few leaders'
 phone  numbers—get more as you go along. Be sure to
 be selective about age; don't encourage children too
 young. You need a firm committment.
   Safety is a concern. Local paramedics with ambu-
 lance were available at  the  control  center. Radio-
 equipped cars and walkie talkies were used for safety
 and  general coordination, proper clothing  for  the
 weather was insisted on—long pants, high socks, and
 long  sleeve shirts. Pass on information to local and
 county police.
   The  leaders  were given  maps and asked to mark
 down the spots where  immovable items were  left
 behind or drain  tiles were polluting the creek.
   The results of these efforts were very rewarding. Ap-
 proximately 200  people  participated and  removed
 three, 5-ton truckloads of  debris during the 4- hour-
 long  cleanup.  It was said that  virtually everything
 needed to furnish a house was removed from the
 creek that day.  The total cost to the agencies to ac-
 complish the cleanup, excluding volunteered services
 and equipment, was estimated at $75.
   Many objectives were accomplished during this
 coordinated volunteer cleanup. The most  important
 was to make community aware  of 'their' creek. The
 cleanup instilled a sense  of pride and value  in this
 neglected community asset. Community awareness is
 a necessary step in promoting and maintaining any
 successful stream preservation  program.  People  in
 the future will  think  before disposing of trash in a
 creek they've spent so much effort in cleaning. They
 will also pay more attention to problems occurring in
 the creek.
   Today in northeastern Illinois we are encouraging
 the adoption of a complete stream preservation pro-
 gram (III.  Div.  Watersources,  1983)  through  com-
 prehensive stream management. This program en-
 courages intergovernmental agreement and local or-
 dinance adoption.
WATER QUALITY EDUCATION PROGRAMS

With my three active children, home life was 4-H pro-
jects, projects, and more projects. The entire two walls
at the Cooperative Extension Service are filled with
projects of every nature, except lake ones. Checking
further, the same topic  was practically avoided by
other youth groups. My question was: How can such
an important issue be forgotten? Thus the birth of our
4-H project manual—the same committee working
toward a better environment in Illinois.
  The main objective of a lakes project manual or
primer on lakes and ponds is to provide youth groups
and instructors with the  material whereby they may
gain basic knowledge of the functions and importance
of bodies of water in their environment (see Exhibit).
   We  anticipate  our  fundamental  course  to be
 separated  into the study sections. One  deals  with
 gathering information in the classroom or home study
 situation on the kinds and uses of ponds and lakes,
 their management, surface water conflicts, sources of
 water, and the factors that affect water quality. The
 second section deals with observations of lakes and
 ponds in the field and conducting certain basic ex-
 aminations of the bodies of water and their water-
 sheds.
   It is intended that the participants in this course will
 develop not only an awareness of the importance of
 waterbodies in the environment,  but  will  be able to
 understand and develop management strategies for
 the  improvement of both water quality and  aquatic
 habitat.
   With the help of the Cooperative Extension Service
 or Soil and Water Conservation District a project such
 as this could be disseminated anywhere to those in-
 terested youth  in  lake management  programs.  The
 following outline has served us  well  and could be
 amended to suit local needs.
   EXHIBIT OUTLINE FOR 4-H PROJECT, STUDY GUIDE
 I. Introduction: Includes basic information about ponds and
 lakes, type use and management.

 II. Kinds of lakes and ponds.

    A. Natural
      1. Glacial origin
      2. Oxbow (cut-off river bends)
      3. Earth faults (Reelfoot  Lake, Tenn.)
      4. Limestone sink pits
      5. Other
    B. Man-made (artificial) lakes and ponds
      1. Farm ponds (most numerous—80,000 in Illinois)
      2. Reservoirs constructed on small tributaries
      3. Reservoirs constructed on main streams
      4. Quarries, gravel pits
      5. Strip mines

 III. Lake management considerations

    A. Water quality by uses
      1. For drinking
      2. For body contact (swimming)
      3. For noncontact recreation (fishing, boating)
      4. For fish and wildlife
    B. Sources of water
      1. Surface runoff
      2. Groundwater seepage
      3. Springs
      4. Aquifer characteristics and groundwater recharge
    C. Surface use conflict
      1. Citizen, landowner opinions on the importance and
        priority of each use
      2. Zoning  in time and space (priorites)

  IV. Uses of lakes and ponds

    A. Recreational
    B. Real estate development
    C. Municipal or private water supply
    D. Hydroelectric or  pump-storage or cooling water for
        electrical power plants
    F. Flood control
    G. Irrigation  for agricuture
    H. Water for industrial uses other than electricity
    I.  Water for livestock and fire control in rural areas

V. Eutrophication process, what is it?

    A. Weeds and part they play in biology of lakes
    B. Nutrients, how they enter and are held
      1. Nitrogen cycle
      2. Phosphorus cycle
                                                 479

-------
 LAKE AND RESERVOIR MANAGEMENT
     C. Sedimentation of a body of water
       1. Filling in by inorganic sediments
       2. Filling in with organic sediments
     D. Influence of  land use in watershed
     E. Poisons that can enter water (insecticides,
       herbicides, and other toxic industrial materials)
     F. Inadequately treated sewage from either sewage
       treatment  plants or septic systems

 Phase II • Visit Lakes and Land Resource Agencies (SCS)

 I. Visit  at  least  one kind of artificial and  natural  lake,
  examples:
     A. Natural, glacial lake - Lake Defiance in Moraine Hills
       State Park
     B. Artificial lake - a typical farm pond or a real estate
       development like Loch Lomond
 II. Review aerial photos from Soil Conservation Service of-
    fice,  measure watershed, review county Water  Fle-
   sources Book (State Department of Conservation)

    A. Map the types of watershed uses
      1.  Farmland (row crops, small grain)
      2. Woodland and pasture
      3. Animals  present (livestock)
      4. Erosion and pollution
      5. Incoming streams
    B. Kind of dam  and spillway
       Presence of drainage device (valve)

 III.  Size (ha),  average depth, maximum depth, volume of
    lake, and configuration of shoreline

    A. Nature of shoreline use (cottage, beaches, etc.)

 Phase III • Sample and Tests on Lakes and Visit to Water
                 Resource Agencies

 I. Do simple water chemistry tests, such as transparency,
  temperature and PH

 II. Collect, identify aquatic weeds, algae (use Department of
  Conservation booklet on how to identify aquatic plants;
  , Learn history of  lake, when built, weed control, winter-
   kill, fishing success, and identify primary lake uses
IV. Collect, identify fish species and other aquatic animals
   (by minnow seine and fishing)

    A.  Requirements of warmwater,  coolwater, coldwater,
      fishes
    B. Visit nearby hatchery

  Phase IV • Determine Management Remedies for Lakes
 Visited that Lake Owners and Resources Agencies Might
                     Follow Up

I. Water pollution abatement

    A. Point sources, such as sewage treatment plants
    B. Nonpoint sources, such as erosion

II. Better aquatic habitat

    A. Deepening  lake (to prevent winterkill)
    B. Weed control (emergent, submergent, algae)
    C. Rehabilitate fish population
    D. Protection of wetlands, spawning areas and wildlife
      areas

III. Management alternatives

    A. Watershed  diversion
    B. Lake aeration
    C. Selective discharge from dam
    D. Lake drainage
    E. Nutrient flocculation
    F. Watershed stabilization
    G. Erosion control on banks, shoreline
    H. Other

IV. Literature sources & responsible agencies


MANAGEMENT OF SMALL  LAKES
SYMPOSIA
That  "water attracts people"  is a  well-known fact.
Well-organized lake symposia attract people, too. This
is what we discovered after organizing six such sym-
posia over the last 7 years. We have attracted over 700
people.
   Community involvement is the major key for a suc-
cessful symposium. A  well-organized committee to
manage and promote the seminar is an important first
step:  it is best if this committee is a mix of local,
State, and Federal agency people along with selected
individuals.  This committee will offer a  stabilizing
force year after year. One member  should serve as
coordinator. Basically these individuals plan the agen-
da, contact the speakers, and handle the logistics.
   Our committee chose to develop a yearly theme for
each of  our  symposia.  The  first  year  watershed
ecology and management was our overall  theme. We
also followed  a year-to-year approach on the sym-
posium since  we did not know what response  we
would get from the community. After 4 years the plan-
ning committee began  to plan in terms of multiyear
programs where one symposium will lead into the
following  year's theme.
   Our basic idea was to have at least a 1 day session
with technical  experts and involved  persons relating
their experiences and to  offer solutions to problems
each year of our symposium. We instructed each per-
son to provide  some general background information
on his topic before selecting a theme. The speakers
keep the  discussion on a nontechnical level to help
the audience grasp concepts and alternatives  before
proceeding  with the description of specific manage-
ment solutions.
   An example  of our proposed 1983 symposium for-
mat Total Lake Management was broken into three
parts:  introduction,  identification  of lake problems,
and solutions to lake  problems. In addition, two in-
dividuals  wrote a paper  on  the legal and financial
aspects of  lake management. A  poster session or
displays by  local agencies and private dealers have
been also a  yearly part of our symposia.
   Costs for  the program varied from year to year,
depending on location, local financial support, and ap-
proval of State and Federal grant money. When these
programs  are held at  community colleges, nominal
amounts can be charged. A registration fee of $15 was
the highest amount we charged: this paid for a catered
lunch.
   The major objectives of our symposia have been to
disseminate information about  lakes   and  lake
management, allow local  management organizations
to exchange ideas, and to encourage  community
cooperation  in total watershed management.
   These symposia have addressed the needs  of the
community and could be readily adapted anywhere.
   There still  is a need to provide nontechnical infor-
mation  on better lake management.  Proceedings are
available  for the 1982  Illinois symposium (III. Dep.
Energy Nat. Resour., 1982). This is one book that every-
one can understand. We need more nontechnical pro-
ceedings for reference.
                                                  480

-------
                                                     ROLE OF LOCAL LAKE ORGANIZATIONS & PUBLIC EDUCATION
LAKE RESOURCE COLLECTION

As a result  of our  symposia,  lake  shore property
owners and  the  interested  public found  that  lake
materials were not readily available. There is not ade-
quate exchange of information about lakes and ponds
between interested groups. We found a solution to
relieve the problem: starting a lakes resource collec-
tion. There are several catalogs available, the most
useful is a reference book County Agents Directory
(Centary Communic., Inc.).  It is  the  most  complete
listing available of agricultural leaders, organizations,
and reference information. To obtain a copy visit your
local  Cooperative Extension Service  Office or write
the publisher. Listings are available for every State in
the Union.
   A list of organizations, agencies and public officials
concerned with natural  resources use and manage-
ment is Conservation Directory, published by National
Wildlife Federation.
   Accumulate resource materials relating to  lakes
and   ponds  management—case  histories,   lakes
recreation, pollution,  aquatic weeds control, dredging,
areation, erosion,  watershed uses, lake uses, fish and
wildlife,  euthrophication,  dams,  landscaping—and
other  related  subjects.
   Store these materials  locally where easily accessi-
ble, perhaps a library, college library, or local agency.
   Be sure to publicize the existence of this material
for two reasons: to let everyone to know of its  avail-
ability and to allow people to add to the collection.
   Advertise with poster boards at local lake symposia
and display  material from lakes resource collection
for viewing, including addresses so people can send
for more information.
   A  reference  catalog  should  be created.  List
publishers addresses, a contact person, and phone
numbers. Then list books and pamphlets and other in-
formation that anyone can send away for.

CARD CATALOG

A successful small lakes program requires a complete
collection of updated interested individuals, agencies,
conservation  groups, homeowners associations, park
districts, villages, unincorporated  areas,  political
groups, and so on. This  list can be as simple as 3x5
cards or  as complex as  a computer  listing. In north-
eastern Illinois, our list has grown to over 2,000 per-
sons.
LAKES MANAGEMENT CORE
MEMBERSHIP/CONTACT LIST

To ensure the free flow of information about other lake
management sessions,  all organizations require  a
 membership list. Complete the names and addresses
 of individuals who are interested in  participating in
 many  lakes  management  programs. These  people
 should represent a variety of expertise and organiza-
 tions.  Our list is circulated every time an announce-
 ment  for a  lakes-related  activity  occurs,  with a
 noticable increase in participation at lake conserva-
 tion activities.
   Such a list is an invaluable source for newsletters,
 symposia, announcements, and the exchange of solu-
 tions to common problems.


 FUTURE ACTIVITIES

 Basically we hope to accomplish these  goals in  the
 immediate future:
   1. Establish a lake managers group  in our area,
 which  will have committees assigned to carry on with
 our current programs. For example: stream preserva-
 tion, small lake symposium, and  resource collection.
   2. Have these groups become members of the North
 American Lake Management Society.
   3. Have our small lakes manual approved and
 published by the University of Illinois Cooperative ex-
 tension Service for 4-H and other youth groups and
 schools.
CONCLUSION

Public awareness requires  promotion  through par-
ticipation in  events. To  increase public  awareness
about small lakes in northeastern Illinois such events
as  cleanup  days,  symposia,  gathering  reference
material, and compiling membership lists have been
used for the last 7 years. The success of one program
was the seed of another. But the overall theme was to
increase public awareness about small lakes and their
management.
REFERENCES

County Agents Directory. Annually. Century Communic., Inc.
  5520-G Tougy Avenue,  Skokie, III. 60077  Phone: (312)
  676-4060.

Conservation  Directory. Annually. National Wildlife Fed.,
  1412 Sixteenth St., N.W. Washington, D.C. 20036 Phone:
  (202) 797-6800.

Stream Preservation Handbook. 1983. III. Div. Water Resour.
  Coord. Policy Comm. III.  Dep. Transport.  Div. Water
  Resour. 201 West Monroe St., Springfield, III. 62706.

Local Self Reliance, Symposium VI. 1982. Doc. No. 82718.
  Pages 108. III. Dep. Energy Nat. Resour. Plann. Policy Div,
  Res. Section, 325 West  Adams,  Springfield, III. 62706.
  Phone: 1-217-785-2800.
                                                481

-------
  GRASS ROOTS  LAKE  AND WATERSHED
  MANAGEMENT ORGANIZATION
  ROBERT BURROWS
  Greenwood  Lake Watershed Management District, Inc.
  Hewitt,  New Jersey


  JOHN D. KOPPEN
  Princeton Aqua Science
  New  Brunswick, New Jersey


             ABSTRACT

             Greenwood Lake is a 777 ha lake in the suburban New York metropolitan area. The lake lies in New
             York State (Orange County) and New Jersey (Passaic County) and comes under the jurisdiction of
             both  States. The 6,000 ha watershed is 80 percent forested and 17 percent in residential land use.
             Greenwood Lake is a headwater of the Wanaque River in the Passaic-Hudson River Drainage Basin
             The lake experiences heavy recreational use including boating, fishing, and swimming. Over the course
             of the last 30 years the lake's water quality has declined appreciably and recreational usage has dropped
             as a  result.  Historically, many different citizens organizatons have been formed to address the pro-
             blems of the lake then gradually died. One major problem with maintaining an active and viable
             organizaton to coordinate and implement lake management activities on the lake is that the lake is
             in two States. Although a single lake and a single watershed, local parochial interests prevented coor-
             dinated action. In 1979 a group of citizens from New York and New Jersey formed the Greenwood
             Lake Watershed Management District, Inc.,  as a bistate committee to address problems of Green-
             wood Lake.  Accepted by the States of New Jersey and New York as a responsible political organiza-
             tion,  the GLWMDI applied for and received a grant under Sec. 314 of the Clean Water Act to carry
             out a lake restoration and watershed management study on Greenwood Lake. The GLWMDI has a
             30-member  board of directors, an executive director as chairman, and several hundred members.
             Throughout the 314 study volunteers contributed up to $100,000 worth of hands-on, in-kind services
             associated with the study. This contribution served as the matching funds needed to obtain the grant.
             Throughout the 4 years of its existence, the GLWMDI has unified the people in the basin into a power-
             ful action-oriented organization with the welfare of the lake as its primary goal.
 INTRODUCTION

 Greenwood Lake is a 777-hectare lake located in the
 suburban New York metropolitan area. The lake is lo-
 cated in both New York State (Orange County) and
 New Jersey (Passaic County). Since the lake is a bi-
 State body of water it comes under the jurisdiction of
 New York and New Jersey (Fig. 1). The watershed area
 is 6,000 hectare with 80 percent forested and 17 per-
 cent residential and commercial. Greenwood Lake is a
 headwater of the  Wanaque River  in the Passaic-
 Hudson River Drainage Basin.
   In addition to two States and two counties, three lo-
 cal governments are within the watershed. These are
 Warwick and Greenwood  Lake Village in New York
 and the  township  of  West  Milford  in New  Jersey.
 Greenwood Lake is recognized as a substantial asset
 to these communities  and water-related  recreation
 and business  forms the major economic base of a por-
 tion of these communities (Raymond et al. 1980).
   During the course of the last 30 years the lake has
 experienced heavy  recreational usage. This has been
 accompanied by urbanization of the watershed and
 the conversion of seasonal homes to permanent dwell-
 ings. Because of the lack of unified basin-wide action
 to provide adequate wastewater disposal, stormwater
 quality management, and a  viable public education
 program, the  lake's water quality has declined and,
 along with it, the attractiveness of the lake for recrea-
tion (Greenw.  Lake  Watersh. Manage. Dist., 1983).
   This situation is not unique to  Greenwood Lake.
 Historically, lake dwellers, developers, planners, and
 governmental officials have not appreciated the fact
 that a lake is a complex, dynamic, and sensitive eco-
  MNNSTIVAMA
Figure 1.—Location of Greenwood Lake in New York/New
Jersey Metropolitan Area.
                                                482

-------
                                                    ROLE OF LOCAL LAKE ORGANIZATIONS & PUBLIC EDUCATION
system that can be defiled so easily, even by the most
distant changes in  the  lake's watershed. In many
cases inaction results from ignorance or economic
consideration. However, in the Greenwood Lake case
this is false economy since the major economic base
in the basin is the lake-related recreational industry.
  In the Greenwood Lake watershed, effective basin-
wide action is complicated by the number of govern-
mental units involved.  The concept of one lake and
one watershed was  not recognized and, historically,
local parochial  interests prevented coordinated ac-
tion. Therefore, it became apparent that an organiza-
tion that could speak for the entire watershed and lake
was required.
  Many different citizen organizations were formed to
address the problems of the lake. However, the effec-
tiveness of these organizations was limited and they
gradually died. The major reason for their lack of suc-
cess was that they did  not represent the entire basin.
Therefore, in 1979 a  bi-State committee of concerned
officials and citizens founded the Greenwood Lake
Watershed Management District, Inc. (GLWMDI). This
case study is a chronology of its predecessors, its
evolution, and its activities.
PREVIOUS LAKE ORGANIZATIONS

In 1768, just 3 years after the first dam on Greenwood
Lake was completed to supply constant water power
to the "Long Pond Iron Works" downstream, the first
weed problem was noted  in the Ringwood Co. reports.
It seems the dam, which was 200 feet long, raised the
lake level four feet above its natural level. The problem
arose when workmen allowed the lake level to drop
back to its natural level to make some repairs to the
dam. The weeds that were exposed began to rot and
the stench was unbearable to the local farmers. These
same farmers formed a  committee and  approached
the owners of the iron works, demanding some relief.
The Ringwood Co. did the only thing possible; they
closed the breach in the dam and the weeds were
once more submerged, thereby ending  the stench.
This was the first example of a citizen committee be-
ing  successful.
  In 1836 a dam was  constructed on the Wanaque
River on the southeastern shore of the lake by the Mor-
ris Canal and Banking Co. Its purpose was to ensure a
constant supply  of water to  the  Morris and  Essex
Canal,  an important route  for  Pennsylvania  Coal
Barges traveling to Newark Bay. It was during this per-
iod  that the  lake and surrounding  valley  became
recognized as an  attractive vacation spot.
  Between 1856 and 1930 the Morris Canal and Bank-
ing  Co. and the railroad  companies that succeeded
them maintained areas of the lake free  from weeds
where there was  a possibility for a steamboat to bog
down.  Therefore, this  operation was limited to the
main docking facilities around the lake. There was lit-
tle or no public involvement.
  The  1960's  ushered in  a new surge of concern for
the lake because a development boom was  at  its
height. More year-round homes were  going up in the
watershed   and  visible  signs  of  accelerated
eutrophication were appearing in the  lake.
  The year 1963 saw the  formation of the Greenwood
Lake Aquatic  Weed Association in Greenwood Lake,
N.Y., to find ways and  means of clearing Greenwood
Lake of excessive weeds, silt, and mud. This organiza-
tion dissolved principally  because  the lake is an inter-
state body of water and the association had no sup-
port in New Jersey. In 1968 West Milford, N.J., resi-
dents formed the Greenwood Lake Conservation Com-
mittee which soon ceased to exist for the same rea-
son—it had no support or representation from New
York.
  Much activity occurred in 1969, including attempts
to contact  manufacturers  of aquatic weed  control
equipment.  West Milford started in the right direction
with the formation of the West Milford Township Inter-
state Clear Water Committee. A lot of research was
done, but loss of interest by some members hindered
the good start. This same year chemical weed control
was tried in the Belcher's Creek area of Greenwood
Lake. It  was  successful  but  the  chemical  used
(sodium arsenate) was banned soon after.
  The 1970's showed increased environmental aware-
ness and phosphate detergents came off the  shelves
in the watershed. The township of West Milford for-
mally asked the State of New Jersey Department of
Environmental Protection to set up a weed control pro-
gram. This started in 1975; however, it involved the use
of herbicides only and the areas treated were restric-
ted to New  Jersey.
  In September 1977 a comprehensive  management
planning meeting of interested citizens and local offi-
cials took place in West Milford, N.J. This was the first
meeting primarily  aimed  at formulating  long-term
planning. This spurred the township of West Milford to
form an advisory committee to the township planning
board. The  Greenwood  Lake Advisory Committee
spent 13  months studying the future of Greenwood
Lake. The committee's report was two volumes of data
and a recommendation that an interstate committee
be formed as soon as possible.
  This committee discovered the Clean Lakes Pro-
gram (Sec.  314) and began preliminary research into
an application for Federal assistance. However, it was
recognized  at this time that any organization would
have to encompass the entire  basin, therefore the
three communities  formed an appointed  committee
named the Greenwood Lake Improvement and Beauti-
fication Committee  which met for some time but end-
ed up reviewing the same things the Greenwood Lake
Advisory Committee had already covered extensively.
  In July of  1979 a combined meeting of governmental
agencies  and concerned citizens recommended the
formation of a single bi-State organization that would
encompass  all the  parties involved with Greenwood
Lake and primarily represent the lake itself. Using the
Greenwood  Lake Advisory Committee's Weed Report
as substantiating information, a preliminary applica-
tion for Section 314 (Clean Lakes) funds for a  Phase  I
study was submitted to the New Jersey Department of
Environmental Protection.
ESTABLISHMENT OF THE GREENWOOD
LAKE WATERSHED MANAGEMENT
DISTRICT, INC.

In September of 1979, the Greenwood Lake Advisory
Committee and Concerned Citizens from the New
York portion of the watershed met at Lakeside Com-
munity Clubhouse and, with the help of a consultant,
formed the Greenwood Lake Watershed Management
District, Inc., as a nonprofit corporation. Within mon-
ths the GLWMDI was recognized by nearly all govern-
mental entities. Under New York State law, Warwick,
Greenwood Lake Village, and the county of Orange
passed resolutions making GLWMDI an Aquatic Weed
Control District in New York State.
                                               483

-------
 LAKE AND RESERVOIR MANAGEMENT
    In 1980 the District, as it was dubbed, adopted a for-
 mal charter  and bylaws. The charter provided for a
 30-member board of directors, 10 of whom can be ap-
 pointed by the various local, county, and State govern-
 ments that have jurisdiction within the District. The
 other 20, along with the officers, are elected  by the
 membership at  large.  Currently  the  District  has
 several hundred members.
    The general  statement of purpose in the charter
 read:
    The  purpose  for which the coroporation is formed shall
   be the protection of the water resources and other natural
  assets of Greenwood Lake, its tributaries and watershed
  from misuse and pollution, the conservation of the scien-
  tific, education, scenic, water resources and recreational
  values of Greenwood Lake, the encouragement of the con-
  tinuation and  development of compatible  land uses in
  order to  improve the overall environmental and economic
  position  of the area;  and the preservation and orderly
  management of [he natural resources of Greenwood Lake
  and its watershed (GLWMDI, 1979).

  To  implement the statement of purpose presented
in the charter various activities were specified. These
included:
                             Compiled by members of

                        GREENWOOD LAKE WATERSHED
                        MANAGEMENT DISTRICT INC
Figure 2.—Promotional map showing camping areas and points of interest.
                                                   484

-------
                                                     ROLE OF LOCAL LAKE ORGANIZATIONS & PUBLIC EDUCATION
  1.  Oversee and assist in studies to develop man-
agement and restoration plans for Greenwood Lake.
  2.  Provide  assistance to local, county, and State
governments concerning those aspects of their plan-
ning  and enforcement activities that will affect the
lake.
  3.  Develop a public education program.
  4.  Develop funding strategies and sources of funds.
  5.  Serve  as the  administrator  for  implementing
Lake and Watershed Management Projects.
ACTIVITIES OF GLWMDI

Oversee and Assist in Studies

In June of 1980 EPA approved $140,233 for a Phase I
diagnostic and feasibility study of Greenwood Lake.
The District's board of directors screened proposals
from,  and interviewed nine  consultants. Following
selection, contracts between  the District, the New
Jersey Department of  Environmental Protection,  and
the consultant were executed.
  The actual 314 study began with special training by
the consultants of laymen volunteers  (of which the
District is mostly composed). This saved thousands of
dollars in labor costs to the study, freeing up profes-
sional hours for other  important activities associated
with this study.
  The volunteers were trained in various skills relating
to stream and lake sampling. They immediately began
daily observations and monthly sampling of their as-
signed locations and continued for a complete year.
Local marinas and individuals  supplied boats for all
the in-lake sampling, at no charge.  This  saved many
dollars over the course of the full year. Some of the
volunteers are still  monitoring their locations to give
continuity to the data  on Greenwood Lake.
Provide Assistance to Local, County, and
State Governments

Probably the most important development is that the
District has learned how to use the tools that are avail-
able to them to protect the lake. They have become in-
volved in the decisionmaking process in the  water-
shed through the public  participation channels avail-
able to them. A watershed group with the broad based
support that this one has, is a potent political  force.
  The District has become a responsible partner in
the decisionmaking process for those issues that can
affect Greenwood Lake.  In every major project plann-
ed for the watershed, the District has had input. Dur-
ing a drought emergency of 1981, the District was call-
ed upon to formally represent the interests of Green-
wood Lake. The communities now are actively seeking
advice from the District. Currently the District is assis-
ting the township of West Milford in its master plan
review, thereby insuring  that the  master  plan will be
sensitive to the needs of Greenwood Lake. Also, the
District is helping explore alternatives for developing
an adequate wastewater disposal system for Green-
wood Lake Village. The organization has served well in
this advisory capacity and in the  long term this inter-
action will be one of the most important  activities of
the District.

Public Education Program

The District has developed a major education program
by using a newsletter  and press releases and  by co-
sponsoring regional and statewide lake management
conferences. Through these activities the citizens of
the watershed have become aware of the problems
and the importance of Greenwood Lake. The 1-day
conferences have featured well known professionals
in lake management, regulatory affairs personnel, and
various  public  officials.  Through cooperation with
Passaic County, N.J., Camp  Hope a  county-owned
summer camp,  has become an education center for
these conferences.
  At the first conference a proposal was presented for
legislative consideration to lake and watershed man-
agement districts throughout New Jersey. This propo-
sal is gradually gaining legislative support throughout
the State.
  The last conference, in April 1983, was attended by
approximately 200 people representing 50 lake associ-
ations and was co-sponsored by the New  Jersey De-
partment of Environmental  Protection. This confer-
ence proved to be a vehicle by which individuals, asso-
ciations and local governments could change informa-
tion on lake restoration and management.
  In 1984 the District will host the 1984 North Ameri-
can Lake Management Society's International Sympo-
sium at  the Americana Great Gorge Resort, McAfee,
N.J.
 Develop Funding Strategies and Sources of
 Funds

 In 1982 the District was granted public foundation sta-
 tus by the Internal Revenue Service. This allows the
 district to receive funding from all sources and donors
 can  get substantial tax advantages.  Initial  inquiries
 are going  out  to  industry  to  financially assist the
 restoration of the lake.
   In May of 1983 the District sponsored the first annu-
 al Miller High Life Cup Regatta on the lake. This activi-
 ty netted over $10,000 for the foundation. The District
 was joined by 35 civic organization volunteers in per-
 senting the Regatta, thus the public participation was
 widespread. Additionally, private industry, i.e.,  Miller
 Brewing and local businesses, contributed money and
 manpower to make the event successful.
   Also, the District  is developing support  from  its
 elected representatives in State and Federal Govern-
 ment for grants to fund the restoration. The District
 recognizes that funds will be needed from  many
 sources,  both public and  private, and they are explor-
 ing all possibilities.
 Serve as the Administrator for Implementing
 Lake Management and Restoration Plan

The greatest challenge lies ahead. The management
plan for Greenwood Lake will call for establishing the
GLWMDI as a  legally constituted bi-State planning
board to implement the restoration and management
plan. This may require special legislation by both New
Jersey and New York. Under this legislation the Dis-
trict would ultimately look for authority to collect user
fees and raise other revenue to support the restora-
tion,  management, and  maintenance  of  the  lake.
Legislation similar to that in Wisconsin is  preferred.
However, intermediate steps can be taken  under ex-
isting legislation both in New York and  New Jersey.
  Although there is a long way to go, throughout the 4
years of its existance the Greenwood Lake Watershed
Management District, Inc., has matured and unified
                                                485

-------
LAKE AND RESERVOIR MANAGEMENT
the people in the basin into a powerful action-oriented
organization with the welfare of the lake as its primary
goal. This organization is an example of the type of ac-    Greenwood Lake Watershed Management District Inc 1979
tion citizens can  take to make things happen in lake      Charter and Bylaws.

     96men '                                         Greenwood Lake Watershed Management District, Inc. 1983.
                                                        Phase I: Diagnostic-Feasibility Study of Greenwood Lake.
                                                      Raymond, Parrish, Pine and Weiner. 1980. Economic Devel-
                                                        opment Plan for West Milford Township, N.J.
                                                 486

-------
LAKE ASSOCIATIONS AND THEIR ROLE  IN THE MASSACHUSETTS
CLEAN  LAKES PROGRAM, 1983
RICHARD GELPKE
Department of Geography & Earth Science
University of Massachusetts
Boston, Massachusetts


            ABSTRACT

            In 1983 the Massachusetts Division of Water Pollution Control began administering a program
            of  matching grants to  municipalities  for lake studies and cleanup/restoration activities.
            Municipalities proposed various projects on 40 lakes and ponds and 35 were able to generate the
            local match. In some cases lake associations assisted in the process. No correlations were
            found to exist between cities/towns, their population or economic characteristics and the
            presence or absence of an association, the size or trophic status of a water body. Apparently
            lake associations are individual or small group efforts but are not particularly a response to iden-
            tifiable, repetitive stimuli. It is likely they result from personal, social, and historical factors that
            are hard to quantify.
INTRODUCTION

In 1982 the Massachusetts Division of Water Pollution
Control  published "Clean  Lakes Program: Rules  &
Regulations, General Application & Administration."
This began the process of granting a State match of
funds for studies and  implementation programs to
restore and maintain the quality of lake waters. Cities
and  towns  applied for this match  from across the
State; some of the lakes for which funds were sought
have active lake associations concerned with the wel-
fare of the local environment.
  This situation provided an opportunity to analyze
what factors are associated with the existence of lake
and pond associations, their role in helping obtain the
matching State funds needed, and whether cities and
towns that  received funds  can  be characterized and
distinguished from those that did not. The program re-
quires  that a public agency make the application;
hence no lake association  appears as the applicant.
One objective of this paper is to assess the role of
associations in the program and to determine their ef-
fectiveness.
 DESCRIPTION OF THE PROGRAM

 The U.S. Environmental  Protection Agency has ad-
 ministered section 314 of the Federal Water Pollution
 Control Act for a decade and funded lake restoration
 projects for half that time (Mackenthun, 1980). Most of
 the effort has been  devoted to scientific analysis of
 the physical, chemical, and biological aspects, lake
 management  and  restoration,  and   cost-benefit
 analyses (U.S. Environ. Prot. Agency, 1980) while con-
 siderably less attention has focused on related legal,
 political, and social  issues. Massachusetts, recogniz-
 ing the decline in  Federal allocations for lake  pro-
 grams, has developed its own funding program which
 first gave matching grants in 1983.
   This program is funded at $30 million for 10 years by
 a bond issue and money not spent in one year can be
 carried  over to the  next year.  For  Fiscal Year 1983
 almost  50 applicants expressed interest in applying
for  the  cost  sharing  within their jurisdictions.  Of
these, 40 were offered a  match and only five were
unable or  unwilling  to generate the  local share  re-
quired. The total amount of funds allocated for FY83
were (in  round figures):
    State share: diagnostic/feasibility     =  $1.3 m.
               implementations        =  $1.4 m.
    local share: both phases              =  $2m.
Three times as many Phase I projects (studies) were
funded as  Phase II projects. The criteria for eligibility
were based on:
  recreational use (active contact/non-
    contact/passive)
  type of public access (beach and boatramp/beach
    or boatramp/undeveloped)
  trophic status rating (oliogotrophic/mesotrophic/
    eutrophic) (Chesebrough and Cooperman, 1979;
    Div. Water Pollut. Control, 1983)
  relative importance
The latter category "shall take into account such fac-
tors as, (a) the proximity of other publicly owned lakes
and ponds and their trophic state and recreational
potential; (b) degree  of public support for the  project;
(c) degree  of political support for the project; and (d)
historical efforts...."  (Dep.  Environ. Qual. Eng., 1982).
LAKES AND PONDS IN MASSACHUSETTS

Massachusetts  is endowed with many natural  lakes
created mainly  by glacial  activities such  as kettle-
holes.  The  State  contains  a  significant  number
developed by early industries and individuals under
the Mill Acts (Water Resour. Comm., 1970) for a variety
of reasons. Also, a number are more recent and pro-
vide water supply, flood control, recreation, etc.
  Of a total 2,859 lakes and ponds more than 1,600 (57
percent) are  10 acres  or  more  (Chesebrough  and
Cooperman, 1979). Many of these lakes have extensive
shallow shorelines caused by damming of a stream
flowing through a relatively flat meadow.
                                                487

-------
 LAKE AND RESERVOIR MANAGEMENT

   With the popularity of a summer house and the ease
 of train and subsequent auto transportation  around
 the turn  of the century, accessible lakes in  Mass-
 achusetts acquired their current settlement patterns.
 It is likely that these  early seasonal lake residents
 developed personal and family ties with neighbors but
 not in a formal way as in an association with officers
 and bylaws. Many people came in the summer to get
 away from that activity. A typical pattern repeated on
 many lakes was a clustering of seasonal  camps be-
 tween the road and water's edge. A visual  inspection
 of U.S. Geological Survey  topographic quadrangles
 reveals this repetitive pattern for more rural lakes.
   In urban areas lakes were probably settled earlier
 and  settlement distributed  more  evenly  about  the
 shoreline  to  suit  development  and aesthetics.
 Whether lakes and ponds were near urban areas or net
 they tended to attract  settlement in  many cases so
 dense that it limited public access (Hardy,  1977).
   The question of public access is important for man/
 lakes since the Colonial Ordinance states  that lakes
 and ponds over 10  acres in their natural state are
 "Great Ponds" and may not be private (Smith, 1950i.
 Dense settlement has prevented the public from being
 able to get to the water without trespassing on  private
 property.
   The resulting problems  of runoff and  nonpoint
 sources of pollution (cultural eutrophication) were ex-
 acerbated by the surge of population growth in Mass-
 achusetts and elsewhere, especially since World War
 II. Although Massachusetts had only an .8 percent in-
 crease between 1970 and 1980, it recorded 10 percent
 a decade increase between 1950 and  1960,  and again
 between 1960 and 1970 (Bur. Census,  1981).
DATA

For the cities and towns considered for funding (ap-
pendix A) the following classification develops:
  expressing interest in the program = 46
  made formal application and received grant = 40
  able to generate local share and match State = 34
  An analysis of the lakes involved with the program
further subdivides them into (1) those lakes that have
a board or commission, and (2) those which do not:
     lakes receiving  funding =  34
      with lake association  =  4
      without association = 30
     lakes not receiving funding = 12
      with lake associaton =  7
      without association =   5
  A  1980 listing  (Dep. Water  Pollut. Control, 1980i
identified 168 lake and pond associations in the State
Of those, almost 6 percent  (10)  responded to  c.
questionnaire survey on their organization and role in
the Clean Lakes Program.
   In reviewing the applicants (cities and towns) and
 those expressing interest the following  data were
 determined for each:
      1980 population
      1960-70, 1970-80 population change
      Median income (U.S. Census, 1982)
      Median value of housing
      % dwellings occupied by owners
      1982 expenditure/person by municipality
 For each lake:
     Size (USDA, 1978; Mass. Water Resour. Res
       Center)
     Severity rating (Chesebrough and Cooperman,
       1979; Mass. Dep. Water Pollut. Control,  1983)
     Amount of award
     % increase in housing on the lakeshore
 The latter information was determined from an inspec-
 tion of USGS  topographic maps between  the  1950's
 and the photorevised  series  published in the  late
 1970's (Thompson, 1981).
   For lake associations located in municipalities that
 expressed interest in the program (FY83) a question-
 naire gathered data on:
     How long they have been  organized
     Perception of lake degredation
     Purpose of the association
     Form of organization
     Membership and dues
     Meetings  and officers
     Level of activity in the application process
 Twelve associatons responded (6 percent)—four were
 in municipalities awarded funds, six were not (mainly
 because of problems of public access), and two were
 associated with lakes  on which an appointed town
 board carried the effort through.
ANALYSIS

In an attempt to ascertain differences between cities
and towns receiving funding or not and whether that
correlates with lake in municipalities with a lake/pond
association, available data was assembled. Simple
correlations (linear regressions  run on a Tl  SR-51-II
hand calculator) were developed between a variety of
factors  and aggregated  according to whether an
association  was present  or  not. Where correlations
were deemed inappropriate, means or  medians were
derived.
  No strong correlations developed between any of
the columns except an expected relationship between
median  income and  median value of  housing (r =
.80 ±).  There  is  no correlation  between  expendi-
tures/person and median values (income and housing)
or any of the other data columns (trophic status, dwell-
ing owner-occupied).
                                 Table 1.—All expressing interest in program.
Factor



Per person expenditure (mean $)
Median income (mean $)
Median value of housing (mean $)
% of dwellings occupied by
owner (median)
1960 to 1980 population increase
(mean)
City/Town
Receiving
Match
(N = 33)
774
19,000
50,300

70
45
37
Not Receiving
Match

(N = 12)
738
19,800
46,800

74
78
67
With an
Association

(N = 12)
111
19,200
44,900

75
71
55
Without an
Association

(N = 33)
722
19,800
52,700

70
68
38
                                                488

-------
                                                       ROLE OF LOCAL LAKE ORGANIZATIONS & PUBLIC EDUCATION
  Table 1 presents the trend of the differences or lack
of differences in the municipalities considered. The
applicants  were combined in two ways on the  table.
The left two columns are all interested, separated ac-
cording to whether they received a State match or not.
The right two columns are all who expressed interest,
divided by the presence of a lake association. The
table  is remarkable for its  lack  of significant  dif-
ferences between those elements  in towns receiving
funds  and those not and  presence/absence  of an
association.
  There seems a substantial difference, however, be-
tween  the  rate of population increase between 1960
and 1980, whether on a  mean or median basis.  An in-
spection of the cities and towns in the group (Table 2)
that received funding and did not have an association
reveals that all of the cities fell in  this category and
these  typically have lost population in the  past 20
years.  This is a common, well known phenomenon in
the Northeast. This  includes seven cities and  towns
that lost as much as 17 percent in  the two decades.
  Turning  to  an assessment of the lake and pond
associations that were reviewed, the sample size (N =
12) and the nature of the questions did not permit  a
quantitative analysis. The following characterizations
can be made:
   • All perceive the degradation  of their waters as
gradual and happening within the  last 10-25 years.
   •  On a 2:1 basis their lakes and  ponds have  receiv-
ed prior help (funds) from some source (local,  State,
etc).
   •  Most  associations have existed about as long as
the perceived environmental problems (mostly weed
or algae problems).
   •  Two thirds state the existence of the organization
is due to lake degradation; four state no  association
with that problem.
   • Most  have a formal organization with bylaws and
find members willing to serve as officers.
   • These organizations  generally  have  somewhat
fewer than 100 members, half to most or all of the lake
residents.
   • Dues are nominal and all have budgets, some  in
the several thousands of dollars range.
   • Most were  not politically active  in  their com-
munities,  i.e., did not attend local  government board
 meetings except when of direct interest.
   • All felt they were the group to  contact concening
lake issues, and that they were the "prime mover"  in
initiating  the  process and developing citizen support
even though the application was carried forward by a
 public agency.
   • Few contributed to the local financial match and
 looked to the city  or town to  provide the  principal
source of  money.

 CONCLUSIONS

 It does not appear that it  is possible, on  the basis of
 readily available data, to distinguish between cities
 and towns receiving funding and whether or not a lake
 association exists in that community. It is fairly clear
 that an association  may well be an integral part of the
 educational process for the public agency and the
 lake  residents,  but whether it is on  the whole the
 decisive presence is not at all clear in Massachusetts.
   Associations   are  not  characterized as  being
 associated  with wealthier or poorer communities,
 those which  spend  more or less per person, or those
 which grew faster  in population (perhaps), or  have a
 larger population. Also there seems no relation be-
tween the size of a lake, its relative location to popu-
lation centers,  its trophic status, or other physical
characteristics.
  Associations perceive themselves as important but
it is difficult to assess the relative weight of this im-
portance in the process. Less than one third of those
funded had an association even listed and several of
those are likely weak or ineffective. In only four were
there active associations that appeared to be signifi-
cant in the process.
                    Table 2.-

A. Municipalities receiving funding with active lake
   associations
 1. Watershops Pond, Springfield
 2. Congamond Lakes, Southwick
 3. Martin's Pond, North Reading
 4. Lake Lashaway, the Brookfields

B. Municipalities receiving funding without active
   associations

 5. Bartlett Pond, Northboro
 6. Porter Lake, Springfield
 7. Long Pond, South Yarmouth
 8. Lake Winthrop, Holliston
 9. North Pond, Hopkinton
10. Ell Pond, Melrose
11. Lake Chauncy, Westboro
12. Quacumquasit, Brookfield
13. Big Alum Lake, Sturbridge
14. Walker Pond, Sturbridge
15. Puffers Pond, Amherst
16. Lake Quannapowitt, Wakefield
17. Forge Pond, Westford and Littleton
18. Metacomet and Acadia Ponds, Belchertown
19. Pequot Pond, Westfield
20. Dudley Pond, Wayland
21. Lake Buel, Monterey and New Marlboro
22. Cedar Swamp Pond,  Milford
23. Chebaco Lake, Hamilton and Essex
24. Floating Bridge Pond, Lynn
25. Lower Mystic Lake, Arlington and Medford (MDC
    applicant)
26. Hardy Pond, Waltham
27. Monponsett  Pond, Halifax
28. Jennings Pond, Natick
29. Dunn Pond, Gardner
30. Willow Pond, Northampton
31. Lake Ripple, Silver Lake, Grafton
32. Morses Pond, Wellesley
33. Sluice and Flax Ponds, Lynn


C. Municipalities not receiving funding, with lake
   assocation

34. Lake Rohunta, Eagle Lake, Athol
35. Lake George, Wales
36. Lake Boon, Hudson  and Stow
37. Red Lily Pond, Hyannis (Barnstable)
38. Silver  Lake, Agawam
39. Lake Wyola, Shutesbury
40. Ashmere Lake, Hinsdale

D. Municipalities not receiving funding, without an
   association

 41. Crystal Lake, Orleans
 42. Lake Nashawannock, Easthampton
 43. Lake Holbrook
 44. Ft.  Meadow Reservoir, Hudson and Marlboro
 45. Mill Pond, West Newbury
                                                   489

-------
 LAKE AND RESERVOIR MANAGEMENT

 SPECULATIONS AND SUGGESTIONS
 FOR FURTHER RESEARCH

 It seems clear that lake associations are more a func-
 tion of social, personal, and historical factors that a-e
 difficult to quantify. It may indeed relate to individuals
 who are either knowledgeable or concerned and act as
 a catalyst for the formation and subsequent activity of
 an organization. This  suggests that no particular fac-
 tors work against the existence of such a group. If this
 is indeed true it may  well help to explain  the  lack of
 vigor that the several  statewide organizations  of lake
 and pond associations have experienced,  not  only  n
 Massachusetts but in Maine and Connecticut as we I.
   It is quite likely that the available census data  en
 municipalities is  too  gross and  needs to be much
 more  site-specific to  be of value in  assessing lake
 quality and resident interest.
   Lakes and ponds, at least in Massachusetts, do not
 seem to act as a focus of concern for the residents.  In-
 deed,  lakes may act more to separate people on op-
 posite shores than bring them together. It  is possible
 that the strong tradition of volunteer town government
 in Massachusetts may substitute for and detract from
 a local association in efforts to improve lake quality.
 In many  cases a city or town will appoint a "lake
 restoration committee" or assign a similar  function to
 an existing board (Board of Health, Public Works, Con-
 servation Commission). The very people who might
 otherwise act through an  association  may well be
 channeled into a municipal reaction to concerns.
   It will  be instructive to identify successful lake
 associations and analyze their common features. This
 likely will be an intensive and lengthy process. Since
 this is the first year of an ongoing process, analyses
 of future applicants to the  program and comparison
 with findings  here will be of interest.
   As of this  date, almost 70 applicants applied fcr
 grants under the Clean Lakes program in Massachu-
 setts for FY84. It is a proving to be a popular program
 and we need to develop indices of potential  success in
 allocating public money.


REFERENCES

Chesebrough, E. and A.  Cooperman. 1979. Massachusetts)
  lake classification program. J. New England Water Pollut
  Control Ass. 13: 19-35.
 Hardy R. 1977. The Impact of Urbanization on New England
   Lakes. New England Council of Water Center Directors.

 Mackenthun, K. 1980. Lake restoration—a historical perspec-
   tive. Pages 162-165 in Int. Symp. Inland Waters and Lake
   Restoration. U.S. Environ. Prot. Agency, Washington, D.C.

 Massachusetts Department of  Environmental Quality Engi-
   neering. 1982.  Clean Lakes Program: Rules and Regula-
   tions, General Application and Administration. Boston.
 Massachusetts Department of Water Pollution Control. 1980.
   Directory of Massachusetts Lake and Pond Associations
   Westboro.

         1983. Massachusetts Lake Classification Program.
   Westboro.

 Massachusetts  Taxpayers  Association.  1983.  Municipal
   financial data. Boston.

 Massachusetts Water Resources Commission. 1970. Com-
   pilation and Summarization of the Massachusetts General
   Laws,  Special Laws,  Pertinent  Court  Decisions  etc
   Relating to Water and Water Rights. Boston.

 Massachusetts Water Resources Research Center. Various
   Dates. An Inventory of the Ponds, Lakes and Reservoirs of
   Massachusetts. 8 vol. Arnherst, Mass.

 Smith, L  1950. The Great Pond Ordinance—collectivism in
   Northern New England. Boston Univ. Law Rev. 30:178-190.

 Thompson, M. 1981. Maps for America. 2nd ed. U.S. Geoloq
   Surv. 55-61.                                     y'

 U.S.   Department of  Agriculture, Massachusetts Water
   Resources Commission. 1978, 1979. Water and Related
   Land Resources  of Berkshire/Connecticut  Valley/Cen-
   tral/Coastal Regions. 4 vol. Washington,  D.C.

U.S. Department of Commerce. 1981. Census of Population.
  Number of Inhabitants—Massachusetts.  PC80-1-A23. Bur.
  Census. Washington, D.C.
	1982. Census of Population and Housing. Summary
  Characteristics for  Governmental  Units and   SMSA's.
  Massachusetts, PCH80-3-23. Bur.  Census.  Washington
  DC.

U.S. Environmental Protection Agency. 1980. Int. Symp. In-
  land Waters and Lake Restoration, Sept. 9-12,  Portland
  Maine.                                           '

U.S. Geological Survey. Topographic Quadrangles. Various
  dates, various sheets. Washington, D.C.
                                                  490

-------
 MICHIGAN LAKE & STREAM ASSOCIATIONS, INC.
 THREE RIVERS,  MICHIGAN
DONALD WINNE
Executive Director
Three Rivers, Michigan


           ABSTRACT
           The history, objectives, and operation of the Michigan Lake & Stream Associations, Inc., is
           reviewed. This paper details the steps in forming a State association, including the original
           motivation and the legal problems involved.
Twenty-two years ago a few  concerned  property
owners on a half dozen lakes in Michigan met together
to consider common lake  problems.  Out of this
meeting grew a statewide organization of riparian pro-
perty  owners who now  number 40,000 and whose
goals go way beyond those of the first organizers. The
goal of  the first organizers was to develop and im-
prove lake  property,  and many of the  associations
organized at that time were incorporated as improve-
ment  associations.
  With more and more people wanting a place on the
water, inland lakes were filled, dredged, diked,  dam-
med, and changed in many ways to accommodate the
wishes of fishermen, boaters, skiers, and swimmers.
Lakes were being channeled in many  parts of the
State to increase waterfront property and the number
of buildable lots. Lots on both sides of channels were
being sold at a rapid pace. The result of man's impact
on these lakes was their degradation and eutrophica-
tion.
  By  1965, the  Michigan legislature, in an effort to
slow  down the  pace and prevent the destruction of
these inland lakes, passed a law requiring a permit
before further  channels could be dug. In 1969, the
legislature mandated public notices of all requests to
channel lake shores.
  At  the same time that the State of Michigan was
taking steps to regulate lake development and usage,
riparian  property owners also  became  concerned
about water quality and the quality of living on lake-
front  property. Michigan Lake & Stream Associations,
which had grown to include over 100 lake and stream
associations with an individual membership of 24,000,
reviewed and redefined its goals. The following goals
were  adopted in 1980:

OBJECTIVES OF MICHIGAN LAKE &
STREAM ASSOCIATIONS, INC.

   1.  To  inform  riparian property owners and the
public at large of riparian rights in Michigan.
   2.  To disseminate information about pending legis-
lation which will have an impact on riparian rights.
   3.  To inform  riparians of application to dredge, fill,
or  change the shoreline of lakes  and  streams  in
Michigan.
   4.  Sponsor  conferences  and  workshops for
riparians and the public to provide information regar-
ding  the protection of lakes and streams.
   5.  To assist riparians to establish an Association to
deal  with problems which call  for unity in action  to
prevent the degradation of the water quality of lakes
and streams and to prevent their misuse.
  6. To assist  associations  in the presentation of
their respective positions regarding riparian rights and
water resource management before courts, municipal-
ities, and government agencies.
  7. To review and submit proposals to administra-
tive and legislative bodies considering statutes, or-
dinances, and regulations impacting riparian property
owners and water resources.
  8. To develop a library  of information  including
books,  pamphlets, documents, and research studies
of Michigan's water resources  and make the same
available to riparians and the public at large.
  9. To sponsor studies and research designed to ex-
pand the fund of knowledge about  Michigan's water
resources.
  10. To  instruct  lake  and  stream  association
members how to monitor land and water development
within the watershed.
  11. To assist local associations  in obtaining help
from local and State government units in their efforts
to protect their water resource.
  12. To support  all efforts of State and Federal
governments  to maintain  water  quality standards
establised by State and Federal  law.
  These goals have also become the goals of many
riparian property owners and lake and stream associa-
tions who not only want to use and enjoy their invest-
ment in their waterfront property, but more than that,
want to protect the lake or stream for the public and
future  generations to enjoy. Everyone  loses when
lakes and streams become polluted. Fishermen will
not venture out on the lake if the fish  are gone. The
bait  and tackle supplier  cannot  sell  his wares.
Workers are laid off when inventories exceed demand.
When the State's  waters  are  unfit for swimming,
bathing equipment remains on the supplier's shelves
and the quality of lakefront living plummets.
  Who will  protect Michigan's lakes and streams in
the 1980's? With reduced Federal and  State monies,
personnel and activities of State  agencies responsible
for  monitoring and enforcing laws passed to protect
Michigan's natural resources have been cut back. At
the same time, pressures to relax air quality and water
quality standards   have increased. These  develop-
ments  place  the  problems of  Michigan lakes  and
streams in the lap of the riparian property owner. Lake
and stream associations must take center stage if the
threat to water quality is to be stemmed. Associations
that have existed on a very limited annual budget will
                                               491

-------
 LAKE AND RESERVOIR MANAGEMENT
 need to consider raising their dues if they are to have
 the funds needed to protect their lake or stream.
   Michigan Lake & Stream  Associations has  made
 some organizational changes designed to make the
 statewide organization leaders more readily available
 when riparians and lake associations need help. One
 of those changes was to divide the State into  15 in-
 stead of five  regions,  and elect  a vice president to
 work directly with lake associations in his region. Th3
 second innovation was to elect regional directors (on
 the  basis  of one  for each  four lake  association
 members) to assist the vice presidents in their work in
 the  region. Forty-six  regional directors have  been
 elected in 11 of the 15 regions of the State.
  A third step, designed  to  increase the annual ir-
 come of the organization, was to promote  individual
 membership in Michigan  Lake and Stream Associa-
 tions.  The  current  individual  and  corporate  dues
 schedule for membership  is as follows:

 Individual Membership
 The minimum annual dues for individual membership
 in Michigan Lake & Stream Associations is $25.00 ex-
 cept  for members of  a local association  that is a
 member of Michigan Lake & Stream Associations who
 may become members for $15.00.
 Individual memberships are as follows:
 Individual Affiliate Membership	$15.00
 Individual Non-Affiliate Membership	25.00
 Sustaining Donor	50.00
 Benefactor	100.00
 Project Sponsor	500.00
 Lifetime Membership	500.00
 Endowment	  1,000.00

 Corporate Membership
 Associations and corporations that desire to suppor:
 and  promote  the objectives  of  Michigan  Lake &
 Stream Associations may become corporate members
 by paying the following dues:
 Lake & Stream Associations	$25.00 to $180.00
(Based upon number of members in local association
Schedule available on request.)
Commercial	100.00
 Public Corporation (Government Unit)	250.00
Private Corporation	250.00
Endowment 	 1,000.00
  Michigan Lake & Stream Associations hopes that
increased revenue will make it possible to establish a
full-time office and staff so that the organization can
better serve the needs of its members. (Michigan Lake
& Stream Associations, Inc., was authorized a tax ex-
empt status by the IRS under rule 501 C-3 in 1980.)
  The  Michigan  Riparian  magazine  is a  quarterly
devoted to the interests of waterfront property owners
and members of the public interested in using and en-
joying Michigan's water resources. The magazine is a
24-page publication printed and mailed to  individual
subscribers the first of  February, May, August, and
November each year. The circulation has continued
over 10,000 since July 1978. Eighty-four percent of the
issues go to year-round residents of Michigan, and the
remaining 16 percent is mailed to nearly every State of
the Union.
  The magazine  is  published by Donald  E. Winne,
Three Rivers, Mich.,  under  the direction and policies
established by The  Michigan Riparian, Inc., a  non-
profit corporation of  the State of Michigan.  The board
of directors of the corporation has eight members and
is chaired by Joseph H. Hollander, an attorney in the
law  firm of Reid, Reid,  Perry and Lasky of Lansing,
Mich.
  The purpose of the corporation is:
 To gather and  disseminate information concerning
 riparian lands in the State of Michigan; to print, publish,
 and circulate from time to time informational bulletins and
 periodicals concerning riparian lands; to own and operate
 facilities for the printing and distribution of such bulletins
 or  to contract for such  services;  to employ officers,
 editors, and researchers for the purpose of gathering and
 disseminating such information; and to do all other things
 consistent with such purposes.
                                                492

-------
                              Restoration   Techniques
CONTROL OF ALGAL  BIOMASS BY INFLOW NITROGEN
EUGENE B. WELCH
MARKV. BRENNER
KENNETH L CARLSON
University of Washington
Department of Civil Engineering
Seattle, Washington
           ABSTRACT

           Dilution water has entered Moses Lake since 1977, usually in amounts exceeding 100 106 m3,
           which has amounted to exchange rales of around 10 percent day-1 in 8 percent of the lake
           volume nearest the input. The principal cause for the marked reduction in algal biomass and in-
           creased transparency, prior to the Mount St. Helen's ashfall, was the dilution of inflow concen-
           trations of NO3. Average summer chlorophyll a was closely correlated with flow-weighted mean
           inflow NO3. Increased NO3-N:SRP in undiluted inflow water following the ashfall caused P limit
           growth in the lake. Increased dilution water input in 1982, compared to the 2 post-ash years,
           reduced inflow NO3 to biomass-controlling levels in spite of continued high ratios of soluble N:P
           in the lake. Dilution water was pumped to previously undiluted Pelican Horn in 1982 at a rate of
           1.4  m3 s~1 during July through  August. Algal  biomass was reduced largely because of cell
           washout, because undiluted water was pumped after July, but chl a actually increased as a
           result of increased N availability to previously N-deficient cells. Transparency remained at the
           same low levels (0.4 m) that existed prior to dilution due to large amounts of nonalgal turbidity,
           which may not improve even if low NO3 water is added and sewage effluent is diverted.
 INTRODUCTION

 Moses Lake has received low nutrient dilution water
 via the Columbia River irrigation system since 1977.
 Columbia River water provides an ideal dilution of the
 very high inflow nutrient concentrations to this previ-
 ously hypereutrophic lake and as a result water quali-
 ty has improved  impressively  (Welch and  Patmont,
 1980;  Welch  and Tomasek, 1981). Nevertheless, the
 cause for the observed effects of the dilution water
 has been poorly understood and the consistency of
 dilution water releases from the irrigation system has
 been  less than optimal.
   While it  was well  known  that nitrogen limited
 growth rate during  the summer (Welch et  al. 1972),
 phosphorus was assumed to be the long-term limiter
 of biomass in the lake because the principal bloom
 former is Aphanizomenon, an N fixer. Although algal
biomass was observed to be related more to total N
than to total P at TN concentrations below about 600
ng 1~1, TN was  not a good predictor of  biomass
(Welch and Tomasek, 1981). Because NO3 was obser-
ved to deplete dramatically as the inflow water pro-
gressively  moved through' the  lake, flow-weighted
mean  inflow NO3 concentration during spring  and
summer was used as an independent variable to pre-
dict summer average biomass. N  fixation  was as-
sumed to be a poor substitute for scarce N03 because
growth rates from fixation were low. The relationship,
however, did not exist during 1980 and 1981  when in-
flow soluble PO4 concentration declined following the
Mount St. Helen's ashfall, making P the limiting nutri-
ent.
  The restoration  project was enlarged in 1982 follow-
                                             493

-------
  LAKE AND RESERVOIR MANAGEMENT
  ing the installation of a pump and piping system to
  transfer diluted lake water from Parker Horn to previ-
  ously undiluted Pelican Horn (Fig. 1). Unfortunately,
  dilution water input ceased about the time the pump
  was operative so algal  biomass was  reduced in Poll-
  can Horn because of cell washout while chlorophyll
  increased  because of greater availability of NO3 to
  previously N-starved  cells.  Because  of shallowness
  and  nonalgal  turbidity the  transparency  did  riot
  change.
  SAMPLING AND ANALYTICAL METHODS

  Sampling: The data  presented  are  from composite
  samples collected from a depth of 0.4 m along tran-
  sects at stations 7 through 19 in the lake and grab
  samples at the mouth of Crab Creek (4), which repre-
  sents the mixed Crab Creek-dilution on water-Rocky
  Coulee wasteway inflow (Fig. 1). The U.S. Geological
  Survey guaged the Crab Creek flow and the U.S.  Elu-
  reau of Reclamation provided the dilution  water input.
  Base flow of Rocky Coulee was measured directly.
  Transparency was determined with a Secchi disk at
  the transect midpoints.
    Analysis: Specific conductance was used as an in-
  dex of dilution water/lake water*fractions in the lake
  and was determined with a resistance meter (Welch
  and Tomasek, 1981).  Samples for chl a and  soluble
  nutrients were stored on ice and filtered within 6 hours
  after collection. The glass fiber filters for chl a and fil-
  trate (0.45 um Millipore) for soluble nutrients were
  frozen for later analysis. The fluorometric method was
  used for chl a (Strickland and Parsons, 1972). Soluble
  reactive phosphate (SRP) was determined by the adic
  molybdate heteropoly blue method and NO3 by cad-
    Lower Porker Horn (7)
       South Lake (9)
                                      iwer Crab
                                      Creek (4)
         Upper Pelican Horn (19]


      Middle Pelican Horn (II)


  Lower Pelican Horn (10]

^Springs (14)
Figure 1.—Moses Lake basins and sampling transect loca-
tions.
                                mium column  reduction  (Strickland  and  Parsons
                                1972).
                                  Algal  cells were counted in transect samples pre-
                                served with acid Lugols. Counts were made using a
                                Palmer-Maloney cell at 200 power after centrifuga-
                                tion. Filaments were  counted using  Olsen's  (1972)
                                method  and with cell measurements; the results are
                                reported as volumes in mm3 1-1.


                                RESULTS

                                Dilution  water: The release of dilution water into Par-
                                ker Horn from the East Low Canal via Rocky Coulee
                                and Crab Creek has been inconsistent over the past 6
                                years (Fig. 2). Quantities of water were substantial ex-
                                cept during 1980 and 1981, after the ashfall, when total
                                inputs were 34 and 69x106 m3, respectively.  Inputs
                                ranged from 117 to 258x106 m3 during the other 4
                                years. As Figure 2 shows, however, a consistent input
                                   20

                                   10



                                   20


                                   10



                                   20

                                   10


                                   20


                                   10



                                   20

                                   10
                                                          20

                                                          10 -
                                                             1977
1 978
1979
1980
                                                             MAR  APR
                                                                         MAY
                                                                                         AUG
                                                                                               SEP
                                                      Figure 2.—Distribution of dilution water additions during the
                                                      6 years of study. Blocks represent idealized average flows
                                                      during the respective periods.
                                                                        DILUTION
                                80


                              or
                              w

                              < 60


                              U)
                              *
                              <
                              J 40
  •	 7, PARKER
   -- 9, LOWER LAKE
   "- 12, ROCKY FORD
                                    MAR  APR   MAY   JUN   JUL   AUG   SEP


                              Figure 3.—Percent lake  water  remaining at stations 7
                              (Parker), 9 (Lower  Lake) and 12  (Rocky Ford Arm), which
                              together comprise 70 percent of the lake volume, in response
                              to dilution in 1982.
                                                 494

-------
                                                                                 RESTORATION TECHNIQUES
 never occurred through the end  of August. Average
 dilution water input rates of 10 and 20 m3 s~1, when
 added to other normal inflows, produced theoretical
 exchange rates of about 10 and 16 percent day-1 in
 Parker Horn and  1 and 2 percent day-1 in the whole
 lake, respectively.
   A pump on the shore of Parker Horn operated in
 1982 and delivered diluted water to Pelican Horn dur-
 ing all of July and undiluted water during August and
 September. The discharge rate was constant at about
 1.4m3s-i.
   The progressive effect of dilution water throughout
 the lake can be seen in  Figure 3. Lake water is dis-
 placed from  Parker Horn first, followed by displace-
 ment from the Rocky Ford Arm and the Lower Lake at
 about the same time. The importance of consistency
 in water addition is illustrated in Figure 3; once dilu-
 tion water input ceases, the undiluted, high nutrient
 inflow water rapidly replaces diluted lake water. The
 shorter the period of dilution input, the quicker the
 high nutrient water returns, permitting increased algal
 biomass.
   Water quality improvement: In  spite of the inconsis-
 tent pattern  of dilution water input over the past  6
              PREDIL POSTDIL
    1969-  1977   1978   1979   1980   1981    1982
Figure 4.—Mean chl a and transparency during May-Sep-
tember at stations 7 (Parker) and 9 (Lower Lake) in pre- and
post-dilution years.
      69-70    77     78     79     80     81     82

Figure 5.—Mean May-September concentrations of NO3-N
and SRP in transect samples from stations 7 (Parker) and 9
(Lower Lake) during pre- and post-dilution years.
years, the levels to which algal biomass has been
reduced and transparency  increased (compared  to
pre-dilution years) have remained rather similar (Fig.
4). An exception was the considerably lower transpar-
ency and biomass during the ashfall year of 1980.
  Nitrogen as  the controlling nutrient: Dilution has
lowered SRP  more than  NO3.  In  Parker  Horn and
Lower Lake, average concentrations of SRP declined
markedly by the second year after dilution began (Fig.
5) while NO3 declined only slightly. The high NO3 dur-
ing the ashfall  year was an exception. The ration  of
TN:TP in the lake remained rather constant during pre-
and post-dilution years alike, ranging from 7.5 to 10 in
Parker Horn. Thus, it might be suspected that the re-
duction in  SRP caused the biomass reduction. How-
ever, the growing season NO3-N:SRP ratio remained
very similar prior to the ashfall, averaging 2.0 in Parker
Horn and 0.4 in  Lower Lake, clearly indicating that N
was the controlling nutrient. In 1980  the ratio in-
creased to  over 30 and near 5  at  the two stations,
respectively. TN:TP did not change.
  The ratio of soluble nutrients indicated that N was
limiting and NO3 was nearly completely removed as
the inflow from Crab Creek (about 1,000 g 1 -1  without
dilution) moved through the lake (Fig.  5). Adding dilu-
tion water  reduced the inflow concentration because
dilution water averages only 10 to 20 ^g 1-1 NO3-N. If
the lake is considered  a simple chemostat system,
then the algal biomass produced should be a function
of the inflow concentration of limiting nutrient. Thus,
the average chl a  in Parker Horn (7),  as well  as the
volume-weighted mean at stations 7,8, and 9, is close-
                                                        100r
                                                         80
                                                         60
                                                         40
    20
                                                         40
                                                         20
                                                              PARKER HORN
                                                                         1979.
                                                                            1977
                                                                         1982
                                                                                             1969
                                                                                             • 1970
                                                                                            11981
                                        • 1980
                                                             STA 7  8
                                                                             -.79
                                       I 81


                                        • 80
                                                           0      200    400     600    800    1000

                                                              FLOW  WEIGHTED MEAN NO3-N, UG L~ 1
Figure 6.—Relationship between mean flow weighted nitrate
concentration in the Crab Creek inflow from  May through
August and mean chl a in Parker Horn (7) and the volume-
weighted mean of stations 7, 8 and 9 during the period when
surface  water temperature  exceeded  20°,  usually
June-August. Correlation coefficients were both 0.97 and
equations were chl a = 0.092 NO3 - 7.4 and 0.052 NO3 - 3.3
for the two regression lines, respectively, excluding the 3
post-ashfall years (1980-1982).
                                                  495

-------
 LAKE AND RESERVOIR MANAGEMENT
  ly related to the flow-weighted inflow NO3 concentra-
  tion. Diluting the inflow NO3 concentration appears to
  account for the decrease in chl a from the pre-dilution
  levels (Fig. 6).
    P apparently became limiting during 1980 and 1931
  following the ashfall. According to Figure 6, the la
-------
                                                                                RESTORATION TECHNIQUES
creased proportionately more than biomass (Fig. 9).
The chl a:C ratio of algae in both upper and middle
Pelican Horn increased  from  around 0.3 percent
before dilution water  input began to a maximum of
around  3.0  percent in  September.  In  Parker Horn,
values ranged between 1.6 and 3.2 during the summer.
Prior to dilution water input,  algae  in  Pelican Horn
would  have  been  nutrient deficient  judging from
bounds suggested by Healy (1978) of 1.0 to 2.0 percent
for severe to moderate deficiency (conversion of chl
ardry weight to chl a:C were based on a C:dry weight
ratio of 0.2).
  Transparency  in upper and middle Pelican Horn re-
mained the same as during 1969-1970, 0.4 m, in spite
of the markedly decreased biomass. The source of the
turbidity was apparently nonalgal.
DISCUSSION

Inflow volume-weighted NO3 concentration is a rea-
sonably good predictor of average summer algal bio-
mass in Moses Lake. The relationship did not hold in
1980 and  1981, the post-ashfall years, because mar-
kedly reduced SRP concentrations in the Crab Creek
inflow caused P to limit in the lake. In spite of contin-
ued low SRP levels in  Crab Creek in 1982, the  much
greater dilution water input apparently reduced inflow
NO3to biomass-controlling levels.
  Although the  N-fixer Aphanizomenon is the princi-
pal bloom-former in Moses Lake, its growth rate when
dependent on N fixation is  probably insufficient  to
permit utilization of the surplus SRP. Through in situ
N-fixation experiments, estimated growth rates aver-
aged 2.4 + /- 1.8 percent day-1 (Brenner, 1983). Com-
pared  to the average water exchange rate in Parker
Horn  during summer (about  10 percent day-1), it is
reasonable to assume that, for Parker Horn at  least,
biomass increase of even the N fixers depends most
on N03 from the inflow.
   40 r
I    30

 CD
    20 -
 I
 O
    10
                                  OPTIMUM
       25       50        100       150


         10  6 M  3   DILUTION  WATER
200
 Figure 10.—Relation between predicted chl a in Parker Horn
 and dilution water input for May through August based on
 limitation by inflow and equation in Figure 6.
  The cause for the markedly reduced SRP concentra-
tions in Crab Creek following the ashfall, resulting in a
three- to fourfold increase in the NO3-N:SRP ration, is
unknown. About 10 cm of ash fell in the Moses Lake
area and much of that material was transported into
Crab Creek through wind and water erosion. The PO4
sorptive capacity of the ash deposit in the Creek may
have redistributed P between soluble and particulate
fractions.
  The  most optimal use of dilution  water for Moses
Lake  would  be  a  moderate  input rate  from May
through August.  Water added too  early (Feb.-Mar.)
would be largely replaced by high-nutrient Crab Creek
water by June when algal blooms begin. Replacement
with Crab Creek  water is also a problem if dilution
water is stopped  in June or early July.  Unfortunately,
the lack of irrigation demands and  storage  space in
the downstream impoundment has  made the supply
of dilution water undependable during  late  summer.
Nevertheless, Figure 10 indicates that around 100x106
m3 during May through August should control chl a to
about 20 ^g M, whether that quantity came as 10 m3
s -1 for the whole period  or was divided up into 25 m3
s -1 for May and 5 m3 s -1 for June through August.
  The reduced biomass in upper Pelican Horn, and ini-
tially in middle Pelican Horn, was apparently caused
by cell washout.  Although  decreased following pum-
ping of Parker Horn water to Pelican Horn, the bene-
fits were  not realized in  improved water clarity. The
persistent nonalgal turbidity probably results from the
shallowness of this water  body (1.3 m) coupled with
wind and a large abundance of carp. It is not  expected
that transparency will improve when sewage effluent
is diverted in 1985, although the severe N deficiency of
algal cells will no doubt decline.

ACKNOWLEDGEMENTS: This work was supported by a
research grant from the U.S. Environmental Protection Agen-
cy from 1977 to 1980 and through two research contracts
from the Moses Lake Irrigation and Rehabilitation District
and Brown and Caldwell Engineers from 1981 to 1982. Work
is currently supported through  a research contract from
MLI&RD and the Washington State Department of Ecology.
Contributions of Barbara Carey, Clayton Patmont, and Mark
Tomasek to the data set are appreciated.


REFERENCES

Brenner, M.V. 1983. The cause  for the effect of dilution water
  in Moses Lake. M.S. Thesis,  Dep. Civil Eng. Univ. Washing-
  ton, Seattle.
Healy,  F.P. 1978. Physiological indicators  of nutrient defi-
  ciency  in algae. Mitt. Int.  Ver. Theor.  Agnew. Limnol.
  21:34-41.
Olsen,  F.C.W.  1972. Quantitative estimates of filamentous
  algae. Trans. Am. Micro. Soc. 69:272-79.
Strickland, J.D., and T.R. Parsons. 1-972. A  practical hand-
  book of seawaer analysis. Bull.  Fish. Res. Board Can.  167.
Welch, E.B. 1979. Lake restoration by dilution. Pages 133-140
  in Lake  Restoration, Proc.  Nat. Conf. EPA 440/5-79-001.
  4U.S. Environ. Prot. Agency, Washington, D.C.
Welch, E.B., and C.R. Patmont. 1980. Lake restoration by dilu-
  tion: Moses Lake, Washington. Water Res. 14:1316-25.
Welch, E.B., and M.D. Tomasek. 1981. The continuing dilution
  of Moses Lake, Washington. Pages 238-44 in Restoration
  of Lakes and Inland Waters,  Proc. Symp. EPA 440/5-81-010.
  U.S. Environ. Prot. Agency, Washington, D.C.
Welch, E.B., J.A. Buckley, and R.M. Bush. 1972. Dilution as an
  algal  bloom control. J. Water Pollut. Control  Fed.
  44:2245-65.
                                                  497

-------
  METHODS AND TECHNIQUES OF: MULTIPLE PHASE DRAWDOWN
  FOX LAKE, BREVARD COUNTY,  FLORIDA
 ROBERT J.  MASSARELLI
 Brevard County Water Resources  Department
 Merritt  Island, Florida
             ABSTRACT

             Multiple phase drawdowns have been suggested as a possible restoration technique for controlling
             the aquatic weed Hydrilla and for consolidating sediments. Brevard County, in cooperation with the
             Florida Game and Freshwater Fish Commiss on, implemented such a program in 1979-80. Fox Lake
             is a small 44.5 ha (110 acre) freshwater lake in Brevard County on Florida's east coast This lake
             the location of a major regional park, had become unusable to boaters and fishermen due to an ex-
             cessive growth of Hydrilla. In addition to the Hydrilla, the lake had minimal fish and wildlife benefits
             due to a thick layer of unconsolidated muck. While the use of proper technique is important, the restora-
             tion of Fox Lake required methods which insure full community support, and the cooperation of other
             agencies and local elected officials' and public involvement are necessary. Techniques must be flexi-
             ble enough to meet unforeseen or changing conditions. For example, during the Fox Lake drawdowns
             lake conditions required innovations such as air boat pull plows and amphibious craft. The Fox Lake
             project demonstrated that lake restoration projects with maximum and innovative use of local resources
             can be completed with minimal impacts on  local government budgets.
 INTRODUCTION

 Fox Lake is a relatively small freshwater lake in Bre-
 vard County west of the city of Titusville. A popular re-
 gional park is located there as is a fish  management
 area of the Florida Game and Freshwater Fish Corn-
 mission. During the 1970's the recreational and wild-
 life value of this lake declined because of almost com-
 plete coverage of the lake by Hydrilla (Hydrilla vertidl-
 lata) and an unconsolidated muck bottom. In 1978, a
 multiple phase drawdown was proposed to control the
 Hydrilla and consolidate the bottom sediments.
   In Florida, Hydrilla reproduces from tubers, turions,
 and plant fragments. Haller et al. (1976) report that the
 formation of these propagules is seasonal. Haller also
 describes the response of Hydrilla  to  water level
 manipulation.  Based on this work a seasonally-timed
 multiple phase drawdown schedule was developed for
 Fox Lake.
   Besides the Hydrilla,  Fox Lake's  unconsolidated
 sediments were also a problem. Therefore, another ob-
 jective of the project was to consolidate the bottom of
 the lake. The dewatering of the lake was expected to
 consolidate and dry out its sediments. Studies at Lake
 Apopka in  Central  Florida  have found  that drying
 muck sediments results in a water loss and shrinkage
 (Foxet al. 1977). Dried sediments remain consolidated
 for long periods of time following refills.
  The  proposed multiple phase drawdown was also
 expected to improve the lake's wildlife habitat. Several
 studies have recognized the need for  water level fluc-
 tuation  for  fisheries management.   Holcomb  and
Wegener (1971) and Wegener and  Williams (1974)
 showed improved invertebrate and fish population fol-
 lowing lake level manipulaton in Florida.
  This paper discusses the techniques and methods
 used in a multiple  phase drawdown  of  Fox Lake  in
 Brevard County, Florida.
  Fox  Lake is a  small 64 ha  lake in North  Brevard
County on Florida's east coast (see Fig. 1). A popular
regional park is located there. Although a large alliga-
 tor population prohibits swimming  and waterskiing,
 boating, canoeing, and fishing are  encouraged. The
 basin is dominated by a cattail (Typhia latifolia) fringe
 marsh and slash pine flatwoods. Three small oak hum-
 mocks  are along the edge of the lake.
   Recently, development pressure is starting to ap-
 pear on the east side of the lake because of the contin-
 ued growth of Titusville and the North Brevard County
 area. This development has taken the form of  large lot
 (1 acre  +) residential development.
   Fox Lake is in the same basin as South Lake, a 440
 ha lake, just  north of  Fox Lake. Historically, these
 lakes were connected  by a cattail marsh. In 1962 a
 canal was dredged between the two lakes.
Figure 1.—Location map.
                                                498

-------
                                                                              RESTORATION TECHNIQUES
   In the 1970's Fox Lake was a relatively shallow lake
with an average depth of 1  meter. The lake's bottom
was sand covered by a layer of unconsolidated muck.
The average  depth  of the  muck was .5 meter. The
source of the muck was unknown; however, it appears
to have been the result of several years of chemical
control of the cattails that dominate the lake's shore-
line. While South Lake had a sewer plant discharge for
several  years ending in  1973, none  existed in Fox
Lake. The unconsolidated nature of the lake's bottom
is undesirable from a fisheries management aspect.
   Hydrilla was the dominant  vegetation within the
lake. Vegetation surveys by the Florida Game and
Freshwater Fish Commission in 1978 indicated that
Hydrilla surfaced in approximately 22.8 ha (McKinney
and Coleman, 1981).  The nature of the Hydrilla made
recreational uses of the lake impossible. Canoes or
airboats were the only possible way of navigating the
lake. The edge of the lake  was dominated  by Naiad
(Najas guadalupensis) and cattail (Fig. 2).
 DEWATERING PROGRAM

 The configuration of the South Lake/Fox Lake drain-
 age basin  made  it ideal  for  a  multiple  phase
 drawdown. Because of its close proximity to  Fox
 Lake, South Lake could be used as a reservoir to pump
 water from Fox Lake. In addition South Lake could be
 used as a source of water to refill Fox Lake at the end
 of the drawdown. The manmade canal facilitated the
 movement of water. Finally, the undeveloped nature of
 the basin would minimize possible complaints during
 the drawdown.
               To accomplish  the  multiple-phase drawdown, an
             earthen dike was constructed across the canal. Fill for
             the dike was obtained from the spoil remaining from
             the dredging of the canal. A metal culvert was in-
             stalled in  the dike so that reflooding could be ac-
             complished by gravity flow. A lift gate structure was
             fitted to the metal  culvert to prevent reflooding during
             dewatering. A 38,000-liter-per-minute centrifugal pump
             (Crissafulli Model CP-16) was used to pump water over
             the dike (Fig. 3). The canal, which was deeper than Fox
             Lake, would act as a sump for the pump.
               The proposed drawdown schedule required a mini-
             mum of two dewaterings with an optional third (Fig. 4).
             The first dewatering was timed to occur in the spring
             to kill the exposed parts of the adult Hydrilla plant and
             to consolidate the bottom sediments. In addition, this
             would  stimulate germination  of the Hydrilla tubers
             and turions.
               The second dewatering, scheduled for the fall, was
             timed to kill the newly sprouted Hydrilla prior to the
             production of new tubers of turions that occurs from
             October to April (Haller et al. 1976).
               A third optional  dewatering was scheduled for the
             following spring to kill the remaining Hydrilla plants
             that might result from late tuber or turion germination.
             DISCUSSION

             A successful lake restoration activity requires close
             coordination and cooperation between all interested
             parties. Interagency cooperation will improve the suc-
             cess of the restoration project and reduce impacts on
             each  agency's  budget.  Although  several  agencies
            WETLANDS


            HYDRILLA
t
 N
     [    )  NAIAD/CATTAILS
                                                                               EARTHEN  DAM
                                                                               8 PUMP  SITE
 FROM  MCKINNEY  a  COLEMAN, i98i

Figure 2.—Fox Lake vegetation map before dewatering.
ALTERNATIVE
PUMP   SITE
             Figure 3.—Dewatering facilities.
                                                499

-------
LAKE AND RESERVOIR MANAGEMENT
were involved  in the Fox Lake restoration project,  it
was primarily a cooperative effort between the Florica
Game and Freshwater Fish Commission and the Bre-
vard County Board of County Commissioners. The
Game Commission  suggested that a multiple phase
drawdown could be used in Fox Lake to control the
Hydrilla, consolidate the bottom, and improve the fish-
eries habitat. Each agency's contribution to the pro-
ject complemented each other.
  The Game Commission provided the technical as-
sistance of their fisheries and aquatic weed biologists
in the design and operation of the project. The Game
Commission staff  conducted vegetation surveys of
the lake to determine the extent and distribution of the
vegetative coverage. The centrifugal pump was pur-
chased by the Game Commission and following the
completion of  this project, the pump would be made
available for other Game Commission projects. Exis-
ting equipment of  the Game Commission was  also
made available.  For example, a metal  shed used in
dove hunting management was loaned to the project
to provide a shelter. To meet the dewatering's 24-hour-
a-day pumping schedule, volunteer members of the
Commission auxiliary manned the  pump during the
weekends and  nights.
  Several departments within the Brevard County gov-
ernmental organization provided support to the restor-
ation project. The Brevard County Planning and  Zon-
ing  Department's Environmental Planner coordinated
the  project and processed the necessary permits. The
County's Water Resources Department staff provided
technical assistance. This included bathymetric map-
ping, groundwater evaluation and monitoring, and ne-
cessary labor.
  The District  1 Road and Bridge Department provid-
ed a great deal of assistance to the project, including
the  necessary  heavy equipment to  construct the ear-
then dam and other earth-moving work, as well as the
necessary tractors  and fuel  used to drive the pump.
County staff manned the pump during the day. A por-
table generator  and radio was provided  to permit
24-hour-a-day operations.
  Other governmental agencies  also provided assis-
tance. The St.  John's River Water  Management Dis-
trict took areal photographs during overflights of the
lake. The  U.S.  Geological  Survey provided sta:f
guages for the  lake level measurements.
  Another group that provided a great deal of assis-
tance was the general public. A nearby landowner pro-
vided access through his property to the dam pump-
site. In addition, two of his wells were made available
for  groundwater  monitoring. Another individual  pro-
vided food and drinks to the workers.
  The first dewatering began Feb. 12, 1979. Dewater-
ing  was completed in 8 days and lasted approximately
11 weeks. The second drawdown, delayed due to Hur-
ricane David and heavy precipitation, began  Oct. 20,
1979. This dewatering lasted 5 months. Thirty days
later, a third drawdown commenced. The last dewater-
ing  lasted 30 days.
  During the implementation of the restoration sever-
al problems became apparent,  necessitating some
flexibility. The  major problem encountered was that
because of the relative flatness of the lake, it became
very difficult to move water by gravity to the pump. At
first this problem was addressed by manpower.  Indi-
viduals with shovels were used to dig ditches to drain
pockets of water. This, however, proved ineffective be-
cause of the long distance to the pump used for the
dewatering.
   Since the bottom was composed of an unconsoli-
 dated  muck, it was hypothesized  that if a shallow
 channel could be formed to the dewatering pump, the
 flow of water would keep the channel open. The Coun-
 ty's District 1  Road and Bridge office has a LARC, a
 lighter amphibious resupply vehicle. The LARC is a
 4,545 kg aluminum-hulled vehicle currently  used for
 rescue. It is propelled on land by four large tires (18:00
 x  25) or in water by a propeller.  The 3-m-wide  boat-
 shaped hull was expected to create the necessary
 channel. However, when the LARC was  put into the
 lake, it quickly became bogged down in the muck.
   An alternative method was developed to provide ac-
 cess through the muck. A v-shaped plow, 91 cm long
 and 76 cm at the widest point, was fashioned out of
 scrap metal.  This plow was then attached to an air-
 boat that  pulled the plow back and forth across the
 lake.  This  method had limited success. The  plow
 would cut a shallow channel which  would be partially
 filled  by adjacent  sediments. While these methods
 provided some  additional  dewatering  capabilities,
 some areas of the lake did not achieve  the desired
 levels of dewatering.
   A second unexpected problem was the existence of
 large  holes located near the  shoreline of Fox  Lake
 Park. Apparently during the development  of Fox Lake
 Park, dredging occurred within the lake to  build up the
 park. These holes were twenty to thirty meters  in di-
 ameter and 1.5 to 2 m deep.
   When the holes were first discovered manpower
 was used  to connect the holes. In  addition, a small
 mud pump was used to move the water from one to
 another. These techniques  proved  to be ineffective
 and time-consuming. Therefore, the County provided a
 dragline to dig a channel to connect the holes and ex-
 tend south to Fox Lake Road. The pump was  then
 moved to an alternative site on Fox  Lake  Road where
 the water was pumped south into an isolated wetland
 area (Fig. 3). The pump was moved between both  pum-
 ping sites on a daily basis. This helped the dewatering
 of the southern area of Ihe lake.
   A third  problem  experienced was Mother Nature.
 During the period between the two dewaterings, Hurri-
 cane David passed  through the area. While David was
 relatively small, it was typical of the high preciptation
 experienced that summer. The Labor Day hurricane
 v>
g
£
    16
    14-
   *
 UJ 12-
         JMMJSNJMM

        I	1979	1— 1980 —
    	  PROPOSED


    	ACTUAL  (2nd DEWATERING)

Figure 4.—Dewatering schedule.
                                               500

-------
                                                                                RESTORATION TECHNIQUES
forced postponement of the second dewatering until
October.
  Returning to normal water levels at the end of the
dewatering schedule  was very important. Fox et al.
(1977)  had shown that cattails and water hyacinth
(Eichhornia) sprouted on drying sediments. Cattails
were the only vascular plant which survived following
flooding of dried sediments. Because of the existence
of the cattail  fringe  marsh, high  water levels were
needed to prevent the movement of cattails into the
lake. At the end of the dewatering, Florida began a per-
iod of severe drought.
SUMMARY AND CONCLUSIONS

The  multiple  phase  drawdown was successful  in
meeting two objectives: the Hydrilla was virtually elim-
inated from the lake, and tuber and turion abundance
was  reduced.  Sediment consolidation, while not as
successful as hoped, occurred throughout the lake.
The sediments remained consolidated after reflooding
(McKinney and Coleman, 1981).
   The lake, however, did experience a tremendous re-
sponse in cattails. During the project, the U.S. Environ-
mental Protection Agency prohibited use of the herbi-
cides commonly used on cattails in the area. This, plus
the low lake level experienced at  the end of the de-
watering schedule allowed cattails to virtually cover
the entire lake (Fig. 5).
   The Fox Lake restoration project was not an experi-
ment or demonstration project. Its purpose was to re-
    fcjXJj WETLANDS


    JUS EELGRASS/OPEN WATER
t
 N
    (    ]  CATTAILS

Figure 5.—Fox Lake vegetation map after dewatering.
store the recreational and wildlife values of the lake.
The  Fox Lake Project does point out three important
considerations that are necessary for a successful res-
toration project. The first is strong local support. Nu-
merous papers have stated the need for public partici-
pation and support. The Federal Clean Lakes Program
(Section 314, P.L. 92-500) requires public participation.
Fox  Lake restoration was successful because locally
it had the support and commitment to follow through.
The County Commissioner from District 1 became per-
sonally involved with the project. With his support, the
Board  of County Commissioners committed the ne-
cessary county resources.
  The second consideration is intergovernmental sup-
port. Today,  public fiscal environment demands the ef-
ficient utilization of the tax dollars. Most budgets have
little room for expensive or time-consuming restora-
tion projects. In addition, at the local level, where there
may be strong support for the restoration project, local
governments may not have the necessary expertise in
restoration  techniques. Intergovernmental  coordina-
tion  minimizes the fiscal impact of a restoration pro-
ject. Involving several agencies spreads the costs. In-
ternally, the involvement  of several departments or
sections within an organization spreads costs among
the various budgets. While the total cost to the public
funds is still the same, this process allows the elected
official to accept the costs. Intergovernmental coordi-
nation also  allows for the efficient utilization of tech-
nical expertise. Many local governments do not have
the need or fiscal resources for a full time aquatic bot-
anist. A State agency can provide this expertise. An
additional benefit is the favorable public relations an
agency receives. Often an elected official will begin to
understand  for the first time an agency's value. This
will improve local support of the agency's budget at
budget time.
  The  last consideration  is flexibility. Apply Murphy's
Law  to any project design. Unexpected situations will
occur during the project. (Strong local support allows
for flexibility). The loss of the  herbicides needed to
control the cattails is an example of what can happen
when flexibility is lost. While not originally planned, if
the herbicides had been available, the project was flex-
ible enough  to meet that need. Successful restoration
will require flexibility in the project  and  project  man-
agement.


REFERENCES

Fox,  S.L, P.L. Brezonik, and M.A.  Kerin. 1977. Lake Draw-
  down as a  Method of Improving Water Quality. U.S. En-
  viron. Prot.  Agency.
Haller, W.T., J.L Miller, and L.A. Garrard. 1976. Seasonal Pro-
  duction and germination of Hydrilla vegetative propagules.
  J. Aquat. Plant Manag. 14:26-9.
Holcomb, D., and W. Wegener. 1971. Hydrophytic changes re-
  lated to lake fluctuation as measured by point transects.
  Proc. S.E. Assoc. Game Fish Comm. 25:570-83.
McKinney, S.P., and W.S. Coleman. 1981. Hydrilla control and
  vegetation response with multiple dewaterings. Presented
  at the 35th  Conf. S.E. Assoc. Fish. Wildl. Agencies.
Wegener, W.,  and V. Williams.  1974.  Fish population re-
  sponse to improved lake habitat utilizing an extreme draw-
  down. Proc. S.E. Assoc. Game Fish Comm. 28:144-61.
                                                  501

-------
 RESTORATION  OF SEBASTICOOK LAKE,  MAINE,  BY
 SEASONAL  FLUSHING
 CHET ROCK
 Department of Civil Engineering
 University of Maine
 Orono,  Maine

 DAVID  COURTEMANCH
 Maine Department of Environmental  Protection
 Augusta, Maine

 THOMAS  HANNULA
 Department of Mathematics
 University of Maine
 Orono,  Maine
            ABSTRACT

            During the past century, increased fertilization and the resultant algal and macrophyte growths have
            severely impaired the quality and use of many lakes throughout the world. A notable case has been
            the deterioration of Sebasticook Lake. The laks began showing signs of cultural eutrophication in the
            early 1950's. By the late 1960's it had become hypereutrophic and has remained in that condition
            despite remedial efforts. Currently a major effort to restore the lake has been undertaken by the State
            of Maine, U.S. Environmental Protection Agency, U.S. Department of Agriculture and local communities.
            Estimates of annual external phosphorus loaoing range from 3,900 to 11,800 kg with an estimate of
            9,000 kg considered as the most accurate estimate of the current load. Sources include 2,300 kg
            and 2,200 kg from the towns of Dexter and Coiinna respectively, and 4,500 kg from nonpoint sources
            of which 85 percent is culturally generated from farmland. Using a mass balance model it is estimated
            that an annual external load of 4,500 kg of total phosphorus would maintain the lake at a desired
            concentration of 15 ^g/l and suppress the chronic algal blooms. Strategy to control external sources
            includes advanced wastewater treatment to reduce the loading from Dexter and Corinna to 300 kg
            and 700 kg, and improved farm practices to reduce nonpoint sources to 3,500 kg. Since the recycle
            of phosphorus from the lake sediments is estimated to contribute 6,900-9,900 kg into the water col-
            umn annually, control of internal recycling was also imperative. Because of the large size of Sebasticook
            Lake (1,798 ha) the most promising means to reduce internal phosphorus was to alter the flushing
            regime of the lake. Epilimnetic phosphorus reaches peak concentration during late summer stratification
            At that time, the lake volume is  reduced by one half, decanting the phosphorus rich epilimnetic water.
            Refill of the lake does not commence until the following spring when phosphorus-poor melt water is
            available. Drawdown of the lake is accomplished through constructing a 4-meter deep canal and gate
            structure at the outlet. Initial trials of the structures were estimated to export about 4,200 kg of phosphorus
            annually.
INTRODUCTION

During  the  past  century,  increased  fertilization
resulting in algal and macrophyte  growth  has im-
paired the use of many lakes throughout the world.
One notable case  has been  the  deterioration of
Sebasticook Lake in Newport, Maine. First studied in
1965 and reported in a technical report (Fed. Water
Pollut. Control Admin., 1966) and by Mackenthun et al.
(1968), the lake began showing signs of cultural eutro-
phication in  the  early fifties.  By  the late  sixties,
Sebasticook Lake had become hypereutrophic and
has remained in that condition despite remedial ef-
forts.  Currently, a major effort to restore the lake has
been  undertaken  to  limit external  sources  of
pollutants and alleviate the effects of years of  sedi-
ment accumulation.
BACKGROUND

Limnology

Sebasticook Lake is located in Newport, Maine, on the
East Branch of the Sebasticook River (Fig. 1). The lake
receives  runoff from three main  tributary  streams:
East Branch of the Sebasticook River, which is the
largest and  contributes about half the total inflow;
Mulligan  Stream;  and Stetson Stream. Mean annual
precipitation is 102 cm. The morphometric data for
this relatively large, shallow lake are given in Table 1.
  Lake water quality reflects conditions associated
with  advanced  eutrophication (Table  2). The phos-
phorus levels are well  above the  15  /^g/l needed to
stimulate phytoplankton in Maine water with spring
turnover values  in the range of 30 to 40 /*g P/l. Conse-
                                                502

-------
                                                                               RESTORATION TECHNIQUES
quently,  the  chlorophyll a values are substantially
above the 8 /^g/l concentration used as a guideline of
eutrophic conditions for Maine waters. Similarly, the
Secchi disk measurements are very low  because of
the high productivity. The dominant summer blooming
algal species are Aphanizomenon flos-aquae and Ana-
baena spp.
   Dissolved oxygen is  depleted  in the thermocline
and hypolimnion  during  stratification, and  during
calm weather conditions, anoxia may extend as much
as 2 meters into the epilimnion. Bottom conditions are
such that only benthic organisms tolerant of low ox-
ygen concentration can survive.  Nolan and Johnson
(1975) recorded midgefly larvae (Chironomus riparius)
and worms (Tubificidae). In recent years, only epiben-
thic Chaoborus could be found in the profundal area
of the lake in late winter because of the anoxic condi-
tion.

Cultural Development

Although located  in Newport, that town has relatively
little influence on the water  quality of Sebasticook
Lake as it is situated at the lake outlet and its waste
discharges are  downstream   of  the lake (Fig.  1).
Development along the lakeshore  includes about 250
residences,  mostly summer  cottages (200).  Other
population centers in the  watershed include  Dexter
(1980 population 4,236) and Corinna (1980 population
1,887). Both are located on the East Branch of the
Sebasticook River upstream of the lake and discharge
their wastes  into the East Branch. The wastewater
from Dexter is currently discharged untreated, while
Corinna has a secondary treatment plant.
  The watershed is primarily  rolling hills of mixed
softwoods and  hardwoods with about 20 percent  of
the land area devoted to agriculture. Sixty-seven farms
are located in the  drainage basin with dairy, livestock,
potato, and poultry production being the principal ac-
tivities (Penob. Valley Reg. Plan.  Comm.,  1980). Agri-
culture has been  identified as a major contributor of
nonpoint source nutrients to the lake.
  At one time, industries were the largest source  of
nutrients to Sebasticook Lake. A potato  processing
plant located in Corinna processed about 160,000 kg
                        of potatoes a day 9 to 10 months of the year until go-
                        ing out of business  in 1968. Two woolen mills are
                        located in Dexter and a third in Corinna. In  1966 the
                        phosphorus discharge of Corinna was approximately
                        5,200 kg P with 66 percent from the potato processing
                        plant, 16 percent from a woolen mill, and only 18 per-
                        cent from domestic and other waste sources (Keup,
                        1968). One of the Dexter mills has now converted to a
                        dry operation and the other has closed. The Corinna
                        woolen mill now treats its wastewater jointly with the
                        town.

                        HISTORY

                        Previous Studies

                        While the first major limnological study of Sebasti-
                        cook Lake was undertaken  in 1965, earlier surveys
                                  Newpo
                       Figure 1.—Sebasticook Lake watershed.
                              Table 1.—Morphometric data for Sebasticook Lake.
 Surface area
 Mean depth
 Drainage area
1798 ha
6.2 m
32,600 ha
Volume
Maximum depth
Hydraulic retention time
111 X1
18.2m
0.58 yr.
                              Table 2.—Water quality data for Sebasticook Lake.
Date
1965
1972
1973
1975
1976
1977
1978
1979
1980
1981
1982
min
0.8
0.8
—
.08
0.6
1.3
0.7
1.1
0.6
1.2
0.9
Secchi disk (m)
mean
1.4(4)*
1.2(3)
—
1.9(10)
1.7(28)
1.6(3)
0.9 (3)
1.8(8)
1.8(9)
1.9(13)
1.4(6)
Total Phos.
mean (^g/l)
55 (15)*
73(11)
50(2)
—
60 (22)
60 (36)
40(9)
—
—
46 (36)
64 (33)
Chi a
mean (p
-------
 LAKE AND RESERVOIR MANAGEMENT
 around  1949 were conducted  by the  Maine Depart-
 ment of Inland Fisheries and Wildlife.  These surveys
 documented a change in fishery from trout and smelt
 to  a warmwater fishery,  a  characteristic  often
 associated with eutrophication. In  the  1965 study,
 Sebasticook  Lake was reported as "plagued  with
 nuisance  algal  growths."  The eutrophication could
 hardly come  as a surprise, based on the conditions
 reported by the  Federal scientists.
   The reach of stream from Corinna to the inlet of Lake
   Sebasticook was deplorably  polluted. The  river sup-
   ported a luxuriant growth of aquatic slimes and contain-
   ed several 'log-jams' of trash from the  dump including
   discarded footballs, dolls, and barrels. The banks of the
   stream were spongy with a mat of wool fibers that has
   accumulated through time.  Occasionally intermingled
   with the fiber were potato sprouts and rotting potatoes.
   The area was revolting to  both the human eye and
   nostril. Proceeding downstream, a mat  of floating wool
   approximately 6 inches  thick and  400  feet long com-
   pletely covered the river.  Boat navigation was complete-
   ly stopped by this 'wool-dam' and birds walked on it as
   conveniently as on land (Fed. Water Pollut. Control Ad-
   min., 1966).

   Not surprisingly,  blue-green  algae  dominate the
 phytoplankton found in the  lake. Mackenthun et al.
 (1968) reported summer blooms dominated by Micro-
 cystis aeruginosa and Anabaena spp.  The technical
 report concluded that  indeed  the lake  was hypar-
 eutrophic, primarily because of the input  of domeslic
 and  industrial  wastes.  They  predicted  that  alcial
 blooms would continue to exist until the lake waler
 phosphorus  concentration  could be  reduced.  Tie
 report recommended 0.02 mg P/l  as  a goal based on
 the concentrations found in Wassookeag  Lake, the
 headwaters of the East Branch  of the Sebasticook
 River.
   The second major study was also conducted by the
 Federal  Government  by Nolan  and  Johnson (197'5)
 from November  1971 to August 1973. They found in-
 creased  phosphorus levels  in both Sebasticook Lake
 (approximately 0.10 mg/l, up from 0.05 mg/l in 1965)
 and its major tributary, the East Branch (approximate-
 ly 0.15 mg/l, up from 0.07 mg/l in 1965). Excessive algal
 growths  were also noted at concentrations similar to
 1965 (chlorophyll a of 11.3 ^g/l versus 11.6 ^g/l in 19615).
 Although low-oxygen-tolerant benthic fauna such as
 midge-fly larvae and tubificid worms continued to ex-
 ist, they did so at significantly lower populations than
 found in  the 1965 study.
   Nolan  and Johnson (1975) concluded that Sebasti-
 cook Lake was  hypereutrophic and the phosphorus
 loading from  the  East  Branch was  still  excessive,
 despite the elimination of the potato processor which
 had been a major contributor of phosphorus prior :o
 1968. Nolan and Johnson also identified the lake sedi-
 ment as a  significant  phosphorus reservoir  and
 predicted that it would  serve as a major source for
 many years to come.  Previously,  Mackenthun, et al.
 (1968) had not recognized the sediments as a source,
 but rather  expected them to serve as  a sink whe>n
 nutrient inputs were controlled.
   Sebasticook Lake was also studied in 1972 as part
of  the  National  Eutrophication  Survey  (1974)  in-
dependently of the Nolan and Johnson study. The  lim-
nological data gathered basically agreed  with  the
previous  studies, i.e.,  Sebasticook Lake  was highly
eutrophic. In addition, an algal assay  was performed,
indicating the lake was nitrogen limited at the time of
sampling. While nitrogen may be rate  limiting, it does
not mean that nitrogen is supply limiting as it can be
 obtained directly from the  atmosphere by  nitrogen-
 fixing blue-green algae.

 Nutrient Budgets
 Estimation of the first nutrient budget for Sebasticook
 Lake was one of the major accomplishments of the
 1965 study (Fed. Water Pollut. Control Admin., 1966).
 However, time has proven it to be rather crude and pro-
 bably in significant error. The budget was based on
 quarterly 1-week sampling  periods.  As a rule, com-
 posite samples were taken, except grab samples were
 made where  water quality did not fluctuate markedly.
 Stream flows were based on actual measurements at
 the time of sample collection.
   For the Dexter area, most of the nitrogen was  con-
 tributed by the  two woolen mills (65 to 85  percent),
 while practically all of the phosphorus came from
 domestic wastewater (87 to 93 percent). However, a
 sampling station (Lincoln Mills) 4 miles below Dexter
 indicated significant nutrient reduction  had  taken
 place in the stream reach below Dexter (about 20 per-
 cent for nitrogen and 30 percent for phosphorus). In
 the Corinna  area, the  woolen mill was the  principal
 contributor of  nitrogen (85  percent) whereas  the
 potato processor was the main source of phosphorus
 (55 percent) (Mackenthun et al. 1968). Nutrient sources
 based on the data presented  in  the  technical  report
 are itemized  in Figures 2a and 3a.
   The technical report concluded that an  80 percent
 reduction in the waste inputs of Dexter and Corinna
 could reduce the loading to about 1,590 kg P/yr. Even if
 this were  accomplished, the authors  predicted  it
 would take 10 years before the lake would be restored.
 It  should be  noted that authors labored under the
 misconception that the hydraulic retention time  was
 3.5 years. More recent computations  have shown the
 retention time to be 0.58 years.
   The nutrient  budget developed for the National
 Eutrophication Survey was  based on monthly grab
 samples  collected from  September 1972  through
 August  1973,  except that biweekly sampling was done
 during April and May. Mean stream flow estimates
 were  provided by the U.S.  Geological  Survey.  The
 nutrient loadings for the Dexter area were based on
 literature coefficients rather than sampling data.  The
 loadings for Corinna, however, were based on actual
 sampling. Although Keup (1968) reported that 29  per-
 cent of  the phosphorus was assimilated by the river,
 this budget assumed that all nutrients reached  the
 lake. The survey showed that 11,800 kg P and 283,200
 kg  N  entered  Sebasticook Lake on an annual  basis
 (Figs. 2b and 3b). Nutrient discharge from the lake was
 measured at 8,350 kg P/yr and 210,300 kg N/yr so that
 phosphorus accumulation  was  estimated at  3,450
 kg/yr and the  nitrogen  level  in the lake increased by
 72,800 kg/yr.  Thus, 8 years  after the  Federal Water
 Pollution Control study recommended an 80 percent
 reduction in phosphorus loading, the input to Sebasti-
 cook  Lake seemed to have  increased, despite  the
elimination of the single largest source, the potato
 processor.
  The National Eutrophication Survey concluded that
at  least a  70 percent  reduction  in  the phosphorus
 loading  from Dexter and Corinna would be required to
significantly improve the trophic condition of Sebasti-
cook  Lake. It noted that an even higher  level of
removal would be necessary to compensate for  the
phosphorus already in the lake.
  The most definitive phosphorus budget was  that
prepared by the Maine  Department of Environmental
                                                 504

-------
                                                                                RESTORATION TECHNIQUES
Protection  (Dennis and Corson, 1979); however, they
did not develop a nitrogen budget. Stream samples
were  collected at 11 stations on a biweekly basis.
Eight nonrecording gauging stations were established
by the USGS, while other flows were established by
the proportional  drainage area  technique  (Morrill,
1975). Mean seasonal phosphorus concentrations and
hydrograph  discharge   estimates  were  used  to
calculate monthly and annual loadings.
  The total  external phosphorus loading was  cal-
culated to be 9,000 kg P from June 1975 through May
1976.  During the same period only 5,500 kg  P were dis-
charged from the lake.  Dennis  and Corson  (1979)
noted that the usual fall drawdown of the lake did not
occur; that would normally have  resulted  in another
3,000 kg P or more being discharged.  The nonpoint
fraction was estimated at about 50 percent based on
an areal export of 18.3 kg/km2. Like Mackenthun et al.
(1968), Dennis and Corson found  some loss of phos-
phorus downstream of Dexter's discharge to a small
impoundment. They estimated this loss at  1,200 kg P.
They also estimated a 500 kg loss of nonpoint phos-
phorus (Fig. 2c).
  In addition to estimating the external phosphorus
loading, Dennis and Corson  calculated the internal
phosphorus loading. During the summer of  1976, inter-
nal loading was estimated through mass  balance
equations to be 9,900 kg P and in 1977 it was 6,900 kg
P. They observed that the summer  internal loading
nearly equaled the annual external loading and that it
is recycled each year. The magnitude of the phos-
phorus source supported  the earlier findings of Nolan
and Johnson (1975)  that the sediments could be a
significant  source of phosphorus for many years.
   In summary, several attempts have been made to
estimate phosphorus and nitrogen loading to Sebasti-
cook Lake. The first attempts (Fed. Water Pollut. Con-
trol Admin., 1966; Mackenthun et al. 1968) have been
found to be in error, specifically the water balance. It
is highly doubtful that the  phosphorus  input to
Sebasticook Lake increased from 3,900 kg/yr to 11,800
kg/yr between 1966 and 1974; rather, it  is likely that
eliminating the 3,500 kg/yr contribution of the potato
processor resulted in  an  actual reduction by  1974.
Despite  this reduction, lake  productivity remained
high because of internal phosphorus recycling. The
phosphorus budgets by the National Eutrophication
Survey (1974) and Dennis and Corson (1979) are actual-
ly quite similar, especially if  1,200 kg P assimilation
loss in the Dennis and Corson budget is applied to
both. Then,  the difference  is only  15 percent which
would be within natural year-to-year variation. The ma-
jor difference is the amount of phosphorus discharged
by the lake, but normal drawdown did not occur in
1979. As  noted by Dennis  and Corson, drawdown
would normally result in an additional 3,000 kg being
discharged,  making the two budgets nearly identical.
  Since  only two  nitrogen budgets  have been cal-
culated and the Mackenthun et al. (1968) budget is
suspect,  it is most difficult to assess the accuracy of
the nitrogen loading. Certainly, the nitrogen loading
has decreased since 1966 as the major sources, two
woolen mills, have either closed, switched to dry pro-
cesses, or treat their wastes. While nitrogen has been
shown  to be limiting  in the short-term, phosphorus
control is believed to be the key to long-term recovery.
Consequently, the major restoration effort has focus-
ed on phosphorus dynamics.
 (a)  Anon  (I966)

(500)
1
(3300)
EAST BRANCH
1

I =• 3900

EAST BRANCH
                                                     RESTORATION

                                                     Despite the criticism of the nutrient budgets of the
                                                     first Federal Water Pollution Control Administration
                                                     study, the report was right in regard to restoration. The
                                                     report noted:

                                                      Prerequisite to any efforts directed toward cleanup of
                                                      Sebasticook  Lake  is the design, construction, and
                                                      operation of secondary sewage treatment plants to ac-
                                                      commodate the communities of  Dexter and  Corinna
 (b)  NES   (I974)

Lost
(0)
>
(0)
*
EAST BRANCH



1=11 800

EAST BRANCH
                                                     (a)  Anon  (1966)

( 54 00)
Lost
UI8900
EAST BRANCH



I = 125400

EAST BRANCH
                                                                                   97OO
                                                                                   precip
 (c)   Dennis  and  Corson  ( 1979)
                                                      (b)  NES  (1974)

Lost
0200)
(
(0)
»
EAST BRANCH
Nonpoint - 3190 	 — , 	 **-
1

Z = 9000

EAST BRANCH

                   t      I
                  (5001     ?      200
                  Lost   Septic tanks  precip

(0)
1
(0,
EAST BRANCH
No.poml ' 122000 	 , 	 —
!


1= 283200


210300
EAST BRANCH

Figure 2.—Phosphorus budget  estimates for Sebasticook    Figure 3.—Nitrogen budget estimates for Sebasticook Lake
Lake reported from three studies.                          reported from two studies.
                                                 505

-------
 LAKE AND RESERVOIR MANAGEMENT


   and their industries. To  demonstrate the effects  01
   reduced  fertilization, it  is proposed that  phosphate
   removal facilities, such as alum or  lime precipitation
   be added to the secondary sewage treatment plants
   Following the installation and functioning of nutrienl
   control procedures, the lake's  water level would be
   lowered during  the summer's maximal  algal  growth
   and the  lake subsequently filled with  nutrient  pooi
   water (Fed. Water Pollut. Control Admin. 1966).
 Obviously,  both the external and internal phosphorus
 loadings must be significantly reduced  if the objective
 of a 10 to 20 /^g/l total phosphorus concentration is to
 be met. Currently, the  springtime  level  of total
 phosphorus is between 30 and 40 ^g/l (Dennis and Cor-
 son, 1979).

 Reduction of Nutrient Input

 The three  major  sources of phosphorus have been
 identified as Dexter, Corinna, and nonpoint inputs. In-
 stallation of secondary treatment  at  Dexter would
 reduce the  present 2,300 kg P/yr by  about 25 percent,
 while  advanced  phosphorus removal  could  be  ex-
 pected to reduce this discharge to 300 kg P/yr. Secon-
 dary land disposal of Dexter's waste could eliminate
 its phosphorus load altogether  and is currently  the
 most cost-effective treatment  alternative. Although
 Corinna installed secondary treatment in 1970, it did
 not operate satisfactorily until about 1978. Since that
 time, the woolen mill, which contributes about 90 per-
 cent of the volume, has installed a number of water
 saving controls and the treatment plant has installed
 covers over the clarifiers to improve treatment during
 cold weather. The Maine DEP currently estimates the
 phosphorus discharge from Corinna at about 790 kg
 P/yr and agreement has been reached  to license  the
 facility at this rate. Dennis and Corson (1979) had esti-
 mated that with  advanced  wastewater treatment,
 phosphorus in the Corinna discharge could be reduc-
ed to 500 kg P/yr.
     The reduction of nonpoint input depends upon im-
   proving farming operations as 85 percent of the non-
   point  phosphorus comes  from  agriculture  (Penob.
   Valley Reg. Plan. Comm., 1980). Runoff from cropland
   and  livestock  operations,  particularly  the  winter
   storage  and spreading of animal manure,  has been
   identified as probable source of phosphorus. Control
   efforts directed at 25 of the 67 farms in the watershed
   are expected to reduce the overall nonpoint inputs by
   25 to 50 percent.
     A phosphorus model (Vollenweider, 1976) was used
   to assess the  change in  lake phosphorus concentra-
   tion from various abatement alternatives for the exter-
   nal phosphorus sources. The results (Table 3) sug-
   gested that phosphorus reduction was needed at both
   Dexter and Corinna, and reduction in nonpoint loading
   was also necessary to reach the desired objective of
   4,500  kg/yr.  Alternatives  B-F are the possible stra-
   tegies which could achieve the objective; alternative C
   was selected  as the most cost-effective technique
   (Table 3).

   Reduction of Phosphorus Recycling

   Since the recycling of phosphorus from the lake sedi-
   ments each summer brings 6,900 to 9,900 kg P into the
   water column, it is  imperative that the cycle be broken
   or at least minimized. While a variety of control tech-
   niques exist, most are economically  infeasible  be-
   cause the lake is so  large. The most promising tech-
   niques are seasonal  drawdown  or hypolimnetic dis-
   charge.
     To evaluate the two techniques,  Hannula (1978)
   developed a computer simulation model of the inter-
   nal cycling of  phosphorus in  Sebasticook Lake. The
   model used seasonally varying  rate  coefficients  for
   eddy  diffusion,  phosphorus  release  (aerobic  and
   anaerobic), and sedimentation to produce water col-
   umn phosphorus profiles  similar to those observed in
    Table 3.—Predicted total phosphorus concentrations in Sebasticook Lake for various external control alternatives.
Alternatives
A.




B.



C.



D.



E.



No change in loading
Dexter
Corinna
Nonpoint sources

Dexter— land treatment
Corinna — land treatment
NPS

Dexter— land treatment
Corinna— flow reduction'
NPS controls

Dexter— land treatment
Corinna— advanced treatment^
NPS control

Dexter— advanced treatment
Corinna— flow reduction
NPS control

Load (kg/P)

2200
2300
4500
9000
0
0
4500
4500
0
800
3500
4300
0
500
3500
4000
300
800
3500
4600
Predicted spring total
phosphorus cone, (^g/l)


30



15



14



13



15


 F. Dexter—advanced treatment
   Corinna—advanced treatment
   NPS controls
 300
 500
3500
4300
14
 1 Flow reduction assumes water savings technology and for reduced production capacity at the expense of the woolen mill
 2 Advanced treatment assumes tertiary treatment with effluent total phos.phorus< 0.5 mg/l
                                                 506

-------
                                                                                RESTORATION TECHNIQUES
 Sebasticook Lake. The  model supported  the hypo-
 thesis that there  is a significant internal  source of
 phosphorus during the summer. Simulation runs con-
 firmed  that  the  current  phosphorus  loading, even
 coupled with the available 1.5 m late  summer draw-
 down, resulted in a buildup of phosphorus in the sedi-
 ments. Increasing the drawdown  to 3.5 m would pro-
 duce a small net export of phosphorus from the lake.
   If drawdown is to have a major impact, the model
 demonstrated  that external  inputs must be  signifi-
 cantly reduced. When the external phosphorus input
 was reduced by 45 percent and combined with a 3.5 m
 drawdown in late summer with spring  refill, a  net ex-
 port of 3,000 to  4,000  kg P/yr was predicted. This
 resulted in a spring phosphorus concentration of 23
 ng/l  at  the end of  year  1. Although the model only
 simulated 1  year, additional reduction in  the phos-
 phorus concentration would be anticipated so that the
 10 to 20 ^g P/l goal could be  reached.
   In addition to the  drawdown alternative, Hannula
 (1978) also modeled the use of a constant  hypo-
 limnetic discharge. This technique was shown to be
 the most effective, resulting in a net export of 4,800 kg
 P. Even so, Hannula cited two reasons for not  recom-
 mending this strategy: (1) the cost of implementation
 would be high since the deepest section of the lake is
 several kilometers from the outlet; (2) the discharge of
 anaerobic water  would  require  treatment  before
 discharge to the outlet stream. In conclusion, Hannula
 decided that the  extended drawdown technology ef-
 fectively exported phosphorus released from the sedi-
 ment and would be easier to  implement.

 Restoration Plan

 The $1.2 million restoration plan adopted by the Maine
 DEP called for constructing a 4.0  m deep, 512 m long
 discharge channel at the lake outlet with a new con-
 trol dam. Coffer dams at the inlets to the lake were in-
 cluded  to protect upstream  wetlands  during  draw-
 down periods. Completed in 1982, the new  lake level
 control structure allows a 3.5 m drawdown,  reducing
 lake volume by 51 percent. In  the initial year of opera-
 tion, 3,800 kg P were  removed by  a partial drawdown
 of 2.7 m compared with 2,400 kg P if the usual 1.5 m
 drawdown had  been used.
   Plans have been  made to control point discharges
 from Corinna and Dexter. While improvements at Cor-
 inna have reduced its phosphorus input to an accep-
 table level, a wastewater treatment plant proposed for
 Dexter remains at the planning stage and actual con-
 struction  is  not  scheduled  until spring 1984. An-
 ticipated design calls  for spray irrigation of the waste-
 water. Engineers estimate the land wastewater treat-
 ment system at Dexter will cost $3.0 million, excluding
 sewers.
   The U.S. Soil Conservation Service and Agricultural
 Stabilization and Conservation Service have instituted
 a nonpoint control program at a cost of another $1.3
 million. The main focus of the  program is construction
of manure storage facilities for area farms along with
contour farming,  buffer strip  projects,  winter cover
crops, and other conservation practices to reduce ero-
sion and surface runoff. By 1982, 20 percent of the
farms (the  largest ones were treated first) had com-
pleted their projects;  it is expected this will treat 30
percent  of controllable  phosphorus  from  winter
spreading of manure.
SUMMARY

It is apparent  that the lake has been receiving ex-
cessive nutrients for at least 30 years, but restoration
is sought more quickly. The unanswered question is:
How much time will the reversal take?
   The restoration project is based on two concurrent
strategies: reduction of external  loading, and deple-
tion of the available phosphorus from the sediment.
Control of all external phosphorus loads is expected
within a few years. Lake drawdown has been chosen
as the most promising technique for flushing sedi-
ment phosphorus from the lake. To a large extent, the
effectiveness of drawdown depends upon the amount
of  phosphorus available from the sediments. Lake
recovery will be relatively quick only if a large fraction
of  the sediment-bound phosphorus remains in  the
sediment. Computer simulation of the internal phos-
phorus cycle suggests that drawdown will be effective
and an initial trial has shown that drawdown can be ef-
fective in remvoing substantially more  phosphorus
than would be expected by natural conditions.
REFERENCES

Dennis, J., and A. Corson. 1979. External loading and internal
  recycling of phosphorus in Sebasticook Lake. Maine Dep.
  Environ. Prot. Augusta.
Federal Water Pollution Control Administration. 1966. Ferti-
  lization and algae in Lake Sebasticook, Maine. Cincinnati,
  Ohio.

Hannula, T.A. 1978. Modeling phosphorus cycling in Sebasti-
  cook Lake  (Newport, Maine). OWRT Proj. A-039-ME, Land
  Water Resour. Center, Univ. Maine at Orono.
Keup, LE. 1968.  Phosphorus in flowing  waters. Water Res.
  2: 373-86.
Mackenthun,  K.M.,  LE. Keup,  and R.K.  Stewart.  1968.
  Nutrients and algae in Sebasticook Lake, Maine. J. Water
  Pollut. Control Fed. R72-R81.
Morrill, R.A. 1975. A technique for estimating the magni-
  tude and frequency of floods. U.S. Geol. Survey Open File
  Rep. 75-292.
Nolan, P.M., and A.F. Johnson. 1975. Comparative Study of
  the Eutrophication of Sebasticook Lake,  Maine.  1965,
  1971-1973. U.S. Environ. Prot. Agency,  Boston, Mass.
National Eutrophication Survey. 1974. Report on Sebasti-
  cook Lake,  Penobscot County Maine, Working Pap. 9. U.S.
  Environ. Prot.  Agency, Corvallis, Ore.
Valley Regional  Planning Commission.  1980. Sebasticook
  Lake watershed preapplication report.  Penobscot Bangor,
  Maine.
Vollenweider,  R.A. 1976. Advances in defining critical loading
  levels for phosphorus in  lake eutrophication. Mem. 1st.
  Ital. Idrobiol. 33:53-88.
                                                 507

-------
 MINNEAPOLIS CHAIN  OF LAKES VACUUM  SWEEPING  AND

 RUNOFF DIVERSION



 JOHN B. ERDMANN
 NORMAN C. WENCK
 Eugene A.  Hickok and Associates
 Wayzata, Minnesota


 PERRY DAMON

 Minneapolis Department of Public Works
 Minneapolis,  Minnesota



             ABSTRACT

             The Minneapolis Chain of Lakes, a series of five lakes covering some 445 ha (1,100 acres) within
             an  urban residential setting, is an important recreational and aesthetic resource The lakes support
             swimming, fishing, sailing, and canoeing and are almost entirely surrounded by public parkland
             However, the lakes exhibit increasing cultural eutrophication. The city of Minneapolis obtained an
             EPA Clean Lakes grant for demonstrating vsicuum street sweeping and first-flush runoff diversion as
             means for improving the lakes. The project was completed in 1981 following 2 years of monitoring
             and pilot implementation.  In 1979 and 1980, the city regularly vacuum swept streets throughout the
             Lake Harriet watershed (46 curb-km in 340 ha). During these years, consultants monitored lake quali-
             ty, weather, runoff flow, and quality in majo- storm drains to Lake Harriet, and quantity and quality
             of vacuum swept materials. Runoff data were used to analyze first-flush diversion. Data analysis resulted
             in estimated runoff coefficients ranging from 0.05 to 0.5 for individual drainage areas throughout the
             chain Seasonal patterns of runoff quality were found. Total phosphorus export from the entire water-
             shed of the chain was estimated to average 5.2 kg per hectare (1 pound per acre) in a climatologically
             normal year (with no allowance for sweeping or diversion). Detailed water and nutrient budgets, in-
             cluding ground water and inter-lake flows, were developed. Runoff and direct precipitation contributed
             nearly equally to the water budget, but runcff accounted for over 95 percent of phosphorus inputs
             Water outflow was predominantly via seepage. Phosphorus retention in the lakes was high, totaling
             over 90 percent for the whole chain. Vacuum sweeping was found to remove 3 kg phosphorus/curb-
             km and 69.5 kg organic matter/curb-km (0.2€ Ib phosphorus/curb-mile and 74 Ib organic matter/curb-
             mile) per sweeping. These are average values: seasonal variations were significant. Weekly sweep-
             ing was projected to remove 38 percent of Luke Harriet's phosphorus load. First-flush diversion  was
             analyzed by taking into account the frequency distribution of storms with respect to total precipita-
             tion. Diversion was found to be cost effective in some areas and capable of reducing the phosphorus
             load to Lake of the Isles by 42 percent. A combination of sweeping and diversion throughout the Cham
             of Lakes watershed was found most cost-effective. This scheme could reduce the whole chain's
             phosphorus load by 27 percent, at an estimated 10-year cost of $4 million. Predicted transparency
             increases averaged .77 m (2 ft.) and ranged from .2 m to 1.9 m (one-half to five feet) for individual lakes.
INTRODUCTION

The  Minneapolis Chain of Lakes is a series of  five
lakes covering some 430 hectares within an  urban
residential setting (see Fig. 1). Maximum depths range
from 9.4 to 27.7 meters, as Table 1 shows. The lakes
are almost entirely surrounded by parkland and are an
important recreational and  aesthetic resource, sup-
porting  swimming,  fishing,  sailing,  canoeing,  and
other activities.
  The lakes exhibit cultural eutrophication. An earlier
study (Shapiro and Pfannkuch, 1973) discussed runoff
diversion  and vacuum street  sweeping  as possible
restoration measures for the Chain of  Lakes. Sub-
sequently, the city  of Minneapolis obtained  an EPA
Clean Lakes research and demonstration grant to in-
vestigate this  further. In 1978, the firm of Eugene A.
Hickok  and Associates  was  contracted  as  con-
sultants. The project was completed in 1981, following
2 full years of pilot implementation and  monitoring,
and  is fully documented elsewhere (Erdmann et al.
1981). This  paper briefly summarizes the research and
demonstration project and its findings.

NATURE  OF PROJECT

The  city implemented vacuum  street sweeping  on a
pilot  basis and  installed  two "first  flush" runoff
                             Table 1.—Morphometiy of Minneapolis Chain of Lakes.
Lake
Harriet
Calhoun
Isles
Cedar
Brownie
Area
(ha)
142
167
46
70
3.3
Volume
(million cum)
12.6
16.3
1.1
4.3
0.1 Ei
Mean Depth
(m)
8.84
9.75
2.41
6.04
4.85
Maximum Depth
(m)
25.0
27.7
9.4
15.5
15.8
                                                  508

-------
                                                                               RESTORATION TECHNIQUES
diverters as part of  the project  (Fig. 2). In 1979 and
1980, the city regularly vacuum swept streets through-
out the Lake  Harriet watershed (114 curb-km  in 340
ha), using a Ford Vac All sweeper. Typically, each area
was swept 13  times from late May to early November,
with sweeping less frequent in summer than  in  spring
and fall.
   The idea of  first flush  diversion  is to prevent
especially  nutrient-rich  storm runoff (thought to be
associated with a first flush effect) from entering the
lakes by diverting it into  sanitary sewers. However,
regional and State authorities would not allow runoff
diversion into  sanitary sewers for this study. In an at-
tempt still to study diversion, the two diverters were in-
stalled in very steeply sloped storm drains,  with the
diverted flow returning to the original drain at a down-
gradient point. Monitoring equipment functioned very
poorly under  these  conditions.  Therefore, diversion
ultimately was studied through analyzing runoff data
obtained at Lake Harriet.
   During 1979 and 1980, the consultant conducted an
extensive monitoring program (Fig. 3). The monitoring
included lake quality, weather, runoff flow and quality
in major storm drains to  Lake  Harriet, and  quantity
and quality of vacuum swept materials. All chemical
analyses were performed  at the consultant's labora-
tory.

LAKE WATER QUALITY
Table 2 presents average water quality data for the
five lakes, based on this and other recent studies. AC-
Figure 1.—Minneapolis Chain of Lakes location map.
  DRAINAGE
  DELINEATION
            FIRST  FLUSH  DIVERTERS
Figure 2.—Pilot implementation of sweeping and diversion.
                                                 509

-------
 LAKE AND RESERVOIR MANAGEMENT
 cording to  lake  classification  data (for  exampte,
 Vollenweider and Kerekes, 1980),  all five lakes are
 eutrophic.  The lake-to-lake variation  in  quality is
 related primarily to depth and watershed size.
   Lake Harriet exhibited some improvement in quality
 during the course of the project. In particular, summer
 total phosphorus decreased from  typical  ranges of
 0.03-0.10 mg/l in 1979 to less than 0.03 mg/l in 1980
 (Fig. 4). The other lakes did not exhibit water quality
 improvement. Thus, the improvement in Lake Harriet
 appeared  to be a beneficial effect  of the vacuum
 sweeping there.

 RUNOFF CHARACTERIZATION

 Runoff to  Lake Harriet  was  monitored with  flow-
 activated Stevens Type F water level recorders and
 Isco model 1680 superspeed  automatic  samplers.
 Rectangular weirs and PVC stilling  wells were install-
 ed for flow measurement, except at two stations hav-
 ing conditions adverse  for weirs but favorable for us-
 ing open channel (i.e.,  partially full pipe) flow equa-
 A LAKE STATION

 • WEATHER STATION
 • RUNOFF STATION
MORGAN AVE
 Figure 3.—Chain of Lakes monitoring stations.
              tions. Weir flow equations were developed to take into
              account  the approaching  velocity and nonvertical
              manhole walls.
                The monitoring yielded 150 interpretable runoff flow
              records. Runoff samples were obtained and analyzed
              in about half these, as well as in additional cases lack-
              ing flow records.
                Flow data were sufficient to determine runoff coeffi-
              cients for seven of the  Lake Harriet subwatersheds.
              The coefficients  were related linearly to the water-
              sheds' percentage  of impervious  area (correlation
              coefficient +.96). Coefficients were then estimated
              for  subwatersheds  throughout the chain based  on
              percentage impervious area. The individual lake water-
              sheds range from 145 to 1,259 ha in area, and runoff
              coefficients are generally low, averaging .138 (Table 3).
                Phosphorus and  other runoff  constituents were
              found to vary seasonally in concentration. Variability
              among subwatersheds could not be related success-
              fully to land use because of the homogeneity of land
              use  throughout  the  whole  watershed.   Thus,
              phosphorus  export was determined from  average
              seasonal concentrations and runoff volumes. The run-
              off  data obtained at Lake Harriet were corrected  for
              the  effects of vacuum  sweeping for this purpose. The
              2 study years closely bracketed a climatologically nor-
              mal  year. Overall for the whole  chain, the  normal
              runoff Total P concentration was estimated  as 1.17
              mg/l, and normal Total P export as 1.2 kg/ha/yr.
                       Table 2.—Average lake water quality.
Lake
Harriet
Calhoun
Isles
Cedar
Brownie
Total P1
(mg/l)
0.054
0.076
0.090
0.054
0.062
Chlorophyll-a*
(ug/l)
9
13
43
10
45
Secchi Depth2
(cm)
226
201
98
178
133
                                                      Shapiro and Pfannkuch (1973).
                                                      "Summer averages; data as above, plus 1973-1980 unpublished data by Shaprio

                                                      Table 3.—Watershed areas and average runoff coefficients.
Watershed
Harriet
Calhoun
Isles
Cedar
Brownie
Total/Area-Weighted Avg.
Area
(ha)
461
1,259
296
659
145
2,820
Avg. Runoff
Coefficient
.159
.138
.225
.098
.081
.138
                                                      Table 4.—Normal water and phosphorus budgets for Lake
                                                                            Harriet.
Figure 4.—Lake Harriet total phosphorus (mg/l).
Component
Inflows:
Precipitation
Runoff
In-Seepage
Total In
Outflows:
Evaporation
Out-Seepage
Sediment Retention
Total Out
Water
(million cu m)

1.03
0.56
0.04
1.63

1.10
0.53
_0 	
1.63
Total P
(kg)

10
649
1
660

0
29
631
~660~
                                                 510

-------
                                                                             RESTORATION TECHNIQUES
WATER AND PHOSPHORUS BUDGETS

Detailed  water and   phosphorus  budgets  were
developed on a seasonal basis for all the lakes of the
chain.  In addition to runoff,  these included ground-
water interactions (in-seepage and out-seepage), bas-
ed on information from a related study (Hickok, 1980),
and inter-lake flows and artificial pumpage for the four
upper lakes.
  Runoff and direct precipitation contribute com-
parable volumes to the water budget, but runoff yields
over 95 percent of phosphorus inputs. (See summary
budgets for a normal year in Tables 4 and 5). Water
outflow is predominantly via seepage. Over 90 percent
of the  annual phosphorus loading is retained in the
lake sediments.
LAKE PHOSPHORUS MODEL

A commonly used mass-balance formulation for lake
phosphorus is Dillon, 1975:
TP =
              - R)/zc
 Table 5.—Normal water and phosphorus budgets for upper
                      lakes.
Component
Inflows:
Precipitation
Runoff
In-Seepage
Artif. Pumpage
Total In
Outflows:
Evaporation
Out-Seepage
Sediment Retention
Total Out
Water
(million cu m)

2.08
2.29
0.47
1.49
6.33

2.20
4.13
0
6.33
Total P
(kg)

21
2,665
9
104
2,799

0
296
2,503
2,799
                                               in which
                                                   TP = average Total P concentration in lake(mg/l),
                                                    L =  areal P loading (g/sq m/yr),
                                                    R  = retention coefficient (fraction of load re-
                                                         tained in lake sediments),
                                                    z =  mean depth (m), and
                                                    Q =  hydraulic flushing rate (per year).

                                               Lake morphometry, together with the water and phos-
                                               phorus budgets, determine the terms in  the right-hand
                                               side of  this equation.  For  example, Lake Harriet's
                                               budget shows normal retention of  631  out of 660 kg,
                                               implying a retention coefficient of .956, as shown in
                                               Table 6. Table 6 shows the terms for each lake, based
                                               on normal budgets, along with the predicted Total P.
                                               The predicted values are in all cases within 0.001 mg/l
                                               of the average data given in Table  2.
                                                 In addition to the phosphorus model,  empirical rela-
                                               tionships among phosphorus, chlorophyll a and Sec-
                                               chi  depth were also determined for each lake.
EVALUATION OF VACUUM SWEEPING
AND RUNOFF DIVERSION

Vacuum sweeping of Lake Harriet watershed streets
removed approximately 190,000 kg of solids, including
110 kg P and 32,000 kg of organic matter in 1980 (all
values on dry weight basis). Overall average removal
rates per sweeping were 0.0733 kg P/curb-km and 20.8
kg organic matter/curb-km. Seasonal removal rates in-
creased from spring through fall.
  It was  estimated that  weekly sweeping (more  fre-
quent than in the pilot implementation) could remove
249 kg P annually from the Lake Harriet watershed, or
777 kg P from the whole Chain of Lakes. The latter
figure is affected by the  assumption that major por-
tions of the upper lakes' watershed outside the city of
Minneapolis  would not be swept.
  To  assess the  potential  effectiveness of runoff
diversion,  73 runoff hydrographs with corresponding
chemical  data were analyzed in detail.  The data
represented seven drainage areas, ranging from 5.6 to
                              Table 6.—Lake model terms and predicted total P.
Lake
Harriet
Calhoun
Isles
Cedar
Brownie
Retention
Coefficient
.956
.892
.870
.877
.662
Mean Depth
(m)
8.84
9.75
2.41
6.04
4.85
Normal
Flushing Rate
(per yr)
0.0425
0.135
0.757
0.323
5.19
Normal Areal
Load
(g P/sq m/yr)
0.463
0.921
1.25
0.856
4.55
Predicted
Total P
(mg/l)
0.054
0.076
0.089
0.054
0.061
 Table 7.—Predicted water quality improvements and estimated expenses for most effective combination of sweeping and
                                              diversion.
Lake
Harriet
Calhoun
Isles
Cedar
Brownie
Total or Avg.
PLoad
Removed
(kg/yr)
249
308
312
51
30
950
Lake
Total P
Decrease
(Percent)
38
20
54
8
20
27
Secchi Depth
Increase
(cm)
145
23
106
16
31
64
Total
10-Year
Expense
$1,212,000
1,207,000
1,164,000
244,000
149,000
$3,976,000
Avg.
Expense
per kg P
$485
392
373
485
485
$419
 Note Expense estimates in 1981 dollars, assuming 10 percent annual interest on capital and 10 percent annual inflation on operation/maintenance expenses
                                                511

-------
LAKE AND RESERVOIR MANAGEMENT
70 ha, and 31  storms, ranging in total  precipitation
from 2 to 47 mm. Diversion effectiveness also largely
depends on the proximity of storm drains and sanitary
sewers, and the sanitary sewer capacity to accept ad-
ditional flow. Diversion is potentially most effective
for Lake of the  Isles and Lake Calhoun, where annual
phosphorus removals  were estimated as 240 kg arid
221 kg, respectively.
  The most effective combination  of measures was
found to be runoff diversion  in all feasible subwater-
sheds  of Lakes Calhoun and Isles, with  vacuum
sweeping throughout the remainder of the watershed
(Minneapolis portion only). For this combination, the
overall reduction in phosphorus would be 27 percent
(Table 7). Predicted Secchi depth increases range from
16 to 145 cm and average 64 cm for the five lakes. The
total 10-year expense  estimate is  approximately $4
million, implying an average expense of $419 per kg P
removed.
 REFERENCES

 Dillon, P.J. 1975. The phosphorus budget of Cameron Lake,
  Ontario: the importance of flushing rate to the degree of
  eutrophy in lakes. Limnol. Oceanogr. 20: 28-39.

 Erdmann, J.B., E.J. Johnson, N.C. Wenck and E.A. Hickok.
  1981. Final report—Minneapolis Chain of Lakes research
  and demonstration project. For  city of Minneapolis by
  Eugene A.  Hickok and Associates, Wayzata, Minn.

 Hickok, Eugene A., and Associates. 1980. Lake level manage-
  ment  program—phase 1  For  Minneapolis  Park and
  Recreation Board.

Shapiro, J, and  H.O.  Pfannkuch.  1973. The Minneapolis
  Chain of Lakes - a study of urban drainage and its effects,
  1971-1973.  Interim Rep. No. 9, Limnol.  Res. Center, Univ.
  Minnesota, Minneapolis.

Vollenweider, R.A., and  J. Kerekes. 1980. The loading con-
  cept as basis for controlling eutrophication—philosophy
  and preliminary results of the OECD programme on eutro-
  phication. Progr. Water Technol. 12: 5-38.
                                                  512

-------
LONG-TERM EVALUATION OF THREE ALUM TREATED LAKES
PAUL J.  GARRISON
DOUGLAS R. KNAUER
Bureau of Research
Wisconsin  Department of Natural Resources
Madison, Wisconsin


            ABSTRACT

            Alum treatment of lakes has been found to be a generally effective method for reducing P con-
            centrations in lakes over a period of 2 to 3 years after treatment. However, the long-term benefits
            of lake alum treatment in most cases have not been determined. In this study, three Wisconsin
            lakes treated 9 to 12 years previously were re-examined in 1982 to evaluate the long-term effects
            of the treatment projects. Results indicated that many years after treatment P levels in all three
            lakes remained lower than before the lakes were treated. Horseshoe Lake, the first  U.S. lake
            treated with alum in 1970, was more eutrophic in 1982 than immediately after treatment but is
            still much improved over pretreatment conditions. The  reduced levels of phosphorus in 1982
            compared with before the alum treatment, indicate the alum layer is preventing the migration of
            phosphorus from the deep sediments. The 1982 phosphorus concentrations in the hypolimnion
            of Snake Lake are similar to levels experienced following the 1972 treatment despite continued
            input of stormwater runoff to the lake. While the previous two lakes are dimictic, Pickerel Lake is
            a polymictic lake. The increased mixing action has redistributed much of the alum toward the
            center of the lake. The alum treatment appeared to have  little effect on the internal phosphorus
            dynamics of the lake.
INTRODUCTION

Nutrient diversion alone does not always bring about
rapid  reduction in lake water concentration of phos-
phorus; nutrient-rich  sediments or  long phosphorus
residence times may delay this. (Larsen et al. 1981;
Cooke et al. 1977; Garrison and Knauer, 1983). To ac-
celerate the reduction of in-lake phosphorus concen-
trations, techniques were developed for phosphorus
precipitation/inactivation. The most widely used has
been aluminum sulfate (alum).
  The first lake treated with alum was Lake Langsjon
in Sweden in 1968 (Jernelov, 1970). While many lakes
have been treated with alum since then (at least 27 by
1980; Cooke and Kennedy, 1981), little is known of the
useful longevity of a treatment. Most reported studies
have been short term, only 2 to 3 years (Peterson et al.
1973;  Dominie, 1980; Gasperino et al.  1980). While
these studies  have indicated that  their treatments
succeeded initially, some studies have indicated the
benefits were only short lived (Born, 1979; Funk et al.
1980). It is unclear from these studies  whether the
treatment  failed because of the alum or because the
alum layer was overwhelmed by high external loading
rates.
  This study was initiated to answer the question of
the longevity of an alum  treatment, and if the treat-
ment  was ineffective, to  explain why. Three of the
earliest treated lakes were chosen as sites. Each lake
was treated  once between 1970 and 1973. The lakes
were  reexamined in  1982  during  the open  water
season.
  All  three  lakes are  located within  the State of
Wisconsin (Fig.  1).  Horseshoe  Lake in Manitowoc
County was the first lake treated in the United States.
It was treated in May 1970, with about 200 mg/l of slur-
ried aluminum sulfate (18 mg Al/l) in the upper 0.7
meters. The lake is a dimictic 8.9 ha lake with a max-
imum depth of 16.7 m and  a mean depth of 4 m (Table
1).  Horseshoe  Lake  is a hardwater lake with an
alkalinity of 230 mg/l.
  Snake Lake is a softwater (alkalinity 13 mg/l), dimic-
tic lake located in Vilas and Oneida counties. It has a
surface area of 5 ha,  maximum depth of 5.5 m,  and a
mean depth of 2 m (Table 1). In May 1972, the lake was
treated with liquid  aluminum sulfate  and  sodium
aluminate in  the upper 0.7  m. An aluminum concentra-
              Table 1.—Data concerning the morphometry and alum treatment of the three study lakes
                                       Horseshoe
               Snake
                                                                                           Pickerel
Maximum depth
Mean depth
Surface area
Volume
Shoreline length
Watershed
Hydraulic residence time
Year treated
Amount Al applied
16.7m
4.0 m
8.9 ha
3.6Xl05fT)3
1.8km
700 ha
0.7 yr
1970
18 mg/l
5.5 m
2.0m
5.0 ha
1.0x1 05m3
1.2km
25.9 ha
1.5yr
1972
12 mg/l
5.0 m
2.6m
20.0 ha
5.2x105m3
1.9 km
50.5 ha
1.6yr
1973
7.3 mg/l
                                                513

-------
 LAKE AND RESERVOIR MANAGEMENT
 tion of 12 mg/l was achieved in 80 percent of the lake
 volume.
   Unlike the previous two lakes, Pickerel Lake in Por-
 tage County is polymictic. It has a surface area of 20
 ha, maximum depth of 5 m, and a mean depth of 2.6 m.
 It is a hardwater lake with an alkalinity of 110 mg/l. It
 was treated with liquid alum on the surface  in April
 1973,  achieving  an  aluminum concentration of 7.3
 mg/l.
 RESULTS

 Horseshoe Lake

 Prior  to  the  alum treatment, Horseshoe Lake was
 eutrophic.  The   major  source  of  nutrients  was
 suspected to be effluent from a cheese and butter fac-
 tory. In addition to agricultural drainage to the lake, a
 tile system from the factory allowed direct drainage
 from the waste lagoon into the lake from 1963 to 196!5.
 Soon  after,  winter fishkills  became  common and
 spring and summer blooms of  the algae Anabaena,
 Microcystls, and Oscillatoria were prevalent (Peterson
 et al.  1973).
    Although effluent from the cheese factory did not
 enter the lake after  1965,  the lake remained very
 eutrophic partially because of  internal loading from
 sediments. In 1970 the lake  was treated with alum.
 Phosphorus  concentrations decreased dramaticaly
 with  average P  concentrations declining from 0.14
 mg/l in 1966 to 0.04 mg/l in 1971 (Peterson et al. 1973;
 Fig. 2). Even  more dramatic was the decrease of P in
 the hypolimnion from 1.5 to 0.1 mg/l (Fig. 2). However,
 by 1972, 2 years following the treatment, P in the bot-
                      HORSESHOE LAKE

                     TOTAL PHOSPHORUS
      16-

      I.4-

      I.2-

      I.O-

  01  0.8-

     0.6-

     0.4-

     0.2-

      0
          Hypolimnion (12m)
1966
                  1971
               M    J    J  '  A  '  S  '  0  '  N  '

Figure 2.—Horseshoe Lake total phosphorus trends. The top
figure is weighted  mean concentration before (1966) and
following the 1970 alum treatment. The bottom figure is P
concentrations in the mid-hypolimnion.
                                                                                   HORSESHOE LAKE
                                                        PICKEREL LAKE

Figure 1.—Location and bathymetric maps of the study lakes. Contours are in meters. The solid circle is the limnolooical
sampling location in each lake.
                                                 514

-------
                                                                              RESTORATION TECHNIQUES
torn waters was higher than the previous year and in
1973 concentrations wsre progressively higher yet.
Nutrients from agricultural runoff were still entering
the lake. We believe the alum layer was  buried by
allochthonous and  autochthonous material  and re-
mineralization of recently sedimented P was returning
to the overlying water  unaffected by the alum layer.
Migration of P from the deeper sediments appears to
have declined since P concentrations in the hypolim-
nion are lower in 1982 than they were in 1966.
  After 1973 phosphorus concentrations seemed to
have reached an equilibrium level. In 1974 and 1975, P
levels both in the bottom waters and the entire water
column were similar to those experienced  in 1973.
Although  P  concentrations  were higher than  im-
mediately following the alum treatment they were
much lower than in 1966.
  Water transparency  was difficult to interpret. The
mean value in 1982 was 2.8 m (Fig. 3). This is similar to
values reported in 1966 when copper sulfate was used
periodically, but much better than the values reported
in 1971 and  1973. Following the spring Oscillatoria
bloom  in 1982, chlorophyll values  declined, averaging
7 mg/m3 for the summer months.  The summer phyto-
plankton community indicated a meso-eutrophic lake
being  rather diverse  with  cryptomonads  such  as
Chroomonas acuta being important.

Snake Lake

Prior to 1942, Snake Lake was of good water quality. In
1942 the water quality rapidly deteriorated as effluent
from a sewage treatment  plant was discharged into
the lake. A large fish  kill occurred in the winter of
1942-43 and large algal  blooms and disagreeable
                   SECCHI DISC DEPTH
o-
" I -
I
0. ?-
LJ '-
O
1
Snake Lake
^r-"^"^7^""
	 » — n —

A ' M ' J ' J '
<^-— _^_
-.^ ^-1982
""•-- .--"" 	 -.
1973-^
A ' S ' 0 ' N '
   2-



   4-

   5
       Horseshoe Lake
                                       1982
   0-

    I-
•J-  O -


£  3H
UJ
Q
   4H
             M
       Pickerel Lake
                   1982-
                      -^.
       A'M'J'J'A'S'O'N

Figure 3.—Water transparency observed in the study lakes.
                                                   odors were common in subsequent years (Born et al.
                                                   1973). Following diversion of the sewage effluent in
                                                   1964, algal blooms continued and low oxygen condi-
                                                   tions persisted. To restore the lake, a dilutional pump-
                                                   ing project was conducted in 1970. Although nutrient
                                                   levels were reduced, they still remained high (Born et
                                                   al. 1973).
                                                     The  year prior to the alum treatment total  phos-
                                                   phorus levels averaged 0.24 mg/l for the lake while at
                                                   times concentrations in the  deeper waters indicated
                                                   that considerable phosphorus entered the water col-
                                                   umn from the sediments. Following  the alum treat-
                                                   ment in early 1972,  P concentrations  decreased from
                                                   0.5 mg/l to 0.1 mg/l  by 1973.  The P concentrations in
                                                   the bottom waters declined from greater than 2.5 mg/l
                                                   to  0.13  mg/l,  even though the hypolimnion was
                                                   anaerobic. Water transparency improved to levels ex-
                                                   perienced prior to the discharge of sewage effluent in
                                                   the early 1940's.
                                                      In 1982 lake phosphorus concentrations resembled
                                                   those experienced  in 1973 after the  alum treatment
                                                   (Fig. 5). Further evidence of the continued success of
                                                   the alum treatment through  1982 was shown by the
                                                   low P concentrations (0.06  to 0.25  mg/l) in bottom
                                                   waters in the presence of anaerobic  conditions from
                                                   May through October.
                                                      Although the major external nutrient source to the
                                                   lake was diverted in 1964, storm sewer  runoff con-
                                                   tinues until the  present time. Consequently, algal
                                                   blooms occur following a large runoff  event. The water
                                                   transparency in 1982 was not as good as the first year
                                                   following the alum  treatment but it  was better than
                                                   1971 (Fig. 3). The adverse effect of the storm  sewers
                                                   upon the limnology of the lake is further exemplified
                                                   by the chlorophyll a levels in 1982 (Fig. 4). An algal
                                                   bloom occurred in July and August dominated by the
                                                   blue-green alga Anabaena wisconsinense.
                                                      Snake Lake is still a eutrophic lake exhibiting algal
                                                   blooms, although winter fishkills occur much less fre-
                                                   quently. The continued eutrophic nature of the lake
                                                   cannot be attributed to the alum's failure to reduce in-
                                                   ternal  P loading, but rather to the elevated nutrient
                                                                          SNAKE LAKE
                                                       60-

                                                       50-



                                                     130-


                                                     o
                                                        10-
                                                            AMJJ'ASON
                                                                             1982


                                                                       HORSESHOE LAKE
                                                        25-
                                                        15-

                                                        10-
                                                         5-
                                                             AMJJASON
                                                                             1982

                                                     Figure 4.—1982  chlorophyll a values  for Snake  and
                                                     Horseshoe Lakes.
                                                 515

-------
 LAKE AND RESERVOIR MANAGEMENT
  loading from storm sewer runoff. Internal loading from
  sediments is low, indicating the alum layer is still ad-
  sorbing  P and  preventing  its migration from  the
  nutrient-rich deeper sediments.
    Although seston  has undoubtedly sedimented on
  top of the alum layer, the alum seems to be reducing
  the amount of the sedimented P returning to the over-
  lying waters during anaerobic conditions.

 Pickerel Lake

 Pickerel Lake is a naturally eutrophic lake, with no sur-
 face inlets or outlets. Hennings (1978) reported that
 the major hydrologic input is ground water. Pickerel
 Lake  had a history of excessive  algal blooms and
 winter fishkills. In 1973 alum was  applied in April. A
 more  detailed  response to  the alum treatment was
 described in Knauer and Garrison (1980). Immediate!/
 following the alum treatment the algal community and
 phosphorus levels were low. These levels were main-
 tained throughout the summer until the lake mixed in
 late July. A large algal bloom resulted, dominated by
 the blue-green alga  Microcystis aeruginosa. A sub-
 sequent  investigation showed that the alum was re-
 distributed to the  center of the lake, thus exposing
 most of the lake's surface sediment to the overlyini]
 water.
   In 1982 phosphorus levels were similar to those
 following  the alum treatment, averaging 0.023  mg/l
 (Fig. 6). Algal blooms continue to  be a  problem. Al-
 though water transparency was good after ice out in
 April (Fig. 3), by June chlorophyll values had increased
 along  with  phosphorus  concentrations (Fig. 6).  In
     05-

     04-

  v  03-
  CT
  E  02-

     0 I-
Whole Lake
                       SNAKE LAKE

                     TOTAL PHOSPHORUS
                  I97I
         I973

Figure 5.—Snake Lake total phosphorus trends. The top
figure is weighted mean  concentration before (1971) and
after (1973, 1982) the alum treatment. The lower figure is F
concentrations just above the sediments in the deepest area.
 September chlorophyll  values  exceeded 20  mg/m3.
 The  dominant phytoplankton  were  blue-green,
 especially Aphanocapsa elachista.


 DISCUSSION

 Snake and Horseshoe Lakes continue to receive ex-
 cessive nutrient  loads even though large  nutrient
 sources have been diverted away from the lakes. Both
 lakes  have similar  phosphorus  concentrations  al-
 though summer chlorophyll values are much higher in
 Snake Lake and the algal community is dominated by
 blue-greens.  Horseshoe  Lake receives  most  of  its
 nutrient load from agricultural runoff, most of which
 probably occurs  during  spring  runoff.  Snake  Lake
 receives  most  of  its nutrients from  storm  sewers.
 While  a substantial  P loading occurs during spring
 runoff, large amounts of phosphorus enter the lake via
 urban drainage following summer rainstorms. This P
 is  readily   available  from   phytoplankton   growth
 (Knauer, 1975).
  Although  Snake Lake  continued to receive ex-
 cessive nutrient loads from the watershed, it did not
 exhibit increased phosphorus concentrations in the
 hypolimnion in  succeeding years following the treat-
 ment. Garrision and Knauer (1983) have shown that
the alum  layer will  either  remain  on top  of the
 sediments  or settle below  the  surface  sediments,
depending  upon sediment density.  The Snake Lake
sediments would be expected to be denser than those
 in Horseshoe Lake. Born et al. (1973) reported that the
sediments in Snake Lake were consolidated following
the dewatering  of the lake in 1970.
  In contrast, Horseshoe Lake sediments have not
been compacted. Petersen et al. (1973) reported that in
the upper 5 cm of Horseshoe  Lake sediments, the
water content was 91 percent. Garrison and Knauer
(1983) found in  two lakes similar to Horseshoe Lake
that if the surface sediments contained more than 90
percent water, the  alum layer settled into the denser
deeper sediments.
                                                                           PICKEREL LAKE
                                                          25-
                                                          15-

                                                          10-
                                                              A    ivf""  J  '  ~J'  A^1  s"1cT"1N~^
                                                                               1982
                                                       0030-

                                                     = 0025-

                                                     -§ 0020-
                                                     Q.
                                                     ^ 0.015-

                                                     g 0010-

                                                       0005-

                                                          0
                                                   AMJ    JASON
                                                                   1982


                                         Figure 6.—Pickerel Lake 1982 trends for chlorophyll a (upper)
                                         and weighted mean phosphorus concentrations (lower).
                                                 516

-------
  After 10 years, the alum treatment in Snake Lake is
still inhibiting internal P  loading, but  because of ex-
cessive loading  from storm sewers algal blooms con-
tinue. This points to the need to reduce the P load to
an acceptable level prior  to an alum treatment. Funk
et al. (1980) also noted that the alum treatment was
only effective 2  to 3 years in Liberty Lake because of
high external loading rates.  In contrast, two lakes in
Wisconsin, Mirror and Shadow  lakes,  first had storm
sewer runoff diverted away from them prior to an alum
treatment. Five years after  the treatment  the  phos-
phorus concentrations are  similar to levels experi-
enced following the alum application (Garrison and
Knauer, 1983). Cooke et  al. (1982)  report that  phos-
phorus concentrations also  remain  low in West Twin
Lake 5 years following an alum treatment. As with Mir-
ror and Shadow lakes, the major external phosphorus
sources were diverted away from the lake.
   Of the three treatments discussed in this paper, on-
 ly the Pickerel  Lake application could be considered
 unsuccessful. This is primarily a result of the lake's
 morphometry. During the summer of 1982 the lake fre-
 quently mixed. However, in 1973 the lake remained
 stratified from  May until late July. The larger P con-
 centrations  and algal  populations in  1973 following
 mixing probably resulted from  P being released from
 the sediments during the longer stagnation period. In
 1982  the lake  mixed  more frequently,  preventing
 anaerobic conditions from  persisting  with the result
 that  less P  entered the water from the sediments.  It
 appears that the mixing regime in Pickerel Lake is
 largely responsible for the phosphorus dynamics and
 the  extent of the algal  bloom. The  alum treatment
 seems to have  had little  effect  upon the algal blooms
 in this lake.
   This study has shown that an alum treatment can
 be successful for a long period of time if certain pre-
 cautions are taken.  The major external phosphorus
 loading sources must be reduced to an acceptable
 level prior to the treatment. Preferably, loading rates
 should  be reduced to near the excessive limits  of
 Vollenweider (1976). The life expectancy of an alum
 treatment for lakes that mix frequently will be shorten-
 ed.  The alum layer is flocculant and  the process of
 sediment focusing (Lehman, 1975)  tends to move the
 alum towards the deepest area of the lake during lake
 mixing, for example, Pickerel Lake. The density of the
 sediments should also be considered. If the alum is
 denser than the surface sediments it  will not remain
 on top. While it will prevent migration of  P from the
 deeper sediments, P may enter the water column from
 the mud above the  alum  layer. High sedimentation
 rates will also  shorten the longevity of an alum treat-
 ment because  the alum will be buried.
 REFERENCES

 Born, S.M. 1979. Lake rehabilitation: a status report. Environ.
   Manage. 3:145-53.
                            RESTORATION TECHNIQUES

Born, S.M. et al. 1973. Dilutional pumping at Snake Lake, Wis.
  Tech. Bull. 66.  Wis. Dep. Nat. Resour.
Cooke, G.D., and R.H. Kennedy. 1981. Precipitation and inac-
  tivation of phosphorus as a lake restoration technique.
  EPA-600/3-81-012. U.S. Environ. Prot. Agency, Washington,
  D.C.
Cooke, G.D., R.T. Heath, R.H. Kennedy, and M.R. McComas.
  1982. Change in lake trophic state and internal phos-
  phorus release after aluminum sulfate application. Water
  Resour. Bull. 18:699-705.
Cooke, G.D., M.R. McComas, D.W. Waller, and R.H. Kennedy.
  1977. The occurrence of internal phosphorus loading in
  two small, eutrophic, glacial lakes in northeastern Ohio.
  Hydrobiology 56:129-35.
Dominie, D.R. 1980.  Hypolimnetic  aluminum treatment of
  softwater Annabessacook Lake. Pages 417-23 in Restora-
  tion of Lakes and Inland Waters. EPA 440/5-81-010. U.S. En-
  viron. Prot. Agency, Washington, D.C.
Funk, W.H.,  H.L.  Gibbons  and  G.C.  Bailey.  1980. Lakes
  assessment in preparation for a multiphase restoration
  treatment. Pages 226-37 in Restoration of Lakes and In-
  land Waters. EPA 440/5-81-010. U.S. Environ. Prot. Agency,
  Washington, D.C.
Garrison, P.J., and  D.R. Knauer. 1983. Lake restoration: a five
  year evaluation  of the Mirror and Shadow Lakes Project
  Waupaca,  Wis.  EPA-600/53-83-010.  U.S. Environ. Prot.
  Agency, Washington, D.C.
Gasperino, A.F.,  et al. 1980. Medical Lake  Improvement Pro-
  ject: Success  story. Pages 424-8 in  Restoration of Lakes
  and Inland Waters. EPA 440/5-81-010. U.S.  Environ. Prot.
  Agency, Washington, D.C.
Hennings, R.G. 1978. The hydrogeology of a sand plain seep-
  age lake; Portage County, Wis. M.S.  Thesis. Univ. Wiscon-
  sin, Madison.
Jernelov, A. 1970.  Aquatic ecosystems for the laboratory.
  Vatten 26:262-72.
Knauer, D.R. 1975.  The effect of urban  runoff on phytoplank-
  ton ecology. Verh. Int. Verein. Limnol. 19:893-903.
Knauer,  D.R., and  P.J. Garrison. 1980.  A comparison of two
  alum treated lakes in Wisconsin. Pages 412-16 in Restora-
  tion of Lakes and Inland Waters. EPA 440/5-81-010. U.S. En-
  viron. Prot. Agency, Washington, D.C.
Larsen, D.P., D.W.  Schults, and LW. Malueg. 1981. Summer
  internal phosphorus supplies in Shagawa Lake, Minn. Lim-
  nol. Oceanogr. 26:740-53.
Lehman, J.T. 1975. Reconstructing the rate of accumulation
  of lake sediment: The effect of sediment focusing. Quat.
  Res. 5:541-50.
Peterson, J.O., J.P. Wall, T.L Wirth,  and S.M. Born. 1973.
  Eutrophication control: Nutrient  inactivation by chemical
  precipitation at  Horseshoe Lake, Wis. Tech. Bull. 62. Wis.
  Dep. Nat. Resour.
Vollenweider, R.A.  1976. Advances in defining critical loading
  levels for phosphorus in lake eutrophication. Mem. 1st.
  Ital. Idrobiol. 33:53-83.
                                                    517

-------
                                         Wetlands  and   Lake
                                               Interrelationships
RESPONSES OF WETLAND VEGETATION
TO WATER LEVEL VARIATIONS  IN LAKE ONTARIO
WOLF-DIETER N. BUSCH

LYNN  M. LEWIS

U.S. Fish and Wildlife Service
Cortland, New York


            ABSTRACT

            Water level fluctuations, a naturally occurring phenomena in the Great Lakes, cause a continu-
            ing rejuvenation of lake-influenced wetlands. Two Lake Ontario wetlands (Campbell and Sage
            Creek Marshes) were mapped for 1 ft. contour intervals and habitat-vegetation type. Historical
            habitat/vegetation conditions were evaluated through interpretation of aerial photography. The
            photography was selected to represent water levels different from the current. Habitat types
            defined at Campbell Marsh and their most important herbaceous species include: (1) narrow-
            leaved persistent emergents, Typha glauca; (2) aquatic bed, Ceratophyllum demersum; (3) grass
            sedge, Calamagrostis canadensis; (4) scrub/shrub, Cornus spp; and (5) flooded deciduous forest,
            Fraxinus spp. Habitat types defined at Sage Creek Marsh and their most important herbaceous
            species include: (1) narrow-leaved nonpersistent emergent, Sparganium eurycarpum; (2) broad-
            leaved nonpersistent emergent, Pontederia cordata; (3) aquatic bed, Elodea canadensis and (4)
            grass sedge, Calamagrostis canadensis. Computerized data analysis showed that vegetation
            types occured within rather distinctive elevational ranges. As water levels changed, the area of
            the various habitat types changed, adjusting to both the new water depth  and to the size of the
            area at that depth. In Sage Creek Marsh a large area of narrow-leaved  nonpersistent emergents
            was lost as water levels increased. The greatest loss in Campbell Marsh occurred to persistent
            emergents; however, this loss did not have a linear relationship to annual mean water depth.
 INTRODUCTION

 Shoreline erosion  problems  are  common in  most
 coastal areas, including the Great Lakes. In the Great
 Lakes area the usual structural and nonstructural ero-
 sion control  methods  are  debated  and tried but
 another alternative receives  much  attention—the
 regulation of  lake water levels. Because the Great
 Lakes are a "closed system" with restricted outflows,
 the concept of controlling these outflows to limit the
 range of water level  fluctuation has  received some
 support. Reviews of various  degrees of  water level
 control have concluded that limited control would not
 be economical (Int. Great Lakes Level Board, 1973; Int.
 Lake Erie Reg. Study Board, 1981). These  evaluations
 of water level regulation proposals usually have in-
cluded some attempt to identify environmental pro-
blems (e.g. loss of wetlands). Unfortunately, few  in-
vestigations have been conducted that are directly ap-
plicable  to  measuring  the impacts  on  natural
resources of water level regulation in the Great Lakes.
  The most easily  identified area affected by water
level regulations is the nearshore zone including the
lake-influenced wetlands. The general habitat type of
wetlands is recognized as very important to fish and
wildlife. For example, direct or indirect evidence has
been obtained which shows that at least 27 species of
fish use some of Lake Ontario's  wetlands for spawn-
ing or nursery grounds. (U.S. Fish Wildlife Serv., 1982;
Ontario Ministry Nat.  Resour.,  1981). Also it is  in-
                                             519

-------
  LAKE AND RESERVOIR MANAGEMENT
  teresting to speculate as to the absolute value of the
  energy contribution of wetlands to the great lakes eco-
  system—a system driven mostly by phytoplankton. It
  would seem that the diversity of the organic material
  produced by these  wetlands  would insure  and in-
  crease the absoute value to the receiving system.
    To start the evaluation of water levels and the im-
  pacts of changes in these levels on some Great Lakes'
  wetlands, we selected two wetlands as pilot study
  areas to try methodologies. We examined the size and
  location of the two wetlands,  mapped the contours,
  and used historical aerial photography to map vegeta-
  tion to study how  the vegetation of these particuleir
  wetlands reacted to water level changes. The mapped
  data were also digitized for further evaluation.


 STUDY AREAS

 Campbell and Sage Creek Marshes, located along the
 eastern shoreline of  Lake Ontario, were selected as
 study  areas  (Fig.  1).  Campbell  Marsh  (longitude
 76°07'W, latitude 43°54'N) is located in  the Town of
 Hounsfield, Jefferson County,  New York. It  is  a 28.3!5
 ha (70 acre), streamside wetland that has developed
 where Bedford  Creek empties into  Henderson Bay
 (Geis and Kee, 1977). At present, wetland  vegetation is
 dominated  by narrow-leaved  persistent emergent!;
 (e.g., Typha glauca) (see Cowardin et al. 1979  for Na-
 tional Wetlands  Inventory   terminology).  Other
 vegetative types found in this wetland are aquatic bed
 (e.g.,  Ceratophyllum  demersum  and Myriophyllum
 spp.),  grass  sedge  meadow  (e.g., Calamagrostis
 canadensis and Carex stricta), scrub/shrub  (e.g., Cor-
 nus spp.), and flooded deciduous forest (e.g., Fraxinus
 spp.). Land use in the area includes permanent and
 seasonal residences,  a golf course, abandoned farms,
 and park land.
   Sage Creek  Marsh  (longitude  76°15W, latitude;
 43°31'N) is  located in the town of Mexico, Oswego
                                 Campbell Marsh


                                 Sage Creek Marsh
      CAMPBELL MARSH
                              SAGE CREEK MARSH
                             Mexico Bay
Figure 1.—Location of Campbell and Sage Creek Marshes.
 County, New York, It is a 12.15 ha (30 acre) flood pond
 system that has developed where Sage Creek enters
 Mexico Bay. Currently, vegetation  is  dominated by
 narrow-leaved  nonpersistent  emergents (e.g.,
 Sparganium eurycarpum). Bands of broad-leaved non-
 persistent emergents  (e.g., Pontederia cordata  and
 Peltandra virginica) border the narrow-leaved  nonper-
 sistent  emergent areas;  there  are  also  areas of
 aquatic  bed  (e.g.,  Elodea  canadensis  and
 Myriophyllum  spp.)  and  grass   meadow (e.g.,
 Calamagrostis canadensis) vegetation. Land  use in-
 cludes seasonal and permanent residences, a wildlife
 sanctuary, farms, and  forest.
 METHODS

 The field survey included measuring the elevations
 and plotting the 0.3 m (1 foot) contour intervals on
 topographic maps of the study areas (Fig. 2).  Eleva-
 tions were determined and benchmarks in each wet-
 land  were  established  based on U.S. Geological
 Survey benchmarks. Baselines in  each wetland were
 stationed at 15 m or 30 m intervals according  to the
 judgment of the surveyor. Cross  section lines were
 marked  and cut at  the 15 m or 30 m  stations.
 Measurements were taken using  a Nikon level or a
 Hewlett  Packard   Total   Station  instrument.
 Topographic data were plotted on .025 m =  15 m
 scale maps showing 0.3 m contour intervals.
   The water-covered  areas,  including those directly
 offshore,  were  also  mapped.  Water  depths  were
 measured by wading  directly into the water up to a
 depth of 1m (3 ft.). Transects were extended into the
 offshore waters  every 50 m (150 ft) or 100 m (300 ft.)
 across  the width of the wetland.  Water depth mea-
 surements were  taken every 4 m along the transects.
 Beyond the  .9 m depth, out to 6 m, water depth mea-
 surements were  taken using a boat and a Lawrance
 depth  recorder.  Bottom  contours  were determined
 from the recorded lake levels at the time of the mea-
 surements and were plotted on the topographic maps.
   Vegetation patterns were mapped using 1978  black
 and white and  1979  color aerial  photography and
 verified by extensive ground truthing. Historical  aerial
 photographs were located through  the National  Ar-
 chives  (Taylor and  Spurr,  1973),  New York  State
 Department  of Transportation (1979), Soil Conserva-
 tion Service, and Agricultural Stabilization and Con-
 servation Service. Two years of appropriate historical
 photography were available for Campbell Marsh (1958,
 1966) and 3 years for Sage Creek Marsh (1938,  1955
 1965).
   Habitat types were classified according to Cowar-
 din et al. (1979). Standard photogrammetric  methods
 were used to interpret the photographs and prepare
 the vegetative maps. Mapping was done at a scale of
 2.54 cm  =  15.25 m. These maps were later photo-
 graphically reduced  to a scale of  2.54  cm  = 61  m.
 Vegetation maps were overlaid on  the topographic
 base maps. For immediate use, the areas of different
 habitat types were calculated using a dot grid.  Each
area was calculated  three  times and the values
averaged. The contours and habitat types were digitiz-
ed later for more detailed computer analysis.


 RESULTS AND DISCUSSION

 By  using historical  aerial photography, it could  be
seen that changes in habitat occurred through time in
Campbell and Sage Creek Marshes (Tables  1 and 2).
                                                520

-------
                                                                WETLANDS AND LAKE INTERRELATIONSHIPS
Data from Campbell Marsh show that narrow-leaved
persistent emergent vegetation (Typha glauca) was
dominant in 1978, while grass sedge meadow vegeta-
tion  (Calamagrostis canadensis and  Carex stricta)
dominated  in  1966 and 1958. Vegetation  in  1978
reflected relatively  high water conditions, which  in-
cluded near record  high levels in  the early to mid
1970's (U.S.  Dep. Comm., 1976). Water levels  were
relatively low in 1958, but historic high levels  were
recorded in  the years immediately preceding. The
vegetation in 1958 therefore probably relfected an  in-
termediate condition between the high levels of 1978
and the low levels of 1966. This is shown by the inter-
mediate position of amounts of narrow-leaved persis-
tent  emergents and grass sedge  meadow  present.
Water levels were still relatively low in 1966, following
near record lows of the early 1960's. These data sug-
gest that under high water conditions, narrow-leaved
persistent emergents, such as cattail (Typha glauca),
will dominate Campbell Marsh, while under low water
conditions, grass sedge meadow vegetation will  be
most prevalent.
  At Sage Creek Marsh, narrow-leaved nonpersistent
emergents,  consisting predominantly  of  burreed
(Sparganium  eurycarpum),   were the  dominant
vegetative type in all years. However, in the high water
years of 1955 and 1978, vegetation consisting of a mix-
ture  of aquatic bed  and nonpersistent emergent (both
narrow- and broad-leaved) species become quite pre-
valent (Table 2). Thus in Sage Creek Marsh, it appears
that  high water levels encourage the development of a
mixture of vegetative types, containing aquatic bed
and  emergent species, while  lower water levels sup-
port monotypic stands of narrow-leaved nonpersistent
emergents.
  Other investigators have reported that water level
fluctuations are important to Great Lakes wetland
communities.  Geis (1979) believes that water regime
may be the most important variable in defining the ex-
tent,  species composition,  and  stability of  Great
Lakes wetlands. Jaworski et al. (1979) contend that the
composition and integrity of these wetlands are main-
tained  by the water level fluctuations  that  occur.
Whereas inland wetlands normally undergo an  aging
     SAGE   CREEK  MARSH
                                     CONTOURS
                         CAMPBELL   MARSH
                                                                        CONTOURS
Figure 2.—One foot contours plotted for Campbell and Sage Creek Marshes.


                                               521

-------
LAKE AND RESERVOIR MANAGEMENT
process, proceeding from open water to  dry land
(Odum, 1971), Great Lakes wetlands do not undergo
this succession  because  of the water level fluctua-
tions that occur. Fluctuating water levels cause a con-
tinuing rejuvenation of the wetlands and a variation,
through time, in the amount of coastal wetlands pre-
sent.
                                                    The International Lake Erie Regulation Study Board
                                                  (ILERSB, 1981) reported that wetland vegetative zones
                                                  shift in response to the water levels. They reported
                                                  that in general the "wetter" vegetative communities
                                                  increased  in  size at  the expense  of the "dryer"
                                                  vegetative communities as the water  level increased.
                                                  However, the Board (1981) found, as did we, that the
             Table 1.—Habitat (vegetative) measurements of Campbell Marsh, Lake Ontario, New York.
Area (acres)
Habitat type 1973
(245.3)1
Persistent— emergent 17.0
Grass sedge — meadow 5.4
Scrub/shrub—forest 5.5
Other 7.8
Total 35.7
'Annual mean water level for the year
Persistent emergent = Typha glauca
Grass sedge meadow = Calamagrostis canadensis, Carex stricta
Scrub/shrub/forest = Cornus spp., Viburnum spp., Fraxmus spp.
1966
(244.6)
4.4
24.1
4.1
5.3
37.9




1958
(243.7)
11.1
14.4
3.2
5.7
34.4




Other = stream bottom (no appreciable vegetation), aquatic bed (Ceratophyllum demersum, Mynophyllum spp.),
tana latifo/ia), floating leaved (Lemna spp.), mixed types, and beach bar
1 acre = 0 405 hectares


1978

47.6
15.1
15.4
21.8
99.9




Percent
1966

11.6
63.6
10.8
14.0
100.0




nonpersistent emergent (Zizanu



1958

32.3
41.9
9.3
16.6
100.1




i aquatica, Sagit-

                             Sage   Creek

                                   1938
                                                             P4f
                                   MWL-244.0
                                                                                      1955
                                                                                      MWL-246.1
                                  1965
                                   MWL-243.6
                                                                                      1978
                                                                                      MWL-245.3
Figures.—
leaved nonpersistent emergent).
                           °f vegetative community at Sage Creek Marsh, shown at different water levels (broad-
                                            522

-------
                                                                       WETLANDS AND LAKE INTERRELATIONSHIPS
              Table 2.—Habitat (vegetative) measurements of Sage Creek Marsh, Lake Ontario, New York.
Habitat type
Aquatic bed
Narrow leaved


nonpersistent emergent
Broad leaved nonpersistent emergent
Mixed aquatic
emergent
Other
Total
bed/nonpersistent



1978
(245.3)1
3.5
16.0
1.0

6.1
5.0
31.6
Area (acres)
1965 1955
(243.6) (246.1)
1.8
23.2
1.7

0.0
5.2
31.9
2.4
6.6
4.9

11.6
6.4
31.9
1938
(244.0)
1.2
21.5
0.5

0.0
9.6
32.8
1978
11.1
50.6
3.2

19.3
15.8
100.0
Percent
1965 1955
5.6
72.7
5.3

0.0
16.3
99.9
7.5
20.7
15.3

36.4
20.1
100.0
1938
3.6
65.5
1.5

0.0
29.3
99.9
 'Annual mean water level for the year.
 Aquatic bed = Elodea canadensis, Myriophyllum spp.
 Narrow-leaved nonpersistent emergent = Sparganium eurycarpum
 Broad-leaved nonpersistent emergent = Peltandra virgmica, Pontedena cordata, Sagittaria latifolia
 Mixed aquatic bed/nonpersistent emergent = homogeneous mixture of above three habitat types; inseparable on aerial photography
 Other = open water (no appreciable vegetation), grass meadow (Calamagrostls canadensis), scrub/shrub (Cornus spp.), mixed types, and beach
 1 acre = 0.405 hectares
                                  Marsh    P5f
earlier than the year of evaluation), the comparisons
were significant (P< 0.05). The use of the 5-year mov-
ing mean water level appears justified since although
plant communities can be destroyed in a short time, it
takes 3 to 5 years to reestablish plant communities.
Current plant communities therefore represent not on-
ly current environmental conditions but also those of
recent years.
                                                        RECOMMENDATION
                                   1966
                                    MWL-244 6
Although the  detailed  evaluations of the two pilot
wetland areas produced very useful information, they
are not adequate to measure system-wide (Lake On-
tario) impacts of water level regulations. We determin-
ed that  13  other Lake Ontario  wetlands need to be
studied. These 13 wetlands areas need to be selected
so that each of the three major wetland types and five
geological provinces identified by Geis and Kee (1977)
is represented.
Figure  4.—Computerized presentation of vegetative com-
munity at Campbell Marsh, shown at different water levels
(persistent emergent).

shifting of the vegetative communities depends large-
ly on the geology and, at the lake boundary,  on the
wave energy  transmitted to the  shoreline. Examina-
tion of the Icoation of the Sage Creek and Campbell
Marsh wetland communities (Figs. 3 and 4) in relation
to prevailing water levels showed that they did  not ad-
just in size at the  previous locations  (pulsation), but
became  established at  new  locations that  provided
the needed conditions.
  Comparisons of annual water levels with the size of
the various vegetative communities did not produce
significant correlations.  However, when the  annual
mean  water level was replaced  by a 5-year moving
mean  (starting  with the annual  water level 4 years
REFERENCES

Cowardin, L.M., V. Carter, F.C. Golet, and E.T. LaRoe. 1979.
  Classification of wetlands and deepwater habitats of the
  United States. FWS/OBS-79/31. U.S. Fish Wildl. Serv. Off.
  Biolog. Serv.
Geis, J.W. 1979. Shoreline processes affecting the distribu-
  tion of wetland habitat. Trans. N.A. Wildl. Conf. 44:529-42.
Geis, J.W., and J.L Kee. 1977. Coastal wetlands along Lake
  Ontario and St. Lawrence River, Jefferson County, New
  York. State Univ. New York, College Environ. Sci. Forestry,
  Syracuse.
International Great Lakes Level Board. 1973. Regulation of
  Great Lakes Water Levels: Rep. to the Int. Joint Commis-
  sion plus Appendices.

International Lake Erie Regulation Study Board. 1981. Lake
  Erie Regulation  Study: Rep. to the Int. Joint Commission
  plus Appendices (Appendix F: Environ. Effects).
Jaworski, E.G., N. Raphael, P.J. Mansfield, and B.B. William-
  son. 1979.  Impact  of Great Lakes water level fluctuations
  on coastal wetlands. Off. Water Resour. Technol. U.S. Dep.
  Inter.
New York State Department of Transportation. 1979. In-
  ventory of  aerial photography and other remotely sensed
  imagery of New York State. Map Inf. Unit, Albany.
Odum, E.P. 1971. Fundamentals of Ecology. W.B. Saunders
  Co., New York.
                                                     523

-------
 LAKE AND RESERVOIR MANAGEMENT


Ontario Ministry of Natural Resources. 1981. Fish  Inventory    U.S.  Department of Commerce.  1976. Great Lakes water
  of Oshawa Second Marsh 1979-1980. Lindsay Dist. Admin.      levels, 1860-1975. Nat. Oceanic Atmos. Admin., Nat Ocean
  Rep.                                                      Surv., Washington, D.C.

Taylor, C.E., and R.E. Spurr. 1973. Aerial photographs in tre    U.S.Fish and Wildlile Service.  1982. Lake Ontario Shoreline
  National  Archives.  Spec. List No.  25.  Nat.  Archives      Protection Study (Sage Creek Marsh). Prepared for the U S
  Records Serv., General Serv. Admin., Washington, D.C.         Army Corps Eng.,  Buffalo, N.Y.
                                                    524

-------
LIMITING  NUTRIENT FLUX  INTO AN  URBAN  LAKE BY
NATURAL TREATMENT AND DIVERSION
WILLIAM  D. WEIDENBACHER

PETER R. WILLENBRING

E. A. Hickok and Associates
Wayzata, Minnesota


            ABSTRACT

            The 48 hectare Lake Josephine was being impacted by urban stormwater runoff flowing directly
            into the lake from 237 hectares of its 341 hectare watershed. Runoff from 90 hectares — or 37
            percent — of this direct drainage area was diverted to a 12 ha wetland treatment system for
            pretreatment prior to its discharge into Lake Josephine. The first year three monitors revealed
            the treatment system had removal efficiencies of 62 percent for total phosphorus, 69 percent for
            ortho-phosphorus, 48 percent for total Kjeldahl nitrogen, and 79 percent for total suspended
            solids. Three-year average water quality data from the lake itself was also compared to data ob-
            tained for 3 years prior to the diversion. The comparison revealed that after the diversion, total
            phosphorus concentrations decreased from .092 to .058 mg/l, total Kjeldahl nitrogen concentra-
            tions decreased from 1.25 mg/l to .7 mg/l, ortho-phosphorus concentrations decreased from .038
            mg/l to .03 mg/l, and Secchi depth transparency increased from 3.83 meters to 5.0 meters.
INTRODUCTION

The 48 ha Lake Josephine in the St. Paul suburb of
Roseville,  Minn., is one of several lakes in the Twin
Cities area experiencing  cultural eutrophication.  By
the late 1970's, nearly two thirds of the lake's entire
watershed had been developed. It became evident that
nutrient loading had affected the lake's water quality,
a problem traced to stormwater runoff generated from
a  high-density  residential  neighborhood  flowing
directly  into  Josephine  via  the storm  sewer.  Lake
Josephine has a public beach and is used heavily  for
recreation. Concern among officials and residents led
to a restoration plan.
  Approximately 237 hectares of Josephine's 341  ha
watershed  is a  developed area  that until  1980
generated most of the stormwater runoff that directly
entered the lake via storm sewers. Though Josephine
is bounded by streets and homes along nearly all  its
shoreline,  a substantial wetland still exists across a
roadway to the south and east. This wetland, known
as Little Josephine, is 12 hectares in size. Prior to the
diversion,  runoff from 104  hectares —  roughly the
eastern third  of the lake's watershed — drained into
Lake Josephine.  After the diversion, an additional 90
hectares drained into this wetland pond rather than
directly  into the  lake, thus reducing the direct drain-
age area by 37 percent. The diversion plan was de-
signed to pretreat runoff in this adjacent wetland prior
to its flow into the lake.
  Construction on the  wetlands  project  began  in
August 1980. The 304 m (1000 ft.) diversion pipe and
outlet system was completed in November that year,
at a cost of $212,000. The storm allignment paralleled
the lake's  southern shoreline. Runoff that entered the
pipe flowed east into the wetland. The outlet from the
wetland was built at a point most distant from the ma-
jor points of inflow. This design discouraged short cir-
cuiting  and   increased  water  contact with  the
wetland's  vegetation  and  soils.
   Monitoring the system's  effectiveness  began  in
March 1981, and has continued to present. Efforts to
monitor the results of this diversion focused on water
quality and nutrient removal rates in Little Josephine.
Project designers theorized that Little Josephine, with
its  estimated 45-day retention  time, would lower
nutrient and  suspended solid levels by acting as a
wetland filter.
MONITORING AND RESULTS

Monitoring  indicated  that the wetland did remove
nutrients from the stormwater runoff diverted through
it. The tests to determine what effect this wetland
treatment process had on water quality began in early
spring 1981. Water was sampled at both the inlet and
outlet to determine the difference in water quality as
measured by four major parameters. After 1 year,
these figures revealed a reduction in total  Kjeldahl
nitrogen (TKN) concentrations from 1.64 to 1.12 mg/l, a
removal rate of 32 percent. Total phosphorus (TP) con-
centrations declined  from .20 to .08 mg/l, or 60 per-
cent. Ortho-phosphorus (OP) levels decreased 67 per-
cent, from .12 to .04 mg/l. Total suspended solids were
reduced from 11.58 to 2.25 mg/l, or 81 percent (refer to
Table 1).
  Chloride  concentrations,  not monitored  in  1981,
were measured in 1982. Chloride levels in the wetland
treatment process were found  to be  falling as well,
some 12 percent, from 43 to 38 mg/l.
  Further testing in all areas over a 3-year period rein-
forced the initial findings. The average TKN reduction
over 3 years was 48 percent.  TP fell an average of 62
percent; OP, 69 percent; TSS, 79 percent; and chloride,
19 percent (for 1982-83).
  Water quality  in Lake Josephine was  also tested
after a year and compared to averages compiled over
a 3-year  period prior to the diversion (Table 2). The
results  suggest  that  diverting  runoff had  reduced
nutrient levels.  TKN  fell 44 percent,  from 1.25 to .7
mg/l.  Nitrite (NO2) dropped  38  percent  and nitrate
(NO3), 83 percent. TP concentrations decreased by 35
percent and Secchi  depth transparency increased
from 3.83 to 4.16 meters.
                                                 525

-------
 LAKE AND RESERVOIR MANAGEMENT
                              Table 1.—Results of Little Josephine wetland treatment.
Parameter
and Year
TKN
1981
1982
1983
Total phosphorus
1981
1982
1983
Ortho-phosphorus
1981
1982
1983
Total suspended solids
1981
1982
1983
Chloride
1981
1982
1983
Avg. Influent
Concentration
mg/l

1.64
2.95
1.56

0.20
0.31
0.26

0.12
0.20
0.20

11.58
18.5
25.0

Not Monitored
43
47
Avg. Effluent
Concentration
mg/l

1.12
1.00
1.09

0.08
0.11
0.10

0.04
0.06
0.06

2.25
4.43
4.75


38
35
Reduction
%

32
66
30

60
65
62

67
70
70

81
76
81


12
26
Avg. Three-
Year Reduction
%


48


62


69



79



1Q
I 
-------
THE EFFECTS OF SHOREZONE DEVELOPMENT ON THE NATURE OF
ADJACENT AQUATIC PLANT COMMUNITIES IN  LAC ST. LOUIS, QUEBEC
T.C.  MEREDITH
Department of Geography
McGill  University
Montreal, Quebec,  Canada
            ABSTRACT

            The island of Montreal is part of an archipelago at the confluence of the Ottawa and St. Lawrence
            Rivers. The mixing of waters from two watersheds, the diversity of channel profiles, substrata, and
            natural riparian communities, and the great length of shoreline give the area a marked biotic richness
            and high rate of primary productivity. Longstanding public concern over deteriorating water quality
            and increasing flood hazard has prompted a comprehensive evaluation of the area's aquatic resources
            by the provincial government. However, these studies have tended to focus on large scale engineer-
            ing problems and proposals. Urban expansion and the associate pressures on waterfront land have
            had a persistent and marked effect on the riparian ecosystem. The loss of upland nesting sites, for
            example, has rendered much of the area sterile to several duck species. The effects on the aquatic
            community are less clear. This study was undertaken to determine the effects of shorezone develop-
            ment on the communities of aquatic macrophytes in adjacent areas. It entailed the study of 24 paired
            sites, one site in each pair being off developed shore, the other site being off nearby natural shore.
            Twenty-two macrophyte species were recorded in all, with Vallisneria americana, Elodea canadensis,
            and Ceratophyllum demersum being by far the most abundant. Samples taken off natural sites were
            significantly higher in average species richness and average biomass values than those taken off
            developed shore. Differences in profile, substratum, and water quality were assessed as possible causes
            and it is concluded that changes in depth and perhaps the removal of natural substratum are most
            likely to have been significant factors. The consequences of the observed changes in terms of habitat
            utility for aquatic fauna are discussed briefly.
INTRODUCTION

The St. Lawrence River is met at Montreal by the Ot-
tawa River (Fig. 1). At the confluence is an archipelago
of hundreds of islands, the largest of  which is  the
Island of Montreal. The Ottawa flows into the Lake of
Two Mountains where it drops much of its sediment
load; large sand  bars and beaches  once occurred
where the Ottawa enters the lake, but were exploited
extensively until the late 1960's. The water leaves the
Lake of Two Mountains in one of four channels, each
Figure 1 .—The confluence of the Ottawa and Saint Lawrence
Rivers in the Montreal archipelago. The study area is in Lake
Saint Louis between the Island of Montreal and Me Perrot.
 of  which has  distinctive  natural and cultural at-
 tributes.
   Although it is reportedly possible for the water from
 the St. Lawrence to flow into the Lake of Two Moun-
 tains during  periods  of very high water  in the St.
 Lawrence and very low flow in the Ottawa, this occurs
 very rarely. More normally the waters of the two rivers
 mix in Lake St. Louis along a very conspicuous line
 which moves north or south depending on the relative
 flows of the rivers.
   The aquatic and riparian habitats are extensive and
 diverse within the archipelago: not only is there a mix-
 ing of waters and very long total waterfront length, but
 there are differences in substratum and marked bathy-
 metric features, as well  as rapids which provide oxy-
 genation and ensure year-round open water. However,
 the factor which may contribute most to  the  main-
 tainance of rich  and diverse riparian communities is
 the flood regime of the rivers.
   The St. Lawrence River, having a large watershed,
 tends to  have a regular and predictable  flow. The
 mean flow rate is 8,000 m3s-2 with the 100-year flood
. flow  rate being  only  123 percent greater at  9,900
 m3s-2.  The Ottawa River, on the other hand, has a
 relatively small watershed in which minor variations in
 winter snow accumulation and spring  temperatures
 can cause significant variations in flow: the mean flow
 rate is 2,100 m3s-2, while the 100-year flood flow rate
 is 10,000 m3s-2,  476 percent  greater. (Project Ar-
 chipel, 1981).
   The great variation in flow and the relatively gradual
 shore profile mean there are large areas which are
 periodically  flooded and which have  until recently,
 been left undeveloped  or been used only for extensive
                                                 527

-------
  LAKE AND RESERVOIR MANAGEMENT
  agricultural  or  recreational  purposes.  The  flood
  regime is a dominant factor in determining the nature
  of  riparian communities (Fig. 2)  (Dansereau,  1945,
  1959). The zonation which characterizes much of the
  undisturbed area on the shores of Lac St. Louis pro-
  vides a diversity of habitats for wildlife and  aquatic
  fauna. In a study of a lake down river from the  present
  study area, Masse (1974) reported the importance af
  seasonally flooded areas for eight species of locally
  spawning fish. The importance of the zones for water-
  fowl, shorebirds, herons, mammals (notably muskrals)
  and amphibious reptiles is discussed by Lagace (1976)
  and Baribeau  et al. (1981).
    Obviously, the water resources around Montreal af-
  ford a great potential resource  for the inhabitants. But
  it is a resource that has been sadly abused. Degrada-
  tion of water quality has forced closure of beaches, re-
  quired limits on fish consumption, and caused public
  concern over  the quality of drinking water supplies.
  Moreover, increased urban expansion onto flood plain
  areas has increased financial loss from flood damage,
  and resulted in a call for comprehensive flood control.
   The government of Quebec has formally recognized
  the need for an effective comprehensive management
  plan for the archipelago, and they have initiated hydro-
  logical and ecological studies  as well as public con-
  sultation under what they have called  "Project Ar-
  chipel." While  the objectives of the program are clear-
  ly laudable and are widely supported, some serious
  questions have arisen about the specific targets of the
  study and even about the motivation for it. The ap-
  parent preoccupation of the government with  capital
  intensive large scale engineering solutions rather than
  with policy-based scluticns,  and the apparent desire
  to tie  conservation  and rehabilitation efforts  to a
  hydroelectric power development scheme has caused
  some observers to suggest that the government's in-
  terest is economic rather than  environmental.
   A power project, or any major engineering work that
  would  alter the natural  flow  patterns of  the rivers
  would  have profound effects  on the riparian corn-
Figure 2.—The consequences of seasonal flood level fluc-
tuations. The letters "a" -  "e" on the  right indicate flood
phenology: "a"  being spring levels, "e" being August or
September low  water. The associations named are  from
Dansereau (1945) while the indications of habitat utility are
from Masse (1974) and Baribeau et al. (1981).
  munities, and this threat has caused concern among
  environmental  groups.  But at the same time it has
  done little to placate those whose immediate interests
  are to protect or develop intensively used waterfront
  property. The reaction  of individual land owners  is
  generally to dike and backfill their land. Their inten-
  tion,  and effect, is to eliminate  the  natural flood-
  regulated terrestrial riparian ecosystem.
    This change in the shoreline has obvious effects on
  the utility of the habitat to organisms that use riparian
  upland sites. For example, Lac St.  Louis was once an
  important site for  breeding  ducks. Now, breeding
  ducks have virtually disappeared: not because of the
  degradation  in  water  quality,  but almost certainly
  because of the loss of upland breeding sites (Titman,
  pers.  comm.). If shorelines are modified by extending
  properties into the water, or through dredging from the
  river basin,  it is  likely that riparian land conversion
  may  have significant  effects on  the  aquatic con-
  ponents  of the ecosystem as  well.
    The work described here was undertaken to deter-
  mine  the effects of local shorezone development on
  adjacent macrophyte communities. By identifying dif-
  ferences between developed and undeveloped shore,
  it  is possible,  from existing  literature, to estimate
  what the effects on  associated fauna may be.


  METHODS

 The area  selected for study is  located in Lac St. Louis
 between  the Islands of Montreal and lie Perrot in one
 of the channels leading  from  the Lake of Two Moun-
 tains (Fig. 1). This is  an area fed almost exclusively by
 water from the  Ottawa River  in which a  diversity of
 riparian land uses can be found. There  are many old
 established waterfront properties dating back to the
 19th century, and there is much recent development.
 At a few  points on the Montreal side, and at several
 locations  on  lie  Perrot there are areas that have
 escaped development. It is presumed that they retain
 a natural  plant community both below and immediate-
 ly above mean water line.
   Specific points were selected for study where two
 adjacent  properties  appeared  to have  been innately
 similar but where one had been developed and the
 other left natural. Sites that showed evidence of re-
 cent disturbance  were avoided, so it was assumed
 that approximate equilibria existed  in all study sites.
 Twenty-four  sites were  located  and  studied. All
 samples were taken during late July or August at a
 time when water levels were low and stable.
  At each point the following procedure was adopted:
 A point on the shore  within the bounds of the property
 was randomly selected and a  20 meter transect run-
 ning perpendicular to the shore was defined, using a
 stretched  line, a floating  buoy, and an anchor.
  One-meter intervals along the line were marked, and
 at each point an  observation  of vegetation directly
 below the point was  made using a specially prepared
 viewer. The viewer was submerged along an imagined
 column of water, and  Ihe  first  plant  species en-
 countered in the cross-hairs of  the viewer was record-
 ed. General notes were taken regarding the nature of
 the plant community  in the immediate area of the col-
 umn. Also at each point the depth was sounded using
 a marked  chain  and weight.  Following  the  initial
 survey  of the transect the field notes were evaluated
and  distinct associations or bands of similar com-
 munity types were defined. From within each of these,
further sampling was done as follows: A Secchi disk
was dropped to record turbidity, a thermometer and
                                                 528

-------
oxygen probe were dropped to record temperature and
oxygen profiles, a benthic sample was taken using an
Eckman dredge, and vegetation  was harvested from
within a circular area of about .25 m2. The harvest was
done with a modified garden rake which when pivoted
on a welded protrusion was found to cut or tear and
entangle most vegetation. Vegetation that floated to
the surface immediately following the  harvest was
also collected and included in the sample.
  Plant matter was taken to a laboratory for specia-
tion and recording of dry weight. Substratum samples
were analyzed for organic content  by loss-on-ignition
and then for sand, silt, and clay fraction.
  Analysis of the results entailed comparing species
richness and harvested biomass  for sites adjacent
disturbed and sites adjacent undisturbed shorelines
and  then attempting to  relate  these differences to
possible  physical  determinants such  as  depth,
substratum, and water condition.
RESULTS
A total of 26 species was recorded (Table 1) in  54
samples harvested. Ten of the species occurred only
once and six occurred fewer than five times. Only four
species occurred in more than 10 samples. Vallisneria
americana was by far the most common with 32 occur-
rences, Elodea canadensis next with 23, followed  by
Ceratophyllum demersum (17) and Numphea tuberosa

       Table 1.—Species harvested and number of
               occurrences (Max = 54).
               Species
Occurrences
 Vallisneria americana
 Elodea canadensis
 Ceratophyllum demersum
 Nymphaea tuberosa
 Sagittaria latifolia
 Myriophyllum exalbescens
 Potamogeton amplifolius
 P. epihydrus
 P. richardsonii
 Drepana claudus
 Lythrum salicana
 P. spirillus
 Scirpus validus
 Chara spp
 Fontinalus spp
 Pontederia chordata
 Typha angustifolia
 Butomus umbellatus
 Potamogeton grammeus
 P. spp.
 Heteranthera dubia
 unidentifiable fragments
    32
    23
    17
    11
    10
     9
     6
     8
     6
     6
     4
     4
     3
     3
     3
     3
     1
     1
     1
     1
     1
     5
              WETLANDS AND LAKE INTERRELATIONSHIPS

(11). There was no consistant pattern of association of
species with distrubed or undisturbed sites per se;
however, marked  differences between disturbed and
undisturbed sites were evident.
  Disturbed sites had a mean species richness per
sample of 1.7 spp while natural sites had a significant-
ly greater (p = .001) mean species richness of 4.5 spp.
Disturbed and  undisturbed  sites also differed with
respect to total standing biomass per sample: disturb-
ed sites had a mean dry matter weight of 18 g while
undisturbed sites had values of 64 g (p = .05).
  The  mean  values  for  the  physical  parameters
recorded (Table 2) reveal significant differences bet-
ween developed and undeveloped sites. Undeveloped
sites had shallower mean depths, a  higher organic
content,  and higher clay content (p <.001) Turbidity,
oxygen and temperature values were found to be so in-
fluenced by extrinsic factors such as wind and sun-
light at the time of sampling that the values were
deleted from the analysis.
  If all data are pooled and correlations are sought, it
is seen that species richness per sample and biomass
both relate negatively (p = .05) with depth (Table 3) and
positively with organic content.
Mean  values for three depth categories (Table 4)
demonstrate the nature of the relationships indicated
in Table 3.
  The  relationships between species  richness and
biomass and depth, organic  content and clay content
are shown in regression equations as follows:

Equation 1:
Species Richness = .18Org. + .20 Clay - .42 Depth
R2 = 73

Equation 2:
Biomass = .74 Org. + 1.0 Clay -  .35 Depth
R2 = 26


DISCUSSION

The  decline in  both species richness and harvested
biomass from natural to developed sites suggests, at
least in some senses, that development impoverishes
the aquatic environment. Species richness is normally
a good   indicator  of  habitat  diversity,  and  for
deciduous  species  standing  biomass  late  in  the
season is a good  indicator of net primary productivity.
Both of these environmental indices are negatively af-
fected by development.
  When all data were pooled, depth, organic  content,
and clay content were seen to account for variation in
both species richness and biomass. Regression equa-
tions based on these factors account  for over 70 per-
cent of the observed variations  in species richness,
but only about 25 percent in biomass.
 Table 2.—Comparison of various parameters for all sites. (Table 2A, N=54) and for sites within 5 m of shore (Table 2B,
                                                N = 16).

A)


B)


State
Nat.
Dev.
P of Dif.
Nat
Dev.
P of Dif.
Species
(no .25m -2)
4.5 ±2.9
1.7 ±2.0
.000
5.7 + 3.6
0.8 + 0.8
.001
Biomass
(g.25m-2)
32.3 + 2.8
9.2 + 14.9
.006
77.8 ±103.9
6.8+ 10.6
.05
Depth
(cm)
80 + 38
128 ±35
.000
40 ±28
111 ±44
.002
Organic
percent
20 ±9
08 ±4
.000
22 ±1
8 + 3
.004
Clay
per-
cent
9 + 7
2 + 3
.001
11+8
2±2
.02
Silt
percent
57 + 15
45 + 32
ns
55 ±20
50 ±33
ns
Sand
percent
34 ±16
53 + 31
ns
34 ±24
48 ±33
ns
                                                  529

-------
 LAKE AND RESERVOIR MANAGEMENT


  Table 3.—Correlation coefficients for various parameters. All samples are included in the variables above the line (N = 54) but
  only those samples from areas without rock bottoms are included in the remaining (N =29). Asterisks indicate levels of pro-
                                         bability o'; values above them.
  Spp. Rich.
  Biomass
+ .06
-.32
                                      + .45
Depth
Organic
Clay
Silt
Sand

+ .43
* *
+ .01
+ .12
+ .07
-.09
Dist.
-.52
* * *
+ .73
+ .61
+ .21
-.35
*
Spp.
-.47
* * *
+ .37
*
+ .28
-.04
-.03
Biom.

-.55
-.31
-.08
+ .16
Dep.


+ .26
+ .49
* *
- .53- .97
* * * * *
Org. Silt
N=54
N = 29

+ .06
-.28
Clay
 Table 4.—Species  richness and biomass per sample by
                   depth categories.
Variable


Depth



Range
(cm)

0-49
50 - 100
>100
P of dif.
Species richness
(no. per sample)

6.0
3.8
2.1
.003
Biomass
(gper
sample)
96
13
9

N


6
19
;>9

 N = 54  N = 29

  The relationships between the content of the sub-
stratum and species  richness  and biomass are in-
teresting and warrant further study. It is difficult to i n-
terpret the nature of cause and effect relationships
solely on the basis of  these data.
As Brinson et al. (1981) note, the multiplicity of factors
acting on  wetlands makes it  unlikely that relation-
ships will  be  found with single environmental para-
meters. However, it is clear that shorezone develop-
ment is associated with changes  in the vegetation,
substratum,  and  profile of areas  immediately  off-
shore. It seems  likely that  the diminished species
diversity among macrophytes will reduce the diversity
of microfauna and perhaps of larger animals as well.
The  reduction in  biomass may  indicate a significant
decrease  in primary  productivity,  and  while in a
eutrophic system this  is unlikely to prove  limiting, the
reduced biomass may offer a less dense vegetation
matrix and afford less shelter from predation for small
fish. This hypothesis is now being  tested by further
research.
  But apart from the effect that an increase in  water
depth may have directly at a specific point, the change
in shore profile will clearly affect the riparian eco-
system. Obviously, the horizontal zonation of vegeta-
tion  associated with vertical changes in water level is
compressed or eliminated as  the  steepness of  the
shorezone  profile  increases.   This   change   may
eliminate components of the riparian community used
                                 in  fish  spawning and  rearing, wildfowl and heron
                                 feeding,  muskrat nesting, and  amphibian  reproduc-
                                 tion. The  loss  of the  habitat offshore  following
                                 development may be as great as that onshore.
                                   Ecologists and conservationists are justifiably con-
                                 cerned about the effects of large scale flow regulation
                                 programs.  However, if waterfront development con-
                                 tinues at its present rate there may be little left to con-
                                 serve when the plans for the large scale engineering
                                 programs are finally unveiled.

                                 ACKNOWLEDGEMENTS: This work was funded by  the
                                 Government  of Quebec through their F.C.A.C. program. The
                                 assistance of the members  of  the McGill  University her-
                                 barium in species identification is appreciated.


                                 REFERENCES

                                 Baribeau L,  J.G. Lanouette,  and C. Tessier. 1981. Proble-
                                  matique des interventions de I'homme dans I'ecosysteme
                                  riverain du Lac Saint-Pierre (Quebec). M.I.C.P., Province de
                                  Quebec.

                                 Brinson M., A. Lugo, and S. Brown. 1981. Primary productivity
                                  decomposition and  consumer activity in freshwater wet-
                                  lands. Ann. Rev. Ecol.  Sys.  12:  123-61.
                                 Dansereau, P. 1945. Essai de  correlation sociologique entre
                                  les plantes superieures et les poissons de la beine du Lac
                                  St. Louis. Rev. Can.  Biol. 4: 369-417.

                                 	1959.  Vascular plant  communities  of  Southern
                                  Quebec. Pages 27-54 in Proc. N.E. Wildlife Conf.
                                 Lagace  M., G.  Pageau  and  J.  Dube.  1977.  Milieux  bio-
                                  physiques, frayeres, vegetation et invertebres.  Vol 1, 2.
                                  M.L.C.P. Province de Quebec.

                                 Masse G. 1974. Frayeres a poisson d'eau chaude du couloir
                                  fluvial entre Montreal at le  lac Saint Pierre. M.L.C.P. Pro-
                                  vince de Quebec.

                                 Project Archipel. 1981.  History and Geography of the Waters
                                  around Montreal.  Province de Quebec.
                                Titman, R. 1982. Pers. comm. McGill Univ., Montreal, Quebec
                                  Canada.
                                                  530

-------
                    Destratification   Techniques
PREDICTION OF LAKE RESPONSE TO
INDUCED CIRCULATION
ROBERT A. PASTOROK
Tetra Tech, Inc.
Bellevue, Washington

THOMAS M. GRIEB
Tetra Tech, Inc.
Lafayette,  California

           ABSTRACT
            The outcome of any lake restoration project depends on numerous variables: e.g., lake morphometry,
            initial water quality, composition of the biological community, and engineering specifications of the
            restoration technique. Consequently, a variety of lake responses to restoration attempts can be ex-
            pected, ranging from complete success to at least partial failure. For example, artificial circulation
            has improved water quality in many cases, but has often caused adverse ecological impacts, such
            as increased turibidity or nuisance algal blooms. The benefits of lake restoration can be realized only
            through accurate prediction of lake responses to alternative management schemes or experimental
            manipulations. Numerical classification of previous case history data can be used to enhance this
            predictive capability and to refine lake restoration techniques By applying Mulitple Discriminant Analysis
            to case histories of artificial circulation, we defined the critical attributes of a successful restoration
            project For each response parameter (e.g., dissolved oxygen, algal density, pH), the initial objective
            was to maximize separation of lake restoration groups (i.e., successful or unsuccessful) by differen-
            tially weighting individual morphometric and mixing-system variables in a discriminant function. When
            adequate discrimination is obtained, the discriminant function can be used to predict the response
            of a lake based primarily on physical attributes of the lake (area, volume, depth) and the aeration
            system (air release depth, air flow rate)
INTRODUCTION

Successful  management and  restoration  of  lakes
depends upon the ability to predict lake responses to
alternative management actions. Previous efforts at
forecasting  water  quality and  biological conditions
have  relied  largely on  mechanistic  or empirical
models, which relate system driving variables, e.g.,
nutrient loading and  mixed  depth,  to response
variables such as algal abundance, chlorophyll a, and
fisheries yield (Jorgensen, 1980; Reckhow and Chapra,
1983).

  In this paper, we describe a multivariate statistical
approach,  which complements previous  efforts at
predicting  lake system responses. The approach ex-
plored here uses Multiple Discriminant Analysis to in-
tegrate previous case history data into a  statistical
classification  scheme,  which  defines  the  key at-
tributes of "successful" and "unsuccessful" restora-
tion  projects  in terms  of lake characteristics and
design aspects of the restoration technique. The deriv-
ed discriminant function  can then be applied to
predict the outcomes of proposed  restoration efforts
in other lakes.  Case histories of lake aeration are used
to illustrate the application of our approach.
                                              531

-------
 LAKE AND RESERVOIR MANAGEMENT
 Predictive Models

 The potential success of  alternative restoration op-
 tions may be evaluated by several predictive modeling
 approaches: (1)  ecosystem  models, (2)  subsystem
 models,  (3) simple water quality models,  and  (4)
 multivariate statistical  models  such  as Multiple
 Discriminant  Analysis.   None  of  the   predictive
 schemes discussed here is mutually exclusive. For fix-
 ample, Multiple Discriminant Analysis may be used to
 complement one or  more of  the  other predictive
 methods.  Results  can  also  be  used  to establish
 design criteria similar to those proposed for aeration
 devices by  Lorenzen  and  Fast (1977) and  Tolland
 (1977).

 Ecosystem Models

 Many limnologists have applied complex simulation
 models to characterize the structure and  function of
 lake ecosystems (Jorgensen, 1980). One advantage of
 the holistic systems approach is that a complex sirr u-
 lation model can be made realistic and precise, and
 thus  useful for  specific management  purposes. A
 good ecosystem model incorporates the key para-
 meters and processes necessary  to produce  an ac-
 curate  description  of  lake response behavior.  Sen-
 sitivity  analysis can help  define the relative  impor-
 tance of  the various driving variables in  controlling
 lake responses to  management actions.  Ecosystem
 models hold great promise for lake management, but
 the cost for sophisticated computer operations and an
 extensive data base for model  validation may  limit
 their practical application in some cases.
   Although ecosystem models have been widely used
 for describing and  managing  the eutrophication pro-
 cess, these models have not  yet been  applied  to
 predicting  lake  responses to  artificial  circulation.
 Models of ecosystem subcomponents, e.g., functioral
 relationships between lake thermal structure  and
 phytoplankton growth  (see Subsystem Models), COL Id
 form the basis for developing a full-scale ecosystem
 model of artificial circulation.  Further work on  modal-
 ing of ecosystem responses to  lake aeration  will re-
 quire coupling  of phytoplankton cycles, grazing pio-
 cesses, and predator-prey  dynamics to physical pio-
 cesses (e.g., Parker, 1976; Kemp and Mitsch, 1979).


Subsystem Models

Various models of ecosystem subcomponents have
been applied to analysis  of processes occurring  dur-
ing artificial circulation.  The most useful  approach
has been to construct a model of net photosynthesis
as a function of depth-specific photosynthetic rates,
light  profiles, nutrient concentrations, and  mixed
depth (e.g., Lorenzen and  Mitchell, 1973; Forsberg and
Shapiro, 1980). The photosynthesis equation is thein
rearranged to predict peak algal biomass (e.g., chloro-
phyll a)  as a function of the depth of the mixed layer
(Fig. 1 and  2). As the  depth of the mixed layer in-
creases, the depth-integrated algal biomass (mg Chi
a/m2)  first  increases  because habitat   expansion
enhances  nutrient availability. Eventually, light  limita-
tion decreases algal   biomass as  mixed  depth  in-
creases.
  Whichever factor limits algal  growth,  the average
chlorophyll a concentration decreases dramatically
with increases in  mixed depth because algal popula-
tions  are  distributed  throughout a larger  water
volume. In practice, the lake manager may want to
  balance the potential benefits of a low chlorophyll a
  concentration (i.e., high  transparency) against  the
  benefits of a large algal standing stock integrated
  over depth (i.e.,  increased food  resources for higher
  trophic levels leading  to greater fish yields).
    Details of subsystem models, their predictions, and
  their  limitations  have been  summarized elsewhere
  (Pastorok et at. 1982).  The peak-biomass models have
  been  subjected to preliminary validation by applica-
  tion to actual lake circulation experiences and experi-
  mental manipulation of mixed depth and algal popula-
  tions  in lake enclosures. Although the results are en-
  couraging, refinement of these or similar models will
  require further data collection and model  validation.
                    MIXED DEPTH. 11 METERS I
 Figure 1.—Generalized plot of peak algal biomass as a func-
 tion of mixed depth  for both nutrient  and light limitation
 (adapted from Lorenzen and Mitchell, 1973).
Figure 2.—Relation of peak chlorophyll concentration (C*)
and areal biomass (C*Zm) to mixed depth (Zm) and total phos-
phorus (TP) (adapted from Forsberg and Shapiro, 1980).
                                                 532

-------
                                                                             DESTRATIFICATION TECHNIQUES
Water Quality Models

Design criteria for mixing systems have been based
on numerical hydraulic models for prototype aeration
systems,  models  of thermal stability, and mass
balance models of dissolved oxygen. Several perfor-
mance criteria have been  used  to  design mixing
systems, but the most commonly used ones are ox-
ygenation  capacity  and  destratification efficiency
(Tolland, 1977). Although these latter parameters may
have   some  empirical  value in  designing mixing
systems, their usefulness in  predicting water quality
responses to artificial mixing is limited.
  Davis (1980) has  provided  a combined theoretical
and empirical approach to designing diffused-air mix-
ing systems. The procedure is based partially on con-
sideration of the total theoretical  energy required to
destratify a stratified reservoir. The theoretical energy
required is estimated from an assumed  density gra-
dient plus  incoming solar radiation. For an  example
reservoir of 20 x 106 m3 volume, 20 m maximum depth
and 1.2 x 106 m2 surface area, Davis' procedures result
in a  recommended  70  I/sec  free air flow  rate dis-
tributed through 250 m of perforated pipe.  By com-
parison, the calculations developed by Lorenzen and
Fast (1977) would result in approximately 120 I/sec as
the recommended free air flow.

Multiple Discriminant Analysis

Most predictive approaches are limited by one or more
problems  commonly  encountered  by ecological
modelers: (1) excessive model complexity  prevents
adequate calibration and  validation; (2)  model  para-
meters cannot be easily measured in the real  world; (3)
response behavior of the lake system is affected by
variables or processes that are not explicitly  included
in the model; and (4) criteria for water quality  improve-
ment or biological enhancement  are not well defined.
  Faced with these potential  problems, we have used
a multivariate  statistical  approach to  complement
other  available  models for  predicting  lake system
responses  to management actions (Pastorok et al.
1982; Grieb et al. 1981). Using this approach, existing
case  history data for lake restoration techniques can
be readily  summarized and  integrated.  Multivariate
statistical  methods, such as Multiple  Discriminant
Analysis, can complement mechanistic  or empirical,
process-oriented models by establishing probabilistic
relationships  directly  between   (1) lake response
variables, and (2) design parameters of the restoration
technique and easily-measured lake characteristics.
  Moreover,  multivariate statistical models do not
necessarily require an extensive data base  for each
lake being considered for restoration. Once adequate
case  history data are obtained for a given restoration
technique, the data base can be used to predict the
success of  applying the restoration technique in each
new lake considered, based on a minor data collection
effort. If extensive data are already available for the
new lake, they can be incorporated into the statistical
model, and the results compared  with those  of other
predictive methods.
  Discriminant analysis is a statistical  method for
determining one or more linear  combinations (func-
tions) of a  set of predictor variables (e.g., lake  area,
volume, depth, air flow rate) whose means show wide
differences among entity (lake) groups. The discrimi-
nant  functions are of the form:
                                                 where Y| is the score of an individual entity (lake) on
                                                 the ith discriminant function; the values represented
                                                 by <»jn's are weights assigned to each variable and are
                                                 called linear discriminant  coefficients; and Xn's are
                                                 the values of the independent predictor variables used
                                                 in the discriminant analysis. The functions are formed
                                                 in such a manner so as to maximize the separation of
                                                 the mean values of each group (group centroids).  The
                                                 statistical criterion of the  linear model dictates that
                                                 the  discriminant  coefficients (variable weights)  are
                                                 chosen so as to maximize the ratio of the between-
                                                 groups sum of squares of Y to the within-groups sum
                                                 of squares of Y. Detailed discussions of discriminant
                                                 analysis are presented by  Green and Vascotto (1978)
                                                 and Johnson and Wichern  (1982).
                                                   Once  variables are  identified  which  effectively
                                                 discriminate  among  the   groups of  entities,  the
                                                 discriminant functions can be used as a classification
                                                 index to place new cases  into the defined groups on
                                                 the basis of measurements for these variables. For ex-
                                                 ample, lake morphometric  characteristics and design
                                                 parameters of the restoration technique (e.g., air flow)
                                                 may be used as predictor variables to define discrimi-
                                                 nant functions for groups  of lakes initially separated
                                                 on the basis of their response to the restoration effort.
                                                   Next, the effectiveness of using the  chosen predic-
                                                 tor variables to discriminate among the defined lake
                                                 groups can be tested by reclassifying the lakes using
                                                 the discriminant functions. Since  group membership
                                                 is defined on the basis of the response to the restora-
                                                 tion effort, the ability to predict restoration success
                                                 using the set of  morphometric parameters  and
                                                 measurements of restoration effort is evaluated.
                                                   In  addition  to  using discriminant  analysis as  a
                                                 classification tool, this approach also provides infor-
                                                 mation concerning the relative importance of the in-
                                                 dependent variables  in discriminating among  the
                                                 groups of lakes. Standardized discriminant coeffi-
                                                 cients can be interpreted in much the same manner as
                                                 standardized partial regression coefficients in multi-
                                                 ple regression analysis.
                                                   The conceptual basis of discriminant analysis is il-
                                                 lustrated by the  simple example shown in Figure 3.
                                                                             NOTE. TIM dtocrimliMnt function
                                                                                 •xl» maximize* Mparatlon
                                                                                 of MM two groups.
Y  =
               +  a-,2 X2  + ...... + ain Xn
Figure 3.—Graphic representation for discriminant analysis
of two lake groups: successful (S) and unsuccessful (u)
restoration projects.
                                                  533

-------
LAKE AND RESERVOIR MANAGEMENT
The two groups of lakes shown represent those that
were successfully restored (S) and those that did not
respond to  the restoration  effort (u),  according to
some criteria of success, e.g., increased  DO content
of the lake, reduced algal blooms, or both. The  two
lake groups have different distributions of the predict-
or variables air flow rate (Q/0 and mean depth (D).  The
discriminant axis is defined by the linear function  ajD
 + a2QA = 0. Since air flow rate contributes more to
the separation of the response groups in the example,
the variable QA would be given more weight than mean
depth in the discriminant function (i.e., a2 is greater
than «i). Discriminant functions can be derived for any
number of response groups and any number of predic-
tor variables or discriminant factors (e.g., QA, D).  The
possible number of discriminant functions is equal to
the number of response  groups  minus  one, or  the
number of predictor variables, whichever is less.

DISCRIMINANT ANALYSIS OF  LAKE
AERATION DATA

Case history data on lake aeration were obtained as
part  of a comprehensive evaluation of  lake  aera-
tion/circulation techniques. The  results from a pre-
liminary discriminant analysis of lakes mixed with
diffused-air systems  are presented here as an exam-
ple of our approach  to predicting lake responses to
restoration efforts. Details of methods and results are
available in Pastorok et al. (1982).
Methods

The analysis of lake responses to aeration was based
on a qualitative assessment of changes in a total of '25
                                      physical, chemical, and  biological  parameters. The
                                      response categories consisted of: ( + ) an increase in
                                      the parameter value during mixing as compared with
                                      pretreatment   or  control  values;  (-)  a decrease
                                      resulting from treatment; (0) no change resulting from
                                      treatment; and (?) a variable or questionable response.
                                      In lakes where mixing experiments were performed for
                                      more  than 1   year, individual years were analyzed
                                      separately. Seasonal treatments within  a year were
                                      considered as a single experiment, however. In this
                                      case,  an overall response was assigned to the year.
                                      Justification  of a  qualitative  approach  is  given  in
                                      Pastorok et al. (1982).
                                        To determine if differential responses among lakes
                                      (or among years within a lake) were related to quan-
                                      titative characteristics of the mixing system or mor-
                                      phometry  of  the   lake basin, a  stepwise   Multiple
                                      Discriminant  Analysis was applied to 41  diffused-air
                                      lakes. For each response parameter, lakes with the
                                      same  qualitative response were grouped.  In  some
                                      cases, a group with a limited number of lakes (<4) was
                                      pooled with another group(s). For each response para-
                                      meter, discriminant  analysis then  attempts   to
                                      separate a given group of lakes (i.e., a response group)
                                      from all other groups on the basis of  quantitative
                                      variables of mixing systems and lake morphometry.
                                      The statistical analyses were run on a PRIME com-
                                      puter  using a package program  available  through
                                      SPSS  (Statistical Packages for the Social Sciences).


                                      Results and Discussion
                                      Table 1  summarizes  lake responses to artificial  cir-
                                      culation. Based on a Chi-square statistical analysis,

Table 1.—Summary of lake responses to artificial circulation, diffused-air systems only.
Parameter

AT Aftera

Secchi depth

Dissolved oxygen

Phosphate

Total P

Nitrate

Ammonium

Iron and manganese

Epilimnetic pH

Algal density

Biomass or chlorophyll

Green algae

BI.-Gr. algae

Ratio Gr:BI-Gr

N

45

19

41

17

20

20

20

22

21

33

23

18

25

21

Lake Responses x2

No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
No.
%
+
15
33
4
21
33
80
3
18
5
25
7
35
3
15
0
0
1
5
6
18
5
22
7
39
5
20
11
52
-
30
67
10
53
1
2
5
29
6
30
8
40
13
65
20
91
9
43
14
42
6
27
4
22
13
52
3
14
0
	
—
2
10
2
5
7
41
8
40
3
15
3
15
2
9
8
4
8
24
6
27
7
39
5
20
6
29
?
	
—
3
16
5
12
2
12
1
5
2
20
1
5
0
0
3
14
5
15
6
27
0
0
2
8
1
5

5.00*

6.50*

55.2***

1.60

0.74

2.33

10.5*

33.1***

6.33*

3.71

0.12

1.00

5.57

4.90

aTemperature differential between surface and bottom water during artificial mixing + means AT> 3"C; - means 4T£ 3"C.
*  P< .05 Goodness-of-fit test to uniform frequency distribution for +, --, 0 responses only
*• P<.01
•" P<.001
                                                534

-------
                                                                               DESTRATIFICATION TECHNIQUES
the parameters that appear most responsive to lake
circulation  include  dissolved  oxygen,   iron  and
manganese,  ammonium, Secchi  depth, epilimnetic
pH,  and  A T-after. Although sufficient data were
available to define lake groups on a qualitative basis
for 14 response parameters (Table 1), inadequate sam-
ple sizes (no lakes) for some parameters constrained
the application of  Multiple Discriminant Analysis to
nine response variables (Table 2). The results of Multi-
ple  Discriminant  Analysis  indicate  that  the  first
discriminant  function  was  significant  for  only
epilimnetic pH (Wilks' lambda test; P <0.05). Air flow
was  an important variable  in many cases, but mor-
phometric variables (surface area or mean depth) con-
tributed the most to separation of groups. Aeration in-
tensity (i.e., air flow divided by area, QA/A, or air flow
divided  by volume, QA/V) was an important discrimi-
nant variable for a number of response variables.
  The percentage of total cases correctly  classified
by the discriminant functions was not uniformly high
(Table 2).  For some response parameters (biomass or
chlorophyll a,DO, green: blue-green algae ratio, Secchi
depth),  a high  percentage  of  lakes was correctly
classified, but the discriminant  function was  not
statistically significant.  Overall, the poorly classifed
lakes were often  in the variable (or questionable)
response  group.   In  many  cases,   the  beneficial
response group (e.g., increase in DO, decrease in algal
density)  was well classified. These  results suggest
that  Multiple Discriminant Analysis holds promise for
prediction of  lake resonses to artificial circulation.
Furthermore,  the  results of Multiple  Discriminant
                 Analysis are in agreement with results obtained from
                 other predictive methods. For  example,  empirical
                 criteria (QA/A) that are currently used to design aera-
                 tion systems (Lorenzen and Fast  1977; Tolland 1977)
                 incorporate variables that have considerable predic-
                 tive value in determining  lake responses. Also,  those
                 response parameters  which  theoretically  would be
                 most affected by circulation were identified by  Multi-
                 ple Discriminant Analysis.
                   Improvement  in predictive power could be achieved
                 by  better definition of  lake  response  indices and
                 refinement  of  data  quality.  In  this  exploratory
                 analysis, all response groups  ( +, - ,0,?) were entered
                 separately in the analysis for  a single response para-
                 meter (e.g., biomass or Chi a).  It may be useful to form
                 only two response groups, defined as "successful" or
                 "unsuccessful"  restoration attempts. Thus, +, 0, and
                 ? lakes would be pooled  to form the "unsuccessful"
                 group for the biomass or chlorophyll a response para-
                 meter.
                   Depending on the specific goals of a restoration
                 project, a response index could be constructed from a
                 composite of several variables (e.g., dissolved oxygen,
                 chlorophyll a, and ratio of green to blue-green algae).
                 Classification  and  other  multivariate  techniques
                 could be  used to help define this composite  index
                 (Grieb et al., 1981). Whenever adequate response data
                 are available, a quantitative response index can be
                 constructed to define group membership for applica-
                 tion of Multiple Discriminant Analysis and to evaluate
                 the results of lake  restoration efforts. Future efforts
                 should  concentrate  on  developing  quantitative
             Table 2.—Results of multiple discriminant analysis of lake responses to artificial circulation.
 Response Parameter
                                    Total
                                    Cases
                                   Correct13
                            Group
                                                                                              Important
                                                                                             Discriminant
                                                                                               Factors
Algal density

Biomass



or Chi a


Blue-greens


AT After



0.07

0.19


0.08


0.23
67

81


72


69
No.
%<=
No.
%

No.
%

No.
6 14
50 93
5 6
100 83

5 13
60 85

15« 30d
8 5 QA/A
50 40 QA/V
4 6 Area
75 67 Volume
QA
7 Max. depth
57 Area
J Mean depth
- - QA/A
DO


Green algae


Epilmnetic pH

Gr:BI-Gr ratio

Secchi
                         0.09
                         0.65
                         0.04
                         0.19
                         0.34
83
72
85
81
84
No.


No.


No.

No.

No.
 47


 33
 94


  7
 71
 11
 82

  4
100
                                                                   80
  4
 75


  9
100
                                                                   10
                                                                   80
 Probability level indicates significance of first discriminant function
 bPercent of all cases correctly classified.
 cPercent of withm-group cases correctly classified.
 dTemperature differential between surface and bottom water greater than 3"C.
 eTemperature differential (AT) between surface and bottom water less than or equal to 3°C.
 'Brackets indicate groups that were pooled to increase sample size
 Note
 QA = Air flow rate
 A = Lake surface area
 V = Lake volume
 8
38


 7
71
          80
           5
          80
                                                                                      11
                                                                                      73
  Volume

   QA/A
 Air depth
   QA/V
   QA/A
   Area
Mean depth
    QA
Max. depth
Mean depth
  Volume
 Air depth
                                                   535

-------
LAKE AND RESERVOIR MANAGEMENT
response criteria (e.g., Porcella et al. 1980) and obtain-
ing data for defining a  "successful" restoration pro-
ject and evaluating various predictor variables.

ACKNOWLEDGEMENTS: We are grateful to D.B.  Porcslla
and  G.N.  Bigham for comments on an early draft of "his
paper. Funds for the initial synthesis of lake aeration data
were provided by U.S. Environmental Protection Agency and
U.S. Army Corps of Engineers.


REFERENCES

Davis,  J.M.  1980. Destratification of  reservoirs—a design
  approach  for perforated-pipe  compressed-air systens
  Water Serv. 84:497-504.

Forsberg,  B.R.,  and J.  Shapiro. 1980.  Predicting the algal
  response to destratification. Pages 134-9 in Restoration of
  Lakes and Inland Waters. EPA 440/5-81-010. U.S.  Environ.
  Prot. Agency, Washington, D.C.

Green, R.H., and G.L.  Vascotto. 1978. A method  for  the
  analysis of environmental factors controlling patterns; of
  species composition  in aquatic communities. Water Res
  12:583-90.

Grieb, T.M.,  D.B. Porcella,  T.C. Ginn, and M.W. Lorenzen.
  1981.  Classification and analysis of cooling impound-
  ments: an assessment methodology using  fish standing
  crop data. Symp. Surface Water Impoundments, Am. Soc
  Civil Eng.  2:482-94.

Johnson, R.A., and D.W. Wichern.  1982. Applied Multivariate
  Statistical Analysis. Prentice-Hall, Inc., New Jersey.
Jorgensen, S.E. 1980. Lake Management. Water Dev., Supply
   Manage. Ser. Pergamon Press, New York.
Kemp, W.M., and W.J. Mitsch. 1979. Turbulence and phyto-
   plankton diversity: a general model  of the "Paradox of
   Plankton." Ecol. Model. 7:201-22.

Lorenzen, M.W., and A.W. Fast. 1977. A guide to aeration/
   circulation techniques  for lake  management. Ecol. Res.
   Ser.  EPA-600/3-77-004.  U.S.   Environ.   Prot.  Agency,
   Washington,  D.C.

Lorenzen, M.W., and R. Mitchell. 1973. Theoretical effects of
   artificial destratification on algal production in impound-
   ments. Environ. Sci. Technol. 7:939-44.

Parker, R.A. 1976. The influence of eddy diffusion and advec-
   tion on plankton population systems. Int. J Sys  Sci
   7:957-62.

Pastorok, R.A., M.W. Lorenzen, and T.C. Ginn. 1982. Artificial
   aeration and oxygenation of reservoirs: a review of theory,
   techniques, and experiences. Tech. Rep. E-82-3, U.S. Army
   Corps Eng., Waterways Exp. Sta., Vicksburg, Miss.
Porcella, D.B., S.A. Peterson, and D.P. Larsen. 1980. Index to
   evaluate lake restoration. J. Environ.  Eng. Div. Am. Soc.
   Civil Eng. 106:1151-69.

Reckhow,  K.H.,  and S.C. Chapra. 1983. Engineering  Ap-
   proaches for Lake Management. Ann Arbor Sci. Publ. Inc.,
   Ann Arbor, Mich.

Tolland, H.G. 1977. Destratification/aeration in  reservoirs.
   Tech. Rep. No  TR50. Water Res. Centre,  Mendenham,
   United  Kingdom.
                                                     536

-------
 THOUGHTS ON SELECTION AND DESIGN OF RESERVOIR
 AERATION  DEVICES
 PERRY L. JOHNSON
 U.S.  Bureau of Reclamation
 Denver, Colorado


             ABSTRACT

             Alternative devices for reservoir aeration are briefly reviewed. It is noted that each device is best suited
             for particular applications and objectives. It is recommended that care be taken by the designer to
             select the appropriate device for the particular application. Advantages and disadvantages of the various
             devices are given along with representative destratification and oxygenation efficiencies.  Design con-
             siderations are discussed including techniques for sizing units, the effects of inflows and releases,
             the effects of reservoir stratification, evaluation of aeration impact on the reservoir temperature regime,
             and the possible development of nitrogen supersaturation
 DEVICE SELECTION

 Many devices have been developed to aerate reservoir
 and lake water. They may be pneumatic, mechanical,
 or use buoyant water or molecular oxygen. However,
 in all cases, selection considerations are the same:
 The appropriate device for use at any site is a function
 of the specific site with its characteristics and the ob-
 jectives of the aeration treatment. Treatment devices
 may destratify or influence the temperature and water
 density distribution within a reservoir or they may not.
 Thus the influence of the various devices on  both
 temperature and oxygen distribution should  be  con-
 sidered. As  an initial step, the characteristics of the
 reservoir to  be treated should be noted. In particular,
 the following should  be considered:
   1. Reservoir volume and the volume of the oxygen-
 depleted reservoir to be treated.  This volume will in-
 fluence the extent of  the problem and thus the size of
 treatment system required. Typically,  a larger reser-
 voir will require a larger aeration system. An option, if
 the quality of release water is of particular interest
 and if the quality of reservoir water is less critical, is to
 treat  only the portion of the reservoir around  the
 outlet. This option,  of course,  yields treatment
 demands that are less than required for the full reser-
 voir and  thus yields  reduced system  size and cost.
 Partial treatment also has, by its nature, limits on its
 potential effectiveness which are a function of dis-
 charge, treatment water quality objectives, and size of
 the treatment zone. Some devices such as wind driven
 aerators,  hydraulic  guns,  and  mechanical jetting
 destratifiers  have  limited  influence  and thus  are
 specifically suited for smaller lakes or for partial treat-
 ment of larger lakes.
  2. Oxygen demand and thus the hypolimnion ox-
ygen decline rate. Oxygen decline rate in conjunction
with minimum allowable oxygen levels and in conjunc-
tion with the initial DO levels, is the other factor that
predominantly influences the extent of the reaeration
problem and thus the size of the treatment system re-
quired. Typically,  the desired unit reaeration  rate
times the volume of the oxygen-depleted water of in-
terest yields a bulk required reaeration or oxygenation
rate. The reaeration  system must then be sized to
meet this bulk rate.
  As discussed under volume, some devices are best
suited  for small applications and thus to meet rela-
tively small oxygen demands. Consequently, total de-
mand to be supplied may be a factor in device selec-
tion. In addition, the reaeration efficiency and effec-
tiveness of some devices are functions of the initial,
pretreatment dissolved  gas levels within  the water.
For example, DO increases of 3 mg/l from an initial
level of 0 mg/l can be relatively easily achieved using
draft  tube aeration.  However, it is quite difficult to
achieve the same  increase using  draft tube aeration
from an  initial level of, say, 3 mg/l. Higher initial DO
levels reduce the oxygen deficit between the  satura-
tion level and existing level which  reduces the driving
force for gas transfer. In particular, this may be critical
for devices for which the  gas transfer occurs under
relatively low pressure.
   3.  Reservoir depth and the depth at which increas-
ed DO is required.  Some devices such as mechanical
surface aerators have a limited range of vertical  in-
fluence.  Likewise,  some devices  may be used  over
limited vertical ranges to push higher DO epilimnion
water down into the hypolimnion (locally lower the
thermocline). With these devices,  for example, if the
withdrawal outlet is at a shallow depth (say 50 feet or
less  below the epilimnion, water may be jetted down
to the  outlet  and   thus  epilimnion water  or a
epilimnion-hypolimnion  water mix is released.  For
deeper outlets, other treatment devices would be ap-
propriate.
   Reservoir  depth  or submergence  depth  on  the
device may also influence operating efficiency. For ex-
ample, a pneumatic  line diffuser that aerates by en-
training hypolimnion water into a rising bubble and
water column and  bringing that water to the surface
for mixing with epilimnion water, functions more effi-
ciently with  a long bubble plume path through the
hypolimnion  and  thus  substantial   entrainment  of
hypolimnion water. In cases where the hypolimnion
depth  is  small relative  to  the  epilimnion  depth
(shallow stratified reservoirs), the bubble path through
the hypolimnion is short. Consequently, entrainment
of hypolimnion water is relatively minor and system ef-
ficiency is substantially reduced. Consequently, pneu-
matic diffusers are more efficient and economically
more competitive in deeper reservoirs.
  4.  Reservoir flowthrough or the size, temperature,
and DO levels of inflows and releases. At  some sites
                                                 537

-------
 LAKE AND RESERVOIR MANAGEMENT
 substantial flowthrough of low DO water occurs. In ef-
 fect, devices may  be required to aerate the  flow-
 through as well  as  the reservoir. Thus, substantially
 larger devices may  be required than those indicated
 by the reservoir volume and oxygen demand. At other
 sites substantial high DO flowthrough may have1 a
 freshening  effect and  may reduce  required hypolim-
 nion aeration. The influence of flowthrough is a func-
 tion of not only discharge and DO concentration, but
 also of the stratified flow dynamics of the reservoir. If
 the inflow is warm and stays on the reservoir surface
 and if withdrawal is also from the surface, then a stag-
 nant hypolimnion may result in which DO decline is
 maximized. On the other hand, if the inflow is cold and
 high in DO and if withdrawals are made from the bot-
 tom of the reservoir, then the inflows will tend  to
 replace and freshen the hypolimnion waters and thus
 minimize oxygen decline.
   Thus inflows and  releases should be considered in
 selecting and sizing treatment systems. As can  be
 seen, it is appropriate to use a hydrodynamic reservoir
 model to evaluate residence times, freshening effects,
 and thus to guide evaluation of expected hypolimnion
 oxygen decline rates. It should be noted that many
 reaeration devices  also yield  destratification.  The
 degree and nature of this destratification  are func-
 tions of the device type, method of device operation,
 and frequently of the strength and profile shape of the
 reservoir  stratification itself. Somehow  this device
 destratification should be incorporated in the hydro-
dynamic model if a clear picture is to be  obtained.
 However,  for  many   devices  the destratification
mechanics  including destratification  efficiency and
resulting destratification circulation patterns are not
known. Consequently, including the destratification in
a hydrodynamic model is very approximate.
   In addition, a treatment option other than partial or
                                   complete reservoir  reaeration  is  to  treat  only the
                                   release water as it is being withdrawn from the reser-
                                   voir. This is beyond the scope of this paper, but nerver-
                                   less should be considered with other treatment alter-
                                   natives.  For this  case, the reservoir water  quality
                                   would be allowed to  deteriorate. Selective withdrawal,
                                   localized aeration of the withdrawal within the reser-
                                   voir, aeration of the  release flow as it passes through
                                   energy dissipators or other types of hydraulic struc-
                                   tures, and/or turbine or  draft tube aeration could then
                                   be used to increase  the DO level in the release water.
                                   It should be recognized that this option may produce
                                   poor quality water that may limit recreational, fishery,
                                   or other uses of the  reservoir. Likewise, substantially
                                   more  treatment of the withdrawal  than just aeration
                                   may be required to  obtain acceptable water  from  a
                                   poor quality reservoir. The Metropolitan Water District
                                   of Southern California has  reported that, in their case,
                                   it is more economically efficient  to  maintain good
                                   quality water within the  reservoir  through  aeration
                                   than to treat poor quality water as it is  withdrawn
                                   (Pearson et al.  1976). At other sites and in particular
                                   for hydroturbine power releases and stream releases,
                                   aeration of the release flow is a common technique.
                                     As  noted  earlier, a final consideration is whether
                                   modification of the temperature structure of the reser-
                                   voir can be tolerated. Many treatment devices function
                                   by mixing the low DO hypolimnion water with the high
                                   DO epilimnion water. This tends to yield efficient aera-
                                   tion in that the  large  reservoir water surface becomes
                                   the  primary oxygen transfer interface. This, however,
                                   warms the hypolimnion and cools the epilimnion.  It
                                   may be that because the fishery, either in the reservoir
                                   or in the downstream channel, some other considera-
                                   tion, these temperature  changes cannot be tolerated.
                                   In these cases  hypolimnion aerators that do not mix
                                   the reservoir may be  used.
                                Table 1.—Comparative reaeration device features.
 Device
 Pneumatic diffusers
   or diffused air
   bubble plumes

 Diffused hydraulic
   buoyant  water
   plumes
 Mechanical pumping
   with free jets
 Hydraulic guns
 Air life-limno (air
   or molecular
   oxygen driven)
   Efficiencies
 Molecular oxygen
   injection through
   fine bubble
   diffusers
                                            Adva ntages
                                                 Disadvantages
                                                        Reference
                                  Devices for in-reaervoir reaeration through mixing or destratification
 Mixing 0-8
  percent
 Aeration 0.6-
  3.9 kg/kWh
 No field proven
  efficiencies,
  may be greater
  than pneumatic
  diffusers
 Aeration less
  than 0.6 kg/kWh
Aeration 1 kg/kWh
Aeration 0.2-
  0.6 kg/kWh
Aeration 14-55
  percent oxygen
  transfer 0.3-
  0.7 kg/kWh
Proven suitable for deep
  reservoirs, relatively
  low capital and operating
  cost for deep reservoirs
May be used in deep reser-
  voirs, potentially offers
  high efficiencies
Simple equipment, may push
  surface water down to
  intake - replaces selec-
  tive withdrawal
Efficient mixing of
  upwelled water with
  surface, may use small
  compressor
May yield nitrogen super-
  saturation, may have
  clogging problems and
  require filtering of air

Unproven, concept has not
  been field applied, may
  cause nitrogen super-
  saturation


Jetting effective only for
  shallow (less than 60 ft
  deep) applications and rela-
  tively small volumes

Moves relatively low volumes
  of water, little gas trans-
  fer from bubbles. Best for
  small applications
                                      Devices for hypolimnion aeration with no reservoir mixing
Allows temperature strati-
  fication to remain
  undistrubed
Allows temperature strati-
  fication to remain undis-
  turbed, no nitrogen super-
  saturation, high transfer
  efficiency
Relatively low efficiency
  with relatively high
  capital cost, air-driven
  units may yield nitrogen
  supersaturation
High operating and capital
  cost
Johnson (1980)
King et al. (1983)
Davis (1980)
AWWA (1971)

Dortch (1979)
Garton et al. (1978)
Toetz (1979)
Holland (1983)


Hydraulic Research
  Station (1978)
Bernhardt (1974)
Fast et al. (1975)
Fast et al. (1976)
Speece (1973, 1976)
                                                  538

-------
                                                                             DESTRATIFICATION TECHNIQUES
   Table 1 contains an incomplete but representative
 list of reservoir reaeration treatment device options.
 Included  are  information on  device efficiencies as
 reported in the literature, a brief description  of the
 potential  advantages  and  disadvantages  of the
 various devices, and a list of useful references on the
 specific device. Figure 1 contains illustrations of the
 devices  mentioned  in  Table  1.  In addition to the
 specific references  cited in Table 1,  several  more
 general references that are quite useful are avail.able.
 They include  King,  1970; Lorenzen and Fast,  1977;
 Pastorok et al. 1981; and Bohac et al. 1983.

 DEVICE DESIGN
 With  the selection  of a device  or devices,  either
 feasibility  or more detailed designs may be  under-
 taken. As an initial step, the size or extent of the pro-
 blems to be treated  must be defined. To do this, the
 size of the impoundment to be treated and  thus the
 volume of water to  be treated should  be evaluated.
 Likewise, the  expected oxygen demand in the un-
 treated impoundment and desired or acceptable ox-
 ygen  decline rates in the treated  reservoir should be
 defined. Expected oxygen demand in the untreated im-
 poundment may be evaluated through observation of
 historical data  for that  impoundment,  through ob-
 servation of the oxygen response in similar  impound-
 ments, or through the use of DO prediction mathe-
 matical models such as the model of Ford et al. (1980).
 It should first be noted that oxygen demand observed
 from historical data or similar reservoirs' demand will
 vary over the short term, for example, because of the
 decay of algae blooms or flooding; and over the long
 term, for example, resulting from reservoir maturing or
 seasonal variations.  A decision must be  made as to
 whether reaeration system design should be based on
 typical expected oxygen decline rates or on  some ex-
 treme value. Sizing a system  based on  an  extreme
 decline rate will yield a system that is oversized for
 most cases and thus may have both excessive capital
 and operating costs. However, sizing a system based
 on a typical or mean decline rate will yield a system
 that  is unable to meet all desired demands. The
 design DO reaeration rate selected generally depends
 on how critical the reaeration is.
   It appears that historical data for the reservoir of in-
terest  may  supply the best estimate of initial un-
treated DO. Similar reservoirs can supply a good esti-
mate of  untreated conditions. However, care should
be taken to ensure sufficient similarity. The  com-
parison impoundment should be in the same vicinity
as the impoundment of interest, should experience
fairly similar climatic  conditions, and should be of
similar depth or at least deep enough to allow similar
thermocline and hypolimnion development. The com-
parison reservoir should experience similar inflow and
release discharges. The relative influence of the flow-
through should be similar and thus the relative magni-
tude of the  discharges versus reservoir volume and
the stratified flow response  of the flows in  the reser-
voirs should  be similar. This also implies  the need,
where  multiple release structures  exist, to  have
similar release operating characteristics for the two
sites.
   Finally, for a good comparison of DO response, the
oxygen demand of  the two reservoir hypolimnions
should be similar. This generally implies that the im-
poundments  have  similar  nutrient  characteristics,
biological productivity, and that they are biologically
managed in  similar ways.
   The final technique for determining the initial or un-
treated state of a reservoir is by using mathematical
models. Numerous models are available for predicting
temperature  and dynamic  response of a  reservoir.
Available models   include  those of  Edinger and
Buchak  (1979),  and Norton et al. (1973).  In recent
years, models such as that of  Ford et al. (1980), have
been developed to predict  biological and  chemical
response including DO. Use of the models requires
substantial data bases; the  models are  best applied
where  sufficient data  exist  for verification.  With
limited input data and with no historic profiles to help
fit the model, only approximate predictions  and guid-
ance can be obtained.
   After establishing the untreated  DO  state of the
reservoir, the next step is to select a desired  minimum
acceptable DO state that could result when reaeration
is  used. The desired uses of the water and their impli-
cation on required DO levels  should be identified. For
example, if the objective is to prevent the development
      PNEUMATIC
      DIFFUSERS
                                           PUMPING WITH
                                            FREE JETS
                                                      Pump
                   FINE BUBBLE
                   MODECULAR
         HYDRAULIC    OXYGEN
            GUN      DIFFUSER
 AIR  LIFT
HYPOLIMNION
 AERATOR
                                                                )  C
                                                                 -
                                        •Air
                                         d iffuser

                                 -'Xv-Water in
Figure 1.—Aeration devices.
                                                 539

-------
 LAKE AND RESERVOIR MANAGEMENT
 of anaerobic conditions, a level of 2 mg/l  might be
 established. However, if the objective is to maintain a
 trout fishery, a minimum acceptable hypolimnion DO
 level of 5 mg/l might be established. Noting then that
 the epilimnion water will tend to be saturated in  DO,
 and considering the degree of destratification or varia-
 tion away from a traditional two-layer density profile
 (discussed later  in the paper) that would result, esti-
 mated  minimum acceptable DO profiles can be ob-
 tained. These profiles are an epilimnion-hypolimnion
 composite with a transition  between the two layers.
 Because hypolimnion DO levels will decline from their
 saturation  value  at the start  of  the stratification
 season to their minimum acceptable values just p-ior
 to fall turnover, the minimum acceptable profiles just
 developed represent the profiles that would exist prior
 to turnover. By comparing the total DO content of the
 reservoir for the saturated spring condition and for the
 minimal acceptable condition, an acceptable total DO
 mass decline is  obtained.  This  acceptable  total  DO
 mass decline is then divided by the expected strati-
 fication season length to obtain an  acceptable  DO
 decline rate. For example, if it is found that an accep-
 table total DO mass decline of 5  x 105 kg O2 could oc-
 cur  over the  stratified season  and if the expected
 stratified  season  length is 200  days, an acceptable
 total DO  decline  rate of 2,500  kg  O2/day could be
 tolerated. It should be noted that if destratification ac-
 companies the reaeration then the stratified season
 will be shorter than it would be in the untreated reser-
 voir.
   A similar computational process can be conducted
 for the untreated reservoir. The total DO mass content
 of the reservoir at the start of the stratified season can
 be computed. Likewise, a total DO content at the end
 of the stratified season or when the hypolimnion goes
 anaerobic  can also be computed. Note that  once  the
 hypolimnion goes anaerobic, the total oxygen decline
 rate that results  in the reservoir declines simply  be-
 cause there is no oxygen left to be depleted. Again, by
 taking the difference between the total DO mass at
 the start  of the  season and the depleted total  DO
 mass, the total DO mass decline that occurs in the un-
 treated reservoir is found. When this is divided by 'he
 time period, the stratified season length, or the time
 length to an anaerobic hypolimnion, the total oxygen
 decline rate in the untreated reservoir is found (for'ex-
 ample, 3,700 kg O2/day). The difference between ihe
 total DO decline rates (3,700 to 2,500 kg O2/day) repre-
 sents an oxygenation or reaeration rate that must be
 supplied by the device.
   It  should be noted that the untreated DO levels in
 the  reservoirs which  were obtained  either  from
 historic data, a  similar reservoir, or  mathematical
 models do include the influence of  inflows a.nd
 releases. One exception, however, is in cases whore
 heavy flowthrough of oxygen-depleted water occurs.
 In these cases, it may be required to size the reaera-
 tion system to treat not only the reservoir, but the flow-
 through as well.  Noting flowthrough  volumes s.nd
 desired DO levels, estimates of  required additional
 reaeration for the flowthrough can be obtained.
   With a knowledge of the required total oxygenation
 or reaeration rate  (1,200 kg O2/day for the example),
 the reaeration system  may then  be sized. Typically,
 some sort of oxygenation efficiency data are available
 for the reaeration devices being considered. These ef-
ficiencies may take the forms of a crude bracketing of
observed efficiencies (as shown  in  Table 1) or may
take the form of  more exact efficiencies such  as
shown on  Figure 2. The Figure 2 efficiencies were ob-
tained by the Bureau of Reclamation for straight line
pneumatic diffusers submerged at a depth of 46 m
(King et al. 1983). Knowing the required reaeration rate
(1,200 kg O2/day) and knowing a device efficiency (for
example,  1.5 kg  02/kWh), a required energy consump-
tion rate  is obtained (1,200/1.5 or  800  kWh per day).
This may then be used in conjunction  with available
literature  on the particular device of interest to size
the required reaeration system. Note that this process
contains substantial potential for error.
  The efficiencies reported in the literature for various
devices show substantial scatter. Not only are device
efficiencies a function of the hardware (the particular
compressors, pumps, motor, and plumbing  used), but
they are likely also a function of the particular applica-
tion geometry.  For  example, diffused  molecular ox-
ygen or air can have very different efficiency charac-
teristics depending  on whether the gas is  uniformly
distributed or concentrated at  local points.  Stratifica-
tion strength and reservoir depth and device sub-
mergence can also substantially affect resulting effi-
ciencies.  In most cases, there are  no guidelines for
evaluating the  effect of these various parameters.
Analytical  techniques  for  calculating  gas transfer,
bubble plume flow entrainment, water jet flow entrain-
ment, and the like are available and  may be used to ap-
proximately  adjust  efficiencies  to  satisfy  the
geometries and  flow conditions of a particular situa-
tion.  Good general analytical  references include
Neilson (1972) and Pastorok et al. (1981).  Studies of
specific devices may also give guidance.
  With the system sized, hardware can be designed. It
is strongly recommended that the designer contact or
visit field  sites where the particular reaeration device
type of interest is in use. Much fabrication, operation,
and maintenance knowledge can be obtained through
experience.
  Two final factors should be considered in reaeration
system design. First, if destratification results from
system operation, a satisfactory reservoir temperature
regime may or may not be obtained. Typically, destra-
tification cools the epilimnion and warms  the hypo-
limnion. If the reservoir is intended for a temperature
dependent use, such as for a coldwater fishery, then a
conflict may result. Sufficient  reaeration to yield the
desired DO levels may produce unacceptable temper-
atures. For some devices, destratification efficiencies
are available. Thus a technique similar to the one
                                                     Figure 2.—Pneumatic diffuser efficiency curves.
                                                 540

-------
                                                                                  DESTRATIRCATION TECHNIQUES
presented in this paper for reaeration may be followed
to evaluate the destratification  influence.  Initial un-
treated temperature profiles  can be determined, un-
treated reservoir stabilities computed, destratification
influence  on stability evaluated (using  system size
determined from the reaeration computations and ap-
propriate destratification efficiencies), and the impact
on temperature profiles  found. If the temperature im-
pact  is unacceptable, a treatment device that would
yield less  or no destratification could be determined.
  A final  consideration is potential nitrogen super-
saturation development within the reservoir.  Super-
saturation may develop either from direct gas transfer
from air bubbles or from  the warming of the water that
results with   destratification.  Warming  the water
lowers the stable saturation concentration. Thus war-
ming  can yield supersaturation even  with no addi-
tional gas transfer. Fast  and Hulquist (1982) show the
degree of  supersaturation development to be a func-
tion of reaeration device influence, density  stratifica-
tion strength, and  position of the reaeration device in
the  water column. In many  cases, nitrogen  super-
saturation development does not  pose  a problem.
Because of submergence, supersaturation levels with-
in  the  reservoirs (with   respect  to  atmospheric
pressure)  are  typically well below saturation  levels.
Likewise, high turbulence releases from the reservoirs
strip  the  excess  gas from the  release water and
alleviate the  problem. Only where  releases with no
free interface  (such as hydroturbine power releases)
are made  does the supersaturation pose a problem.
Where supersaturation is a problem, hypolimnion in-
jection of molecular oxygen may be required.
REFERENCES

American Water Works Association. 1971. Artificial destrati-
  fication in resevoirs. Committee Rep. 63:597-604.

Bernhardt, H. 1974. Ten years experience of reservoir aera-
  tion. Seventh Int. Conf. Water Pollut. Res., Paris, France.
Bohac, C.E., J.W. Boyd, E.D. Harshbarger, and A.R. Lewis.
  1983. Techniques for reaeration of hydropower releases.
  Tech. Rep. E-83-5. Prepared for U.S. Army Corps Eng.
Davis, J.M. 1980. Destratification of reservoirs—a  design ap-
  proach for perforated-pipe compressed-air systems. Water
  Serv. 84:497-504.

Dortch, M.S.  1979. Artificial destratification of  resevoirs.
  Tech. Rep. E-79-1. U.S. Army Corps Eng.
Edinger, J.R., and  E.M. Buchak. 1979. A hydrodynamic two
  dimensional reservoir model: development and test  ap-
  plication to Sutton Reservoir, Elk River, W.Va. Prepared for
  U.S. Army Corps Eng., Ohio River Div.

Fast, A.W., V.A. Dorr,  R.J. Rosen. 1975. A submerged hypo-
  limnion aerator, Water Resour. Res. 11:287-93.
Fast, A.W., M.W. Lorenzen, J.H. Glenn. 1976. Comparative
  study with costs of hypolimnetic aeration. J. Environ. Eng.
  Div. Am. Soc. Civil Eng. 102:1175-87.
Fast, A.W., and R.G. Hulquist. 1982. Supersaturation of nitro-
  gen  gas  in  reservoirs  caused  by artificial  aeration.
  Prepared for U.S. Army Corps Eng.
Ford,  D.E.,  K.W. Thornton, A.S.  Lessem,  and J.L Norton.
  1980. A water quality management model for reservoirs.
  Proc. Symp. Surface Water Impoundments. Am. Soc. Civil
  Eng. 624-33.
Garton, J.W., R.G. Strecker, and R.C. Summerfect. 1978. Per-
  formance of an axial flow pump for lake destratification.
  Pages 336-46 in W.A.  Rogers, ed. Proc. 13th Conf. S.E.
  Assoc. Fish Wildl. Agencies.
Holland, J.P. 1982. Parametric investigation localized mixing.
  Draft Tech. Rep. U.S. Army Corps  Eng.

Hydraulics  Research Station.  1978. Air bubbles  for water
  quality improvement. OD/12. Wallingford, England.

Johnson, P.L. 1980. The influence of  air flow rate on line dif-
  fuser efficiency and impoundment  impact.  Proc. Symp.
  Surface Water Impoundments. Am. Soc. Civil  Eng. 900-12.

King, D.L. 1970. Reaeration of streams and  reservoirs - analy-
  sis and bibliography. REC-OCE-70-55. U.S. Bur. Reclam.

King,  D.L.  et al. 1983. Destratification research at  Lake
  Casitas, Calif. Draft rep. U.S. Bur.  Reclam.
Lorenzen, J.W., and A.W. Fast. 1977. A guide  to aeration/
  circulation techniques for lake management. Ecol.  Res.
  Ser.   EPA-600/3-77-004.  U.S.  Environ.  Prot.  Agency,
  Washington, D.C.

Neilson,  B.J.  1972. Mechanisms of oxygen transport and
  transfer from bubbles. Dissertation. The Johns Hopkins
  Univ.

Norton, W.R., I.P. King, and G.T. Orlob. 1973. A finite element
  model for Lower Granite Reservoir. Prepared for U.S. Army
  Corps Eng. Walla Walla District.
Pastorok, R.A., M.W. Lorenzen,  and T.C. Ginn. 1981. Artificial
  aeration and oxygenation of reservoirs: a review of theory,
  techniques, and experiences. Final Rep. TC-3400. Prepared
  for U.S. Army Corps  Eng.

Pearson, H.E., J.W. Sundberg, and J.O. Beard. 1976. Facing
  consumers' expectations on taste  and odor problems.
  Paper presented at Annu. Conf. Calif. Sec.
Speece, R.E.,  F.  Rayyan, and G. Murfee.  1973. Alternative
  considerations in the oxygen of reservoir discharges and
  rivers  Pages  342-61  in R E  SIX-PCC ed Application:, of
  Commercial Oxygen to Walci and Wastcv-'iter Systems
  Univ Texas, Austin
Speece, R.E., R.H. Siddiqi, R. Auburt, and  E. DiMond. 1976.
  Reservoir discharge  oxygenation demonstration of Clark
  Hill Lake. Final Rep. Prepared for U.S. Army  Corps Eng.,
  Savannah District.
Toetz, D.W.  1979. Effects of whole lake mixing on algae, fish,
  and water quality. Tech. Comple. Rep. A-078-OKLA. Okla.
  Water Resour. Res. Inst, Okla. State Univ.
                                                    541

-------
 EFFECTS OF AERATION ON LAKE CACHUMA, CALIFORNIA
 1980-1982
 JOHN  R. BOEHMKE
 U.S.  Bureau of Reclamation
 Denver,  Colorado


            ABSTRACT

            Lake Cachuma, Calif., has historically experienced severe hypo-limnetic oxygen depletion during
            summer stratification. To alleviate this prob em and improve the water quality of the reservoir, a
            diffused-air aeration system was installed in May 1981. A limnological study was conducted
            from April 1980 through November 1982. Results showed a weakening of the stratification, in-
            creased oxygen and temperature in the hypolimnion, increased green algae, and increases in the
            populations of Bosmina and Chironomids. Lake turnover in 1981  and 1982 occurred approx-
            imately one month earlier than in 1980. Water quality problems of manganese and  hydrogen
            sulfide were effectively controlled.
   The limnological investigation of Lake Cachuma,
 Calif., was initiated to study the effects of reaeration
 upon a reservoir whose hypolimnion becomes anoxic
 during summer stratification. During summer months,
 the   reservoir's  hypolimnion  normally  becomes
 depleted of oxygen, causing heavy metals and sulfur
 compounds in the reservoir sediments to go into solu-
 tion. This deterioration of the water quality makes the
 water undesirable for domestic water users of  the
 Santa Ynez Water District for several months of the
 year.
   One method of eliminating or lessening this prob-
 lem is to aerate  the lake during periods of stratifica-
 tion, generally April through October. As this method
 is far less costly than installing a water treatment
 plant, the Santa Ynez Water District agreed to install a
 diffused  air aeration unit with the design assistance
 of the Bureau of Reclamation (USER). The unit the.t
 was designed consists  of a 40-horsepower electric
 compressor that forces air through  small  (approx-
 imately 1 mm) holes. These holes are positioned on 0.6
 meter centers along both sides of four 30 meter PVC
 (polyvinyl chloride) pipes. This manifold is suspended
 2 to 3 meters above the lake's bottom adjacent to the
 dam.  Installation was completed in  May 1981; it was
 operated from then until October 1981 and from April
 through October  1982.  During  this time, the Water
 District experienced improved water quality.
   Study Area. Lake Cachuma is part  of the USBR's
 Cachuma Project located on the Santa Ynez River ap-
 proximately 30 kilometers northwest  of  Santa Bar-
 bara,  Calif. The  project supplies water to the area
 around the city of Santa Barbara via the Tecolote Tun-
 nel, in addition to the Santa Ynez Valley. Two other
 reservoirs, Gibraltar and Jameson, are also located on
the Santa Ynez River upstream from Lake Cachuma.
   Lake Cachuma is formed by Bradbury Dam, a zoned
earthfill structure 1,021  meters in length constructed
 in  1953. The lake is 230 meters above sea level, has
surface area of 1,255 hectares, a maximum depth of 45
meters, a total capacity of 2.53 x 108 m3i anc| a
68-kilometer shoreline.
  The region surrounding  Lake  Cachuma   is
dominated by mountains rising  1,000 to 2,000 meters
above the lake. They are made up mostly of shales;,
sandstones, and  siltstones. The vegetation is varied,
 consisting of a live oak (Quercus agrr//o//a)-dominated
 community in the  canyons and  lake basin,  and
 grasses and  chapparral  on the steeper,  exposed
 slopes. The climate is mild, with wet, cool winters and
 hot, dry summers.
   Methods and Materials. The study consisted of two
 separate phases. The first, from April  1980 through
 April 1981, obtained baseline data on the lake with no
 aerator in operation.  Phase two,  from  May 1981
 through November 1982, obtained postaerator opera-
 tion data.
   Three locations were sampled each month (Fig. 1).
 Station 1 was in the deepest part of the lake adjacent
 to the dam and aerator. Station  2 and 3 were located
 to monitor the up-reservoir effects of the aerator. This
 was important, because the aerator was designed to
 concentrate its effect within the lower basin.
   Physical-chemical  profiles were made  using  a
 multiparameter probe. Readings  were taken at the sur-
 face and then at odd meters from 1 to the bottom.
 From the oxidation-reduction potential (ORP) values,
 E7 values  were derived by adding 20 mV (to standar-
 dize to a platinum probe) and then adding 58 mV times
 the difference between the pH value and 7.0 (to stand-
 ardize to a neutral pH).
          LOS PADRES
          NATIONAL
           FOREST
Figure 1.—Lake Cachuma sampling stations.
                                               542

-------
                                                                            DESTRATIFICATION TECHNIQUES
   Water samples were taken at 1 meter, middepth (the
thermocline, if one was present), and 1 meter above
the bottom. These samples were analyzed for metals,
nutrients,  nitrogen-phosphorus,  and  major  ions.
Plankton hauls  were collected  in duplicate with
80-micron mesh closing net from 0 to 5 meters, 5 to 10
meters, 10 to 25 meters, and 25 meters to the bottom.
Duplicate benthos samples were collected with an
Ekman dredge (231 cm2) from each station.
 RESULTS AND DISCUSSION

 Temperature

 Aeration  changed  the  thermal  regime  of  Lake
 Cachuma by weakening the stratification and increas-
 ing the temperature of the hypolimnetic waters. Dur-
 ing the study, temperatures ranged from 24.5°C at the
 surface of Station 3 in August 1981, to a low of 11.7°C
 in the bottom waters of  Station 1 in February 1982.
 Stratification generally began in February, increasing
 through the summer to a maximum in August and then
 disappearing in  November with fall turnover. Figures
 2, 3, and 4 show temperature profiles for Station 1 for
 1980 through 1982. During maximum stratification, the
 differences  in temperature  between the  epilimnion
 and hypolimnion are 9.4°  C in 1980, 8.0° C in 1981, and
 5.1° C in 1982.
   This decreasing  temperature difference between
 top and bottom  layers during maximum stratification
 is caused mainly by a warming of the hypolimnion.
 The decrease from 1981 to 1982, both years of aerator
 operation, may be attributable  to the earlier startup
 date in 1982 of April, compared to May for 1981. Maxi-
 mum temperature variations in the hypolimnion are
           o JAN
           • FEB
           o MAR
          " • APR
           A MAY
           » JUNE
           x JULY
          4 0 AUG
             SEPT
           X OCT
           I NOV

        __f*DEC
                       12      16
                    TEMPERATURE CO

           TEMPERATURE  PROFILES AT C-l FOR  I98I
                 LAKE CACHUMA, CALIFORNIA

 Figure 3.—Temperature profiles  at  C-1  for 1981 Lake
 Cachuma, Calif.
                       12      16
                    TEMPERATURE CO

           TEMPERATURE  PROFILES AT C-l FOR I960
                 LAKE  CACHUMA, CALIFORNIA

Figure  2.—Temperature profiles  at  C-1  for  1980  Lake
Cachuma, Calif.
                   I0  12  14   16   IS
                   TEMPERATURE-CO
          TEMPERATURE  PROFILES AT C-|  FOR I982
                LAKE CACHUMA, CALIFORNIA
Figure  4.—Temperature  profiles at  C-1  for  1982  Lake
Cachuma, Calif.
                                                 543

-------
 LAKE AND RESERVOIR MANAGEMENT
 2.0° C in 1980, 5.0° C in 1981, and 5.5° C in 1982, while
 the maximum variations  in the epilimnion for 1980,
 1981, and 1982 are 10.3° C, 10.5° C, and 10.5° C, respec-
 tively. This  increase in maximum temperature varia-
 tion in the hypolimnion during years of aerator opera-
 tion, while epilimnetic variation remained essentially
 constant during periods of both operation and non-
 operation, shows the hypolimnetic heating caused by
 aeration.

 Dissolved Oxygen

 The primary purpose of the aerator was to oxygenale
 the hypolimnion to improve water quality for domest c
 use. Without aeration, low dissolved oxygen concen-
 tration at the water-sediment interface  results in the
 formation of a reducing environment in  which metaIs
 and compounds are found in solution, causing water
 quality problems.
   Figures 5, 6, and 7 show the dissolved oxygen pro-
 files at Station 1  for 1980,1981, and 1982, respectively.
 In 1980, the bottom of the lake became anoxic in the
 month of July. By September 1980, the entire hypolim-
 nion,  or approximately  one  half  the  volume,  had
 become anoxic. During the same time, the epilimnetic
 waters exceeded 20°C, making survival  of the lakes
 yearly stocked rainbow trout population very difficult.
 During the same period, the epilimnion  dissolved ox-
 ygen ranged from 6 to 9 mg/l.
   Concentrations of dissolved oxygen in May 1981 at
 beginning of the stratification season were 1 to 2 mg/l
 lower than in 1980. Without aeration, this would have
 meant poor water quality sooner than in  1980.  How-
 ever, by July 1981, the aerator, which had begun opera-
 tion on  May  10,  had subsequently slowed the deox-
 ygenation process. From May to July 1980, nearly 4
                                     o JAN
                                     • FEB
                                     n MAR
                                     • APR
                                     A MAY
                                     * JUNE
                                     x JULY
                                     0 AUG
                                       SEPT
                                     X OCT
                                     I NOV
                                     •k DEC
                        6      8      10
                  DISSOLVED OXYGEN (mj/L)

         DISSOLVED OXYGEN PROFILES AT C-l FOR I98I
                 LAKE  CACHUMA, CALIFORNIA

 Figure 6.—Dissolved oxygen profiles at C-1 for 1981  Lake
 Cachuma, Calif.
                  DISSOLVED OXYGEN (mg/L)

         DISSOLVED  OXYGEN  PROFILES AT C-l FOR I980
                 LAKE CACHUMA, CALIFORNIA

Figure 5.—Dissolved oxygen profiles at C-1 for 1980  Lake
Cachuma, Calif.
    0     2      4      6      8      10
                 DISSOLVED OXYGEN (mg/L)

         DISSOLVED OXYGEN PROFILES AT C-l FOR I982
                 LAKE CACHUMA,  CALIFORNIA


Figure 7.—Dissolved oxygen profiles at C-1 for 1982 Lake
Cachuma, Calif.
                                                  544

-------
                                                                              DESTRATIFICATION TECHNIQUES
mg/l of dissolved oxygen were lost in the hypolimnion,
while with aeration in 1981, this loss was only 2 mg/l.
With  aeration in 1981,  dissolved oxygen concentra-
tions averaged approximately 1.5 mg/l in the hypolim-
nion during September.  This increase in dissolved ox-
ygen may or may not help fish survival because of the
concurrent heating of the hypolimnion.  Turnover in
1981 occurred in November,  a  month earlier than in
1980.
   In 1982, the aerator was in operation  in April, one
month earlier than in 1981. The rationale for doing this
was to begin aeration before the stratification became
too strong and to increase the total dissolved oxygen
throughout  the  summer.  Comparing 1981 and 1982,
the  July, August, and  September  dissolved  oxygen
concentrations  were less in 1982, even  though the
June concentration was  higher.  The reason for this
decrease in  DO, despite  the earlier startup date,  is
unknown, but may be  related  to increased hypolim-
nion biological oxygen  demand (BOD). More research
should  be conducted to determine the time at which
aeration should be started  to  optimize the DO con-
centrations  later  in the  stratification season. Con-
centrations  approaching zero were noted at the bot-
tom in  August and September. Turnover in 1982 was
also in  November.
   To compare aeration effects on the DO concentra-
tion throughout the lake, Figures 8,9, and 10 illustrate
the change  in DO at 21 meters for all three stations.
Figure  8 shows a 20  percent improvement  in DO
saturation at Station 1 in September with aeration. At
              i I mill	I, m, I, u, i km.TV...., I,,/I In,,, I i, ,,, I
              !5 5 IS 25 5 IS 25 5 15 25 5 15 25 5 15 25 5 IS 25 5 15 25 5 15 25 5 15 25
                APR  MAY  JUNE  JULY  AUG SEPT  OCT  NOV  DEC
             PERCENT SATURATION AT 2IM- STATION 1
                  LAKE CACHUMA. CALIFORNIA

Figure  8.—Percent  saturation  at  21m—Station  1  Lake
Cachuma, Cali.
   ' 5 15 Z5 5 J5 25 5 15 25 5 15 25 5 15 25 5 15 35 5 15 25 5 15 25 5 15 25 5 15 25 5 15 ?5 5 15 25
     JAM   FtB  MAR  APR  WAY  JUNE JULY  AUG  SEPT  OCT   NOV  DEC

             PERCENT SATURATION AT 21 M-STATION 2
                   LAKE CACHUMA, CALIFORNIA

Figure 9.—Percent  saturation at  21m—Station 2  Lake

Cachuma, Calif.
                                                     Station 2 (Fig. 9), the improvement decreased to 8-15
                                                     percent. As one moves farther away from the aerator,
                                                     and gets to Station 3 (Fig. 10), all 3 years are identical
                                                     in August and September, showing no improvement in
                                                     DO from aeration at 21 meters. All  three stations are
                                                     fairly equal through the month  of June.

                                                     Major Ions

                                                     Table 1  presents average values for the major ions,
                                                     pH, and conductivity for the period  of thermal stratifi-
                                                     cation  (April through November) for both  1980 and
                                                     1981. These data show the lake to  be high in sulfate,
                                                     bicarbonate, calcium, and total dissolved  solids,  ex-
                                                     ceeding the limits for some domestic water quality
                                                     standards. No significant changes in ionic concentra-
                                                     tions were noted between the period of aeration and
                                                     no aeration. Major ion data were taken only in 1982 in
                                                     November.  Comparing these  data with  November
                                                     1981, the only significant change is a  50 percent in-
                                                     crease in chloride level from 11.3  mg/l to  16.1 mg/l.
                                                     This is still a low level.

                                                     Oxidation-Reduction Potential

                                                     From  the  oxidation-reduction  potential readings, a
                                                     value of E7 is calculated. This is a measure of the abili-
                                                     ty of certain elements or compounds to go into solu-
                                                     tion. The lower the E7, the more likely a certain heavy
                                                     metal, etc., will be found in solution.
                                                        Figure 11 compares bottom E7 values  for  1980,
                                                     1981,  and  1982.  The ranges on the left side  of the
                                                     figure show the approximate values where ammonia,
                                                     manganese, iron, and sulfides are  found in solution.

                                                         Table 1.—Averages of water chemistry parameters
                                                            comparing the period of April-November for
                                                            both 1980 (no aeration) and 1981 (aeration),
                                                                   Lake Cachuma, California.
                                                           Parameter
                                                                            1980 Average
1981 Average
Calcium (mg/l)
Magnesium (mg/l)
Sodium (mg/l)
Potassium (mg/l)
Bicarbonate (mg/l)
Sulfate (mg/l)
Chloride (mg/l)
TDS/105°C (mg/l)
PH
Conductivity (S/cm)
83
38
36
3.2
197
242
13.1
550
7.9
754
77
40
41
3.1
191
257
13.6
581
7.9
799
                                                                  I960

                                                                  I98I

                                                                  1962:
                                                                                                I
                                                                              5 25 5 15 25 5 15 25 5 15" 25 5 15 25 5 15 25 5 15 25
                                                                              JUNE JULY  AUG SEPT OCT  NOV  DEC
                                                                   PERCENT SATURATION AT 2|M- STATION 3
                                                                         LAKE CACHUMA, CALIFORNIA
                                                      Figure 10.—Percent saturation  at  21m—Station 3 Lake
                                                      Cachuma, Calif.
                                                  545

-------
 LAKE AND RESERVOIR MANAGEMENT
 Values from 1980 decline rapidly from nearly 600 rrV
 in April down to a -25 mV in August. Data in Figure "I
 indicate that during this decline one would  expect
 manganese and iron in solution in July and August;
 sulfides (H2S) should  be found  in solution  during
 September  and October, and iron  and manganese
 should be found just  before turnover in  November.
 This scenario was confirmed by the Santa Ynez Water
 District, which experienced  problems  with these
 metals and  H2S in its water, which comes from an in-
 take at the lake bottom.
   Data collected during 1981 indicate that, although
 the  initial  E7 values were lower than in 1980,  tne
 overall rate of decline was less. The low E7 value in Oc-
 tober may have been caused by some large releases
 of water from the dam's  bottom outlet structure into
 the  river to  recharge downstream water tables.  This
 brought poorer quality up-reservoir water into the area
 adjacent to  the aerator. Turnover in late October cor-
 rected this situation.
   Values of E7 in 1982 were similar to those of 1981,
 declining slowly from  January through August and
 then dropping  sharply in September nearly into the
 sulfide range.  No samples  were taken in October,
 where 1981 shows its low value.
   The effects of turnover can be seen on  Figure 11.
 The November E7 values for 1981 and 1982 were bolh
 high, increasing with the  mixing of turnover after low
 values  during stratification. The November 1980  EE7
 value was still low, not rising until the turnover took ef-
 fect in December.

 Metals

 Six  metals, cadmium, copper, iron, manganese, lead,
 and zinc, were investigated in the Lake Cachuma
 water samples. Of the six, cadmium, copper, lead, arid
 zinc were found in concentrations low enough not :o
 be considered water quality problems. The maximum
  CO
     -200
             JFMAMJJASOND
                         MONTH
       AVERAGE   BOTTOM  "E7"  VALUES

           LAKE  CACHUMA,  CALIFORNIA

Figure 11 .—Average bottom E?values Lake Cachuma, Calif.
values for these four metals are found in Table 2. Ex-
cept for lead, all concentrations have decreased dur-
ing the study.
   The manganese concentrations found in the hypolim-
nion  confirm the E7 trends.  In 1980, the  maximum
average concentration of 605 jug/l occurred  in August.
This average dropped to 40 ^g/l in December following
turnover. The E7 values during these times were 150
mV in August and 500 mV in December. The maximum
during 1981 was 393 ngl\. This occurred in October, the
same time that the E7 had dropped dramatically  from
400 to 180 mV.
   Iron concentrations show some interesting trends.
The high values associated with November-February
1981  and November-May 1982 show the input of par-
ticulate iron that accompanies the runoff season. The
long  period  of zero iron in  the summer of 1980 was
caused by the extremely low E7  values  that went
below the iron range into the  sulfide range. This tied
up the iron in the insoluble form of iron sulfide. The
summers of 1981 and 1982 show low iron values that
correspond to E7 values in the iron range.

Nutrients

Nutrient concentrations data reflect the aerobic  con-
ditions during 1981 and 1982. Table 3 lists the  nutrient
averages for Lake Cachuma;  1979 Twin Lakes,  Col-
orado, nutrient  averages are given as a comparison
with an oligotrophic lake. The reduced percentage of
ammonia in the total nitrogen of 5.3 percent in  1981
(aeration) versus 31.9 percent  in 1980 (no aeration) is
indicative of the absence of the anaerobic  portion of
the reaction transforming organic nitrogen into am-
monia (no total nitrogen data were taken in  1982). The
subsequent oxidation of ammonia into nitrite and then
nitrate also indicates a more aerobic environment. Be-
cause this oxidation reaction  is aerobic in nature, a
larger percentage  of available ammonia  was  con-
verted to the stable nitrate during aeration in 1981 and
1982  (85  versus 53 percent in 1980).  Lower ortho-
phosphate concentrations in 1981 and 1982,  and lower
percentage of  total  phosphorus during aeration in
1981 and 1982 (55 and 63 versus 73 percent in 1980), in-
dicate the lack of phosphorus release from the sedi-
ments that occur under anaerobic conditions. Effects
of runoff, inflow, and outflow were negligible during all
3 years. Lower  nutrient  concentrations in  1981  and
1982 indicate a trend toward decreased trophic status
of the lake resulting from aeration.
                                                         Table 2.—Maximum heavy metal concentrations,
                                                                  Lake Cachuma, 1980-1982.
Metal
Copper
Zinc
Lead
Cadmium
Maximum Concentration (^g/l)
6.7
20.0
1.8
0.44
Date
June 1980
November 1981
July 1981
September 1980
  Table 3.—Lake Cachuma nutrients (/xj/l) lake averages.

                    Total Ortho NH3  NO2 N03  TKN
                     P    P
4/80-12/80 (84 values)
1/81-11/81 (98 values)
1/82-11/82 (60 values)
31.05 22.54 60.42 4.65 31.85 202.5

26.34 14.80 17.040.80 14.46282.40
14.6   9.2  16.6  1.1  25.3   —
Twin Lakes, Colo., 1979  1.8   —   20.0
                  46   106
                                                546

-------
                                                                            DESTRATIFICATION TECHNIQUES
Benthos

Tables 4 and 5 summarize the effect of the aerator on
the benthic populations  at Lake Cachuma. Table  4
shows the trend in population toward chironomlds, and
away  from  oligochaetes.  From  1980  to   1982
chironomids  increased   nearly 300   percent  while
oligochaetes decreased 20 percent.
From  Table 5, several  general statements can be
made in comparing the three sampling stations. Sta-
tion 3 has the greatest concentration of chironomids,
and it  has increased with aeration. Station 2 has, by
far, the greatest concentration of oligochaetes, and  it
has decreased with aeration. The  biomass has cor-
respondingly declined at Station 2 while increasing at
Station 3. At Station  1,  both chironomids and oligo-
chaetes have increased, although  not  by enough to
have the greatest concentration of  either organism.

Phytoplankton

Phytoplankton  populations  at Lake  Cachuma  were
found  to peak in the spring and early  summer. This
pattern existed both with and without aeration. In 1980
(Fig. 12), when the study began, April and May showed
a large bloom of the blue-green algae Aphanizominon
of  approximately  10,000  cells/l.  As  the  water
temperatures increased, Aphanizominon disappeared
and was replaced by the green algae Mougeotia. The
maximum  level  for  Mougeotia was  3,500  cells/l in
June.  The remainder of  1980 saw  low, near  zero,
values for phytoplankton.
   In the spring of 1981 (Fig. 12), before the aerator was
in operation, Mougeotia peaked in  March to 7,800
    Table 4.—Average yearly concentrations of benthic
  organisms/m2 (April-November averages in parenthesis).
                Chironomids
                    Oligochaetes
 1980
 1981
 1982
100(110)
380 (267)
497 (396)
        887 (846)
        886 (673)
        572 (683)
             cells/l. In  April, the phytoplankton dropped  to  near
             zero before Aphanizominon peaked in May to 7,000/1.
             Also in May, a diatom, Asterionella, showed a small
             peak of 1,200/1. As the May data were collected just 1
             week after the aerator was put into operation,  it is
             doubtful that they had any effect on this May peak.
             The  remainder of 1981, as in 1980, showed very low
             phytoplankton levels. A  small  peak of the diatom
             Synedra occurred in November, which corresponds to
             lake turnover.
               As is illustrated in Figure 12, more changes occur-
             red in 1982 in the phytoplankton populations than in
             1981. As in 1981, the green algae  Mougeotia peaked
             first  in April to a value of 9,800 cells/l. This is 1 month
             later than  in 1981. The Aphanizominon peak occurred
             in May, as in 1981, rising to a value of 9,800 cells/I. The
             total phytoplankton  (Mougeotia  and Aphanizominon)
             at this time reached nearly 17,000 cells/l, the highest
             value of the study. In July, instead of the usual decline
             to near zero level, Mougeotia showed a second peak
             of 7,350 cells/l. The substantial green algae population
             from April to July in 1982 is showing a change in the
             phytoplankton populations from the lake's aeration.

             Zooplankton

             Figure 13 depicts the changes in zooplankton popula-
             tions during the study. In 1980 the copepods in April
             were at a  concentration of 5.0 no/I, increased to 6.2
             no/I  in June, declined to  zero in September and Oc-
             tober, and then increased to 1.8 no/I in January 1981.
             During the same time, Daphnia concentrations in April
             1980 were 12.0 no/I, declined in May and then rose to a
             second peak of 5.8 no/I in June, and then fell to  near
             zero values through January 1981.
               Bosmina in 1980 didn't appear until June, increased
             to a peak of 10.4 no/I in July, decreased to near zero in
             September 1980, and then rose to 12.0 no/I in January
             1981.
               Copepod populations in 1981 fluctuated between 2
             and  4 no/I during the spring, peaking to 6.9 no/I in
             June. Populations fell  to  near  zero  from  August
             through November before a second smaller  peak of
             3.9 no/I appeared in December. The Daphnia popula-
         Table 5.—Yearly average benthic population numbers and biomass for Lake Cachuma, Stations 1 to 3.
       Year
Chironomids (no/m2)
Oligochaetes (no/m2)
Biomass (g/m2)
                            Station 1
                        80
                               81
       31
      306
199
357
                                                           Station 2
                                       82
                                                      80
                                           81
                                                                     82
384

401
  17

2208
 335
2024
 342
1194
    0.0683  0.4355   0.2072
                                  1.3449   1.2186   0.7014
                                                                     Station 3
                                                                80
                                                          81
251

148
604

277
                                                          82
765
120
                                                 0.1629  0.3822   0.4542
Figure 12.—Phytoplankton, 1980-82, Lake Cachuma, Calif.
                                                 547

-------
LAKE AND RESERVOIR MANAGEMENT
Figure 13.—Zooplankton, 1980-82, Lake Cachuma, Calif.

tion again had two peaks, 7.5 no/I in March and 2.2 no/I
in May, before dropping to zero values for the re-
mainder of the year. Bosmina fell from the January
1981 peak of 12.0 no/I to 1.8 no/I in February. They re-
mained at this low level until May, when a peak of 10.0
no/I appeared. From this peak, Bosmina decreased to
near zero values from August until the end of the yesir.
  In 1982, copepods ranged between 2 and 4 no/I un:il
August, when the concentration fell to near zero. It re-
mainded  near zero through September and then rose
to 5.9 no/I in  November. Daphnia populations remain-
ed near zero throughout 1982, while Bosmina peaked
in April-May  at 11.9 no/I before also dropping to near
zero values in August.
  Throughout the study,  four major changes can be
seen for the zooplankton of Lake Cachuma. First, the
copepod population remained similar in trends with or
without aeration. Second, the  Bosmina population
tended to increase with aeration and peaked just once
in 1982 instead of the double peak in 1981. Third, the
Daphnia population declined each year until, in 1982,
the numbers decreased to near zero throughout the
year.  Fourth  is the trend of  all the  zooplankton  to
decrease to  very low  numbers during the summer
months, regardless of aeration.
CONCLUSIONS

Lake Cachuma historically experienced anoxic watsr
in the summer, which degraded the quality of water for
domestic  and  irrigation  purposes.  To eliminate  or
alleviate the problem, the Santa Ynez Water District
installed an aerator to mix and oxygenate the hypolirn-
nion. The Bureau of Reclamation's study to determine
the effects of the aerator operation began in 19HO
(preoperation) and continued through 1981 and 19B2
(operation). The study found that the hypolimnion in-
creased in dissolved oxygen and temperature and
decreased in heavy  metals and available nutrients.
These and other effects are summarized in Table 6.
Further  studies would be needed to  determine the
aerator's effect on all the parameters. Some of these
 effects would have to be based on the operation of the
 reservoir  and the needs of the water users.  Future
 studies should include fish population information.
   The Santa Ynez Water District has been very pleas-
 ed with the operation of the aerator and its resulting
 improvement in the quality of water delivered to its
 constituents. Proposed future plans for the reservoir
 include a  multiple-level outlet works to further improve
 the delivered water.
   Table 6.—Summary of effects of aeration on the water
   quality and biology of Lake Cachuma, Calif., 1980-1982.
 Parameter
With Aeration    Without Aeration
Dissolved oxygen
Temperature
ORP
Iron
Manganese
PH
Conductivity
Major ions
Nitrate
Ammonia
Orthophosphate
Blue-green algae
Green algae
Diatoms
Copepods
Daphnia
Bosmina
Chironomids
Oligochaetes
+
*
+
+
+
0
0
0
*
*
*
0
*
0
0
*
*
*
*
—
*
_
_
—
0
0
0
*
*
*
0
*
0
0
*
*
*
*
Key:
+ positive effect
- negative effect
0 no effect
* effect not determined
                                                548

-------
REVIEW OF DESIGN GUIDANCE ON HYDRAULIC
DESTRATIFICATION
J. P. HOLLAND
U.S. Army Corps of Engineers
Waterways Experiment Station
Vicksburg, Mississippi
            ABSTRACT

            Two general methods exist for artificially destratifying an impoundment: (a) pneumatic destrati-
            fication using compressed air jets, and (b) hydraulic destratification using water jets. Results
            from laboratory parametric investigations of hydraulic destratification are presented for a varie-
            ty of test conditions. A practicable limit of destratification, the 80 percent mixed state,  was iden-
            tified and regression analysis used to define the time required for development of this 80 percent
            mixed state in terms of reservoir geometry, stability,  and pumping rates. Results of the regres-
            sion analysis showed the dimensionless mixing time to the 80 percent mixed  state to be a func-
            tion of a "destratification" densimetric Froude number. The use of these results in a design pro-
            cedure is discussed.
INTRODUCTION

When density stratification  develops  in a reservoir,
vertical  mixing between the  epilimnion and hypolim-
nion is limited or negated and little if any dissolved ox-
ygen (DO) is transferred into the hypolimnion. Subse-
quently,   biochemical   oxygen  demand  gradually
reduces  the  DO  level  in the  hypolimnion, often  to
anoxia.  Complete or partial  artificial destratification
has been suggested as  a method  for enhancing bulk
hypolimnetic DO. Artificial destratification, in either
complete or partial form, enhances hypolimnetic DO
by the continuous mixing of hypolimnetic waters with
the well-oxygenated epilimnion, thereby reducing  or
eliminating impoundment stratification.
  Two general methods  exist for artificially destratify-
ing an impoundment: (a) pneumatic destratification
using compressed air jets and  (b) hydraulic destratifi-
cation using water jets. The results of site-specific ap-
plications of pneumatic destratification were reported
by Fast and Hulquist (1982). Results of initial research
conducted at the U.S. Army  Engineer Waterways Ex-
periment Station (WES)  on hydraulic destratification
were given by Dortch (1979).  Research has continued
at WES on design guidance for hydraulic destratifica-
tion  systems; the results  are  presented  here along
with  a   review of  the  initial  research  efforts  on
hydraulic destratification.
REVIEW OF INITIAL RESULTS

Jet injection orientation  (vertical or  horizontal) was
found  to be a major design criterion for hydraulic
systems. The vertical orientations (shown in Fig. 1)
resulted in the most efficient mixing characteristics.
Mixing results also suggested that the destratification
system should be designed  for an 80 percent mixed
state rather than a 100 percent mixed state that would
require prohibitive pumping rates and/or mixing times.
The time required for an impoundment to become 80
percent mixed with  a vertical injection system was
found  to be a function of two dimensionless group-
ings: dimensionless time, t* (mixing time multiplied by
flow rate and divided by the  volume of the reservoir),
and the product of densimetric Froude numbers of the
reservoir (FDR) and the jet (Fj). Regression analysis of
results of  laboratory tests yielded the  following rela-
tionship
    t*80o/0 = 0.0031 (FjFDR)-0.54
(D
  a.  Horizontal injection in the epilimnion.
  b.  Vertical injection in the   row nuue utter* R
     hypolimnion.
                                                                      I
T7
r
q^N —

	 ^_ j
                  I

  c.  Horizontal injection in the hypolimnion.
  d.  Vertical Injection in the epilimnion.

Figure 1.—Schematic of diffuser-intake orientation.
                                                 549

-------
  LAKE AND RESERVOIR MANAGEMENT
  where the ratios of reservoir depth to length (aspect
  ratio) for these tests were in the range 0.036 to 0.05C!.


  OVERVIEW OF ADDITIONAL RESEARCH

  To investigate the applicability of Equation 1 outside
  the range  of  reservoir aspect ratios  investigated  in-
  itially, additional  laboratory tests were conducted
  with aspect ratios ranging principally from 0.036 to
  0.09; six tests with ratios of approximately 1.0 were
  also conducted. Figure 2 shows a plot of the Froude
  number product (FjFDR) as a function t*80o/0 for both
  the Dortch (1979) initial tests (which are described  by
  Equation 1) and the additional  tests.  Examination of
  the figure clearly shows that the functional relation-
  ship between  t*80% and the  Froude number product
  expressed  in Equation 1 is an  inadequate predictivs
  formulation for the later tests.
   Analysis of the Froude number product showed that
  the grouping is actually a composite variable approxi-
  mately equal to the square of  a "destratification" den-
  simetric Froude number (Fr) times the ratio of the area
  of the diffuser ports to the reservoir  cross-sectional
  area in the lateral plane. The "destratification" den-
  simetric Froude number, defined as
              V
Fr =V2gAP
                                                 (2)
 where
       V = average port exit velocity, ft/sec
       g = acceleration due to gravity, 32.18 ft/sec2
       dr = depth of the reservoir, ft
       p = reference density of water, approximately
            1.0 g/ml
      Ap = absolute density difference of epilimnion
            and hypolimnion, g/ml

 represents the  ratio  of momentum  and  buoyancy
 fluxes, both of which had been  shown  in previous in-
 vestigations to affect mixing time. The area ratio, how-
 ever, represents the effects of  site-specific reservoir
 geomorphology  and diffuser design. To produce  a
 more general formulation for the prediction of dimen-
 sionless mixing  time, regression  analysis was  again
 used to develop the following  relationship between
 the "destratification" densimetric Froude number and
 dimensionless time to the 80 percent mixed state for
 vertical injection:
                                                           t'80o/0 = 0.204 (Fr)-
                                                 (3)
 0
 F

 D
 I -a s
 M
 E
 N
 S
 I
 0 -I 2
 N
 L
 E
 S

 S-!6
                                            ,  -8
                                    188X = 8 B03KFr>
     -5       -42     -34     -26     -I 8     -I
             LOGia OF DENSIMETRIC FROUDE NO  PRODUCT


 Figure 2.—Predictions of dimensionless time to 80 percent
 mixed state as a function of the densimetric Froude product.
 D
 1-88
 S

 0 -I Z
                          - diirorait>nl«« 188X = B 284CFr>
            -84     82      88      I
              LOG 18 OF DENSIMETRIC FROUDE NUMBER
Figure 3.—Prediction of dimensionless time to the 80 per-
cent mixed state as a function of the destratification den-
simetric Froude number.
 A plot of this fit for the same data given in Figure 2 is
 shown in Figure 3. This regression fit produced a cor-
 relation coefficient of 0.83 and does a quite adequate
 job in general in reproducing the laboratory data.
   From Equation 3, the effects  of reservoir and dif-
 fuser conditions on mixing times can be  parameter-
 ized. For a given time required to produce an 80 per-
 cent mixed state in a reservoir of known stratification
 and depth, pumping requirements for a given diffuser
 can be  obtained. Conversely, mixing times can be
 established given specific diffuser and reservoir  con-
 ditions.  Thus, the ability of Equation 3 to  accurately
 predict 80 percent mixing times  is essential for  sys-
 tem design.


 COMPARISON  OF LABORATORY AND
 FIELD  RESULTS


Predictions for the time to the 80  percent mixed state
using Equation 3 were compared with values observed
for the conditions  specified in Table 1 for Boftz Lake
(Symons et  al.  1967) and Vesuvius  Lake (Symons,
1969). As shown in Table 2, the observed dimension-
less  times to the  80 percent mixed  state are quite
similar to those calculated with  Equation 3 for  the
Table 1.—Site-specific diffuser and lake details for Boltz and
 Vesuvius Lakes (from Symons et al. 1967; Symons, 1969).
                          Boltz Lake   Vesuvius Lake

Volume, acre-ft            2930        1230
Maximum depth, ft           62          30
Initial density difference        0.00282     0.00360
Pump discharge, cfs           6.5         6.4
Discharge velocity, fps         8.3         8.4
Discharge port diameter, ft      1.0         1.0
Pumping duration, days       36.0         8.7*

'Almost fully mixed at 8.7 days; this time assumed for 80% mixed state.
                                                  550

-------
                                                                              DESTRATIFICATION TECHNIQUES
 Table 2.—Comparison of predicted and observed dimen-
sionless times to the 80 percent mixed state for Boltz and
                  Vesuvius Lakes.
Bolvz
Vesuvius
t* 80%
Calculated
0.12
0.11
Observed
0.16
0.09
conditions of Table 1. The error of prediction is quite
likely traced to site-specific effects such as wind and
solar radiation which were not included in the labor-
atory tests upon which Equation 3 is based. However,
Equation 3 does give a  good first-approximation for
the  design of a  hydraulic destratification system.
Operation of this initial  design could be further ana-
lyzed with a reservoir water quality model which simu-
lates the effects of meteorology upon the impound-
ment to produce a more accurate system design.
SUMMARY

An integrated research program has been conducted
to  establish  design  guidance  on  pneumatic  and
hydraulic  destratification  systems.  The  principal
results presented herein relate research which deter-
mined the effects of  injection orientation upon  the
rate of hydraulic destratification.  From analysis of the
study results, it was concluded that vertical injection
that penetrated the thermocline was a more efficient
mechanism   for   hydraulic  destratification  than
horizontal  injection.
   Hydraulic  destratification systems should be  de-
signed to produce an approximately 80 percent mixed
state.  Design of a system to produce a thoroughly
homogeneous reservoir would require pumping rates
and/or times that are prohibitive. The 80 percent mixed
state is essentially equivalent to the fully mixed state
except at the vertical reservoir boundaries and can be
achieved with realistic pumping rates and times.
  A  dimensionless  description  of  the   hydraulic
destratification process was developed. This descrip-
tion correlates normalized mixing time to the 80 per-
cent mixed state with a "destratification" densimetric
Froude Number.  The  latter  dimensionless grouping
addressed the effects of momentum and buoyancy as
they pertain to mixing time.  This description may be
used in the preliminary stages of system design to in-
vestigate the feasibility  of  destratification. Further
operation of initial system designs may be simulated
with more sophisticated water  quality models to
assess the effects of site-specific conditions, such as
meteorology, on system performance.

ACKNOWLEDGEMENTS:  The  tests described  and  the
resulting data presented, unless otherwise noted, were ob-
tained  from research  conducted under the Environmental
and Water Quality Operational Studies of the United States
Army Corps of Engineers, Waterways Experiment Station,
Vicksburg, Miss. Permission was  granted by the Chief of
Engineers to publish this information.


REFERENCES

Dortch, M.S. 1979. Artificial destratification of reservoirs.
  Tech. Rep. E-79-1. Hydraul. Lab. U.S. Army Eng. Waterways
  Exp.  Sta., Vicksburg, Miss.
Fast, A.W., and R.G. Hulquist. 1982.  Supersaturation of
  nitrogen gas caused by  artificial aeration in reservoirs.
  Tech. Rep. E-82-9.  U.S. Army Eng. Waterways Exp. Sta.,
  Vicksburg, Miss.
Symons, J.M. 1969. Water Quality  Behavior in Reservoirs,  A
  Compilation of Published Research Papers.  Pub. Health
  Serv. Publ. No. 1930, 256-8, 266-8, 316. Cincinnati, Ohio.
Symons,  J.M.,  et  al.  1967.  Impoundment destratification
  for raw water quality control using either mechanical or
  diffused air-pumping.  J. Am. Water Works Ass.  59 (10):
  1268-91.
                                                   551

-------
  ENHANCEMENT OF RESERVOIR RELEASE QUALITY
  WITH LOCALIZED MIXING
  JEFFREY P. HOLLAND
  U.S. Army Corps of Engineers
  Waterways Experiment Station
  Vicksburg, Mississippi
             ABSTRACT

             Density stratification limits or negates vertical mixing in lakes and reservoirs with the result that
             vertical strata of decreasing water quality are formed. Releases from the lowest of these strata
             the hypolimnion, may be of generally poor quality due to oxygen deficiency resulting from the
             coupling of limited vertical mixing and hypolimnetic oxygen demand. A simple, cost-effective
             method to enhance  these releases, localized mixing, utilizes the effects  of jet mixing to
             transport high-quality epilimnetic water down to the hypolimnetic withdrawal zone and dilute the
             release. To effectively enhance downstream release quality, the localized mixing system must
             produce a jet of sufficient quantity and initial momentum so that it will both penetrate into the
             hypolimnion and adequately dilute the release. Laboratory investigations showed jet penetra-
             tion into the hypolimnion to be a linear function of the densimetric Froude number at the ther-
             moclme. Dilution was observed to be a function of effective pumping ratios. An example design
             based on these laboratory results is given.
 PROBLEM DEFINITION

 During the late spring or early summer months many
 reservoirs  become thermally  stratified. The subse-
 quent density stratification inhibits vertical mixing in
 these reservoirs, resulting in the formation  of thres
 vertical strata in the reservoir. The epilimnion, the up-
 per region, contains warm, low-density water general-
 ly  high in  dissolved oxygen (DO)  concentration be-
 cause of surface exchange and wind mixing;  it is
 usually considered high  quality. The region  of rapid
 temperature  change  just  below the  epilimnion  is
 called the thermocline or  metalimnion. The hypo-
 limnion, the lowest region of the reservoir, consists of
 cooler high-density water which, because  of stratifi-
 cation and oxygen demand, is often low or deficient in
 DO.
   Stratification may present a water quality  problem
 for downstream  releases from reservoirs  with  low-
 level  release outlets. The water released from these
 outlets, which will be either predominately  or com-
 pletely hypolimnetic, may be of generally poor qualit/
 because of relative oxygen deficiency. Further, during
 certain periods of the year, these waters may  become
 anoxic, resulting in the release of high concentrations
 of reduced iron, manganese, and hydrogen  sulfide.
   Several solutions have  been considered to  improve
 the  water  quality  of   these  releases:  Artificial
 destratification, hypolimmetic oxygenation, structural
 modification, and localized mixing are all feasible ap-
 proaches. However, artificial destratification of the
 entire  reservoir destroys either most or all  of the
 stratification  within the reservoir, and  hypolimnetic
 oxygenation and structural modifications are genera -
 ly expensive alternatives. Conversely, localized mixing
 is designed to destratify the reservoir in the  vicinity of
 the release structure and field applications  of  this
 concept for small impoundments have shown it to be
 a simple, cost-effective approach to improve the qual -
ty of  low-level releases (Garton and Peralta, 197£i;
Dortch and Wilhelms, 1978). It is in fact, the simplicity
of localized mixing which  promotes its cost effec-
tiveness.
CONCEPT

Localized mixing is illustrated in Figure 1. A down-
ward vertical jet of epilimnetic water transports high
quality water downward into the hypolimnion. This jet
is formed near the release structure in a number of
ways, ranging from the use of an axial flow propeller
in the epilimnion (Garton and Peralta, 1978) to a sur-
face pump with an epilimnetic intake and outflow as
shown in Figure 1. Regardless of the mechanism, the
jet  is designed with adequate initial momentum to
penetrate to the level of the release outlet in the hypo-
limnion. A portion of the transported epilimnetic water
will then be withdrawn from the reservoir along with a
quantity of hypolimnetic water, thus diluting the hypo-
limnetic outflow and improving the release quality.
The quantity  of epilimnetic water required for trans-
port will depend on  the quality desired  for the given
release and the qualities of both the hypolimnion and
epilimnion. Obviously,  the poorer the  hypolimnetic
quality,  the  more  epilimnetic  water  required  to
enhance, or dilute, the hypolimnetic release.
  While a number of important parameters have been
identified for localized mixing (Busnaina et al. 1981), it
is imperative that two general conditions be met. First,
the epilimnetic jet must penetrate to the level of  the
release outlet. If not, the release quality will be less
improved. Further, penetration of the jet beyond  the
outlet represents a waste of energy which reduces the
cost-effectiveness of the  method. In certain cases,
such as for  bottom outlets, over-penetration may
disturb bottom sediments and degrade rather than im-
prove release quality (Garton, pers. comm.) Second,
                                                 552

-------
                                                                            DESTRATIFICATION TECHNIQUES
the volume of epilimnetic water jetted into the hypo-
limnetic withdrawal zone must be sufficiently large so
that the flow-weighted  average of epilimnetic and
hypolimnetic qualities is equal to the desired quality
of the total release. This process of augmenting the
quality of the release with an epilimnetic component
is referred to as dilution of the release.
   An effective localized mixing device, then, must pro-
vide adequate initial momentum and volume flux to
transport an adequate epilimnetic volume to a pre-
scribed hypolimnetic elevation.  Holland (1983), using
laboratory test  results, quantified penetration of the
epilimnetic  jet  as measured from the thermocline
down  into  the  hypolimnion.  Moon  et  al.  (1979)
developed  initial  guidance on the required  ratio of
pumping rate to release  required to  achieve a given
dilution factor. These research results may be used to
obtain the  initial design of a localized  mixing device
for a given set of reservoir conditions. A brief  overview
of these results will be presented in this article. For a
thorough review of the assumptions inherent in these
works the references cited should be consulted.

JET PENETRATION AND DILUTION

The quantification of jet penetration depth was deter-
mined from  laboratory experiments that idealized
reservoir stratification as two-layer. The upper layer
was considered to be low-density (A) epilimnetic water
which resided over a high-density (A + ZA) hypolimnion.
Epilimnetic water was assumed to initially comprise
the downward vertical jet. Within the epilimnion, the
jet was  characterized as a nonbuoyant jet  and the
classical analysis of Albertson et al.  (1950) was used
to quantify the densimetric Froude number of the jet
at the thermocline. Subsequently, using dimensional
arguments and the results of 100 laboratory tests, the
penetration of the jet into the hypolimnion was related
to the densimetric Froude number at the thermocline
by the expression
= 1.66Fr - 0.66
                                               (1)
     DT
 where
     ZH =  depth of penetration into the hypolimnion
           as measured from the -top of the thermo-
           cline, ft
     DT =  diameter of the jet at the thermocline, ft
     Fr = densimetric Froude number of the jet at the
          thermocline
                                             Moon et al. (1979) have given guidance on the prac-
                                           ticable dilution of a hypolimnetic release by an epilim-
                                           netic jet. Based upon a comparison of the ratio of the
                                           volume flux of the epilimnetic water released  to the
                                           total release  volume flux  (DF)  and the ratio of the
                                           epilimnetic flux pumped to the total release flux (Q*),
                                           Moon et al. (1979) showed that a maximum dilution of
                                           80  percent  was practicable  (Fig. 2).  Further, for
                                           epilimnetic pumping rates greater than one half the
                                           release rate,  no increase  in  dilution was observed.
                                           While the range of data and test conditions is limited
                                           in Moon et al. (1979), their results combined with the
                                           penetration results discussed above are sufficient for
                                           an initial design.

                                           DESIGN OF A LOCALIZED MIXING SYSTEM

                                           For a given reservoir stratification, release rate, re-
                                           quired downstream  release quality, and hypolimnetic
                                           and epilimnetic qualities,  a localized mixing system
                                           may be designed. This design is incorporated  in the
                                           following general procedure:
                                             1. Specify  the downstream release  water quality
                                           objective.
                                             2. Determine the thermal distribution in the reser-
                                           voir.
                                             3. Predict the withdrawal zone established by the
                                           known downstream release  flux (Smith and Dortch,
                                           1983).
1.0
u. 0.8
Q
§0.6
i-
d 0.4
Q
0.2

1 1 1 1 1
- ' f -gg -

-/°?
f /i
7/V
~F
/ i i i i i
                                                  0123

                                                            FLOW RATE  RATIO Q*

                                           Figure  2—Dilution DF as a function of Q*, the ratio of
                                           epilimnetic  volume flux  pumped to  total  volume  flux
                                           released.
                                                              DEFINITION OF TERMS USED IN FIGURE 1:
                                                                 , V0, D0 • VOLUME FLUX. VELOCITY, DIAMETER OF
                                                                        JET AT THE OUTLET
                                                                    DT - DIAMETER OF JET AT THERMOCLINE
                                                                    ZT - DISTANCE FROM JET OUTLET TO TOP OF
                                                                        THERMOCLINE
                                                                    ZH • PENETRATION DEPTH BELOW TOP OF
                                                                        THERMOCLINE
                                                                     f - EPILIMNETIC DENSITY
                                                                  f * Ap - HYPOLIMNETIC DENSITY
                                                                    Ap - ABSOLUTE DIFFERENCE BETWEEN
                                                                        EPILIMNETIC AND HYPOLIMNETIC
                                                                        DENSITIES
Figure 1 —Definition schematic of localized mixing application (the withdrawal zone from the outlet is not shown).
                                                  553

-------
 LAKE AND RESERVOIR MANAGEMENT
    4. Compute the epilimnetic volume flux which must
 be withdrawn to result in the specified quality of the
 given downstream release based on mass  balance
 criteria.
    5. Compute  a dilution  factor,  DF,  based upon
 results from  step d. If  all the  epilimnetic volume
 pumped is assumed withdrawn, DF equals the value
 from step d divided by the total release flux.
    6. Compute the initial  pumping rate from Moon et
 al. (1979).
    7. Specify depth of penetration required. This depth
 is usually to either the outlet centerline or an elevation
 representing  a  percentage of  the withdrawal zone
 thickness.
    8. Compute  the  initial  jet  characteristics from
 Equation (1) and Albertson et al. (1950) work.
    Holland (1983) used this procedure to size a localis-
 ed mixing system for the reservoir conditions shown
 in Figures 3  and 4. For these conditions,  Holland
 determined that a pump located .6 m (2 ft) below ths
 surface should pump .62 m3/sec (22 cfs) of epilimnetis
 water  into the hypolimnetic  withdrawal  zone  to
 achieve  a  6 mg/l quality for a 2.8 m3/sec (100 cfs;)
 release.
 Figure  3.—Definition  schematic for example  reservoi'
 localized mixing design where the thermocline is 20 ft below
 the water surface and the jet was designed to penetrate 38 ft
 below the thermocline.
          TEMPERATURE. °C
                                  4681)
                                 DISSOLVED OXYGEN. mg/\
Figure 4.—Detailed vertical temperature and dissolved ox-
ygen distributions for example reservoir.
 SUMMARY

 Localized mixing is often a  viable  method of en-
 hancing the  water quality of releases from low-level
 hypolimnetic outlets. A jet of good-quality epilimnetic
 water penetrates the zone of withdrawal where it is
 released with a quantity of hypolimnetic water thereby
 enhancing the quality of the total release volume. Al-
 though  enhancement  is a function of several para-
 meters, two conditions  are necessary to promote suc-
 cessful  localized mixing: (1) the epilimnetic jet must
 penetrate to  the outlet  or  well within  the withdrawal
 zone; (2) the jet  must  provide sufficient  volume  of
 epilimnetic water to effectively dilute the hypolimnetic
 release  component  and thereby  enhance the total
 release   quality.   A  design  procedure   has  been
 developed which provides guidance for both condi-
 tions. The procedure provides guidance on  a first ap-
 proximation  for  the   design  of  localized  mixing
 systems. Certain  site-specific effects (such as the ef-
 fects reservoir  geomorphology on near-field mixing
 and withdrawal) have not been quantified in this exam-
 ple   and  may  require specific  physical/numerical
 modeling to complete the design.
                                                        ACKNOWLEDGMENT: The tests described and the resulting
                                                        data presented, unless otherwise noted, were obtained from
                                                        research conducted  under the Environmental and  Water
                                                        Quality Operational Studies  of the  U.S. Army  Corps of
                                                        Engineers by the U.S. Army hngmeer Waterways Experiment
                                                        Station, Vicksburg, Miss. Permission was granted by the
                                                        Chief of Engineers to  publish this information.
 REFERENCES

 Albertson, J.L,  et al. 1950.  Diffusion of submerged jets.
  Trans. Amer. Soc. Civil Eng. 115: 639-97.

 Busnaina, A.A., et al. 1981. Prediction of local destratifica-
  tion in  lakes. J. Hydraul. Div., Am. Soc. Civil Eng. 259-72.
 Dortch, M.S., and S.C.  Wilhelms. 1978.  Enhancement  of
  releases from stratified impoundment by localized mixing,
  Okattibee  Lake, Miss. Misc. Pap.  H-78-1. Hydraul. Lab.,
  U.S. Army  Eng. Waterways Exp. Sta., Vicksburg, Miss.
 Carton, J.E. 1979. Pers. comm. Okla. State. Univ.
 Carton, J.E.,  and R.C. Peralta. 1978. Water quality enhance-
  ment by point destratification,  Gillham Lake, Ark.  Spec.
  Rep., Okla. Water Resour. Res. Inst.

 Holland,  J.P. 1983. Parametric investigation of  localized
  mixing. Draft Tech.  Rep.,  Hydraul. Lab.,  U.S. Army Eng.
  Waterways Exp. Sta., Vicksburg, Miss.

 Moon, J.L, O.K. McLaughlin, and P.M. Moretti. 1979. En-
  hancement of reservoir release water quality by localized
  mixing-hydraulic model  investigation. Final  Draft Rep.
  First Phase of Contract DACW39-78-C-0045. U.S.  Army
  Eng. Waterways Exp. Sta., Vicksburg, Miss.

Smith, D.R., and M.S. Dortch.  1983. Freudian scaling criteria
  for selective withdrawal from stratified  impoundments.
  Draft Tech. Rep. Hydraul. Lab., U.S. Army Eng. Waterways
  Exp. Sta., Vicksburg, Miss.
                                                   554

-------
                           Watershed   Management
ILLINOIS SOIL AND WATER CONSERVATION DISTRICTS
ACTION PROGRAM FOR LAKE WATERSHED IMPROVEMENT
HAROLD HENDRICKSON
WARREN FITZGERALD
ROGER  ROWE
Association of Illinois Water Conservation Districts
Springfield, Illinois


           ABSTRACT

           Illinois Soil and Water Conservation Districts (SWCD's) working through the Association of Illinois Soil
           and Water Conservation Districts (AISWCD) have developed a strategy for nonpoint source pollution
           abatement which recognizes SWCD priorities and assigns available information, technical assistance,
           and incentive resources to solving soil erosion and sediment problems in lake watersheds. Elements
           of that strategy include: (1) SWCD and AISWCD staff field reviews of lake watersheds designated
           by the Illinois Environmental Protection Agency; (2) classification of watersheds based on watershed
           resource information and SWCD priorities; (3) encouragement and assistance for SWCD's to work
           with appropriate local, as well as, state and Federal agencies to solve nonpoint source pollution pro-
           blems through  a local work plan; (4) development of cost efficient land-operator incentives such as
           the conservation tillage risk share program being used in one lake watershed; (5) encouragement
           for lake managers and monitors to work through their SWCD to put soil and water conservation prac-
           tices in the watershed; and (6) reinforcement of SWCD informational and promotional materials to
           improve lake watershed management. AISWCD has developed an innovative program based on grass
           roots support for SWCD watershed programs. Benefits of this approach will be projected for other
           areas, and this will  prove useful, particularly for persons in areas experiencing funding reductions.
There are 3,000 soil conservation districts nationwide,
generally along county lines. The exact names vary
somewhat, but districts are special-purpose units of
Government  having a  broad array  of  powers  and
responsibilities assigned by the State government.
   District programs are based on the following:
   1. Local needs are determined by locally elected
directors. In  Illinois most  directors are farmers who
know what is happening in their neighborhoods. In Il-
linois districts, the five elected  unpaid directors are
usually from different parts of the county and many
have associate directors who keep in touch with other
localities, or interest  groups. The local  needs are
reviewed annually and a work plan for soil and water
conservation personnel is developed to address them.
   2. Sound resource information is needed. One of
the  most  basic  pieces  of information  is  the
cooperative modern soil survey. The Federal share
from the  Soil Conservation Service is matched  by
State and local funds, and districts are very closely in-
volved in  raising these funds. Another piece of  re-
source  information has  recently been completed—
The National Resources Inventory.  SCS and district
personnel analyzed the state of our soil resources.
This survey will be repeated about every 5 years.
  Districts employ a variety of  methods to gather
resource information: landowner surveys, windshield
surveys, aerial photography, and satellite imagery. In
Illinois, our water information is collected by a variety
of agencies with the State Water Survey and the State
Environmental Protection Agency playing key roles.
  3. Education and information. Districts rely on the
Extension  Service  as their educational arm. Many
districts also have  education programs geared  to
                                             555

-------
  LAKE AND RESERVOIR MANAGEMENT
  school or civil groups. Virtually every Illinois district
  has a newsletter to keep land operators informed and
  interested. Some  of our  promotions have included
  conservation tours tailored to a variety of groups, con-
  servation tillage contests, no-fall-till  campaigns, and
  lady landowner programs.
    4. Technical service. With a landowner's goals in
  mind, a trained conservationist can help design con-
  servation practices that will reduce soil loss,  keep
  nutrients where they belong, and prevent water pollu-
  tion. Applying these practices is the key work of
  districts. Illinois districts rely very heavily on the Soil
  Conservation Service for technical help but also utilize
  the services  of district employees, foresters, exten-
  sion specialists, land improvement contractors, and,
  most importantly, trained landowners. Many conserva-
  tion practices are now designed by young  farmers
  through their enrollment in community college conse'-
  vation planning sessions.
    5. Incentives are  needed  in many cases  to en-
  courage conservation practices. Incentives vary  from
  a compliment to an award to cost sharing with land-
  owners. The cost sharing percentage depends on the
  degree of public  benefit  from a  practice.  There is
  relatively little benefit  to  a  farmer, for example, h
  stabilizing a streambank leading into lake, and a hign
  cost share level is needed. We believe there is a hign
  private benefit to conservation tillage and we are star-
 ting to  move away from cost sharing on it, except fcr
 first timers. We rely very heavily on various State pro-
 grams and the Federal Agricultural Stabilization and
 Conservation Service for the monetary incentives. Our
 Association has numerous award and recognition pro-
 grams.
   6. Securing cooperation is the most important ele-
 ment in district work.  The key element in working
 together is planning in such a way that everyone has
 an opportunity to participate. The meetings and ac-
 tivities  of a State Association of Districts is  an ex-
 cellent  method of assuring cooperation of State and
 Federal agencies.
   The National Association of Conservation Districts
 represents  districts, provides them services,  and
 helps secure the  consensus needed  to get  things
 done.  Nationally,   we  need   the  understanding of
 Western States, for example, on our water quality pro-
 grams and they want the East's understanding on ir-
 rigation programs.
THE AISWCD WATER QUALITY PROGRAM

In late 1981, we negotiated a contract with the Illinois
Environmental Protection Agency to help districts im-
plement their soil erosion standards and watershed
protection programs. The project started in February
1982. The  contract was recently  renegotiated and
signed in July 1983.
  Under this contract we have hired two water quality
coordinators, both of whom have experience in work
ing closely with conservation districts, Governmenl
programs, and particularly watershed programs. Thev
report to the president of the  association and the
association's Water Resources Committee.
  Look briefly at Illinois' water resources. Lakes anc
reservoir watersheds are emphasized in our work foi
these reasons:
  1. Owing to a  lack of adequate  ground  water  ir
many parts of Illinois, reservoirs or  lakes supply over
700 public water supplies serving around 600,000 peo
pie with drinking water.
    2. Water-based  recreation accounts  for an  enor-
  mous outflow of Illinois dollars. Minnesota, Wiscon-
  sin, and Michigan  are the principal beneficiaries. We
  would like to keep some of these dollars at home.
    3. Illinois is not blessed with many  lakes, so we
  need to emphasize maintenance of the ones we  have.
    4. Our lakes and reservoirs are efficient sediment
  traps. About two thirds of Illinois' land is devoted to
  crop production and the State ranks second in the Na-
  tion in gross cropland soil loss. We have a great deal
  of sediment available to trap; we need to keep this pro-
  ductive  soil on the land.
    The program we  have come up with has several ma-
  jor features which we believe are unique.
    First,  and most important, AISCWD provides a vari-
  ety of services to  the districts, where conservation
  work gets done. We provide SWCD's  with onsite
  watershed evaluation of lake watersheds targeted by
 the Illinois Environmental Protection Agency (IEPA)
 for potential  watershed programs. We  developed a
  unique three-part rating system based on the water-
 shed evaluation  and  how the program would fit into
 the District program.
    In the first category, a SWCD project is not needed.
 Examples  include watersheds composed entirely of
 municipal governments (for which we have no authori-
 ty  in Illinois), or  those that have no  land treatment,
 such as a  lake with an entirely forested watershed.
   In category two, the watershed needs land treat-
 ment that  can probably be handled  by  the ongoing
 district  program. Perhaps only 1,000 acres of the
 watershed needs land treatment. We believe districts
 can achieve a great deal through developing a local
 land treatment watershed project and focusing avail-
 able resources on that specific area. Some of these
 local watershed projects will involve several years and
 several other units of government. Two Illinois cities,
 for example, have bought conservation tillage equip-
 ment for the district to use with farmers in watershed
 protection  programs above their water supply reser-
 voirs.
   In the  third category, the watershed projects need
 major assistance beyond  the scope of  the normal
 SWCD program.  These land treatment  projects are
 suitable  for the P.L 566 land treatment  watersheds
 program, the Rural Clean  Water Program, the Agri-
 cultural  Conservation  Program  Special  Projects or
 possibly the 314 Clean Lakes Program. These projects
 take more time, manpower, and special funding. We
 encourage  SWCD's to  maintain or  accelerate  the
 regular program in these watersheds until major  fun-
 ding is available—which may take years. We are en-
 couraged by one SCS, P.L. 566 land treatment project,
 which  took just  2  years from  inception to imple-
 mentation—we believe this is a national record for
 similar programs. Credit goes to the municipality, the
 district, and SCS for excellent cooperation and strong
 commitments to the project.
  We have provided assistance to SWCD's with com-
 munication and information efforts. About 500,000 Il-
 linois citizens will be reached through  this effort over
 a 3-year period. Films, slide sets, displays, and award
 programs are featured and  maintained at the district
 level. AISCWD also hired an artist to do conservation
 clip art  for district  newsletters to help  make them
 more contemporary.
  The association's coordinators provide a liaison be-
tween SWCD's and  State and Federal agencies  and
act as advocates  for districts.
  Finally, the  coordinators have assisted  districts
with developing  watershed project work plans  for
local projects and we expect this activity  to expand.
                                                 556

-------
                                                                               WATERSHED MANAGEMENT
  We also assist State agencies with various pro-
grams. An example is the IEPA Volunteer Lake  Moni-
toring Program. Four districts have provided monitors
and others have  assisted with organizational steps.
This is an excellent program that relies on volunteers
and  local initiative. We will give two  examples of
watershed projects that grew out of strong local in-
itiatives  including  volunteer  monitoring.  The  Lake
Kinkaid watershed project will be the next Illinois P.L.
566 land treatment project; JacKson District personnel
have been monitoring for the past 3 years. In the Lake
Sara watershed,  an ingenious local cost  share pro-
gram based on soil saved is being developed by the Ef-
fingham  District,  and they expect to see a long-term
improvement through the monitoring.
  Finally, our water quality coordinators provide sup-
port,  assistance, and  information  related  to the
general soil erosion control program throughout the
State. Unlike most States, we have water quality goals
expressed in "T"  or tolerable soil loss as a substitute
for a water quality based goal. Achieving our districts'
and State goal  of "T" by the year 2000  will require a
great deal of work. AISWCD assists  districts, par-
ticularly in planning to meet those goals and develop-
ing a consensus on the resource needs. We have
developed liaisons  with many organizations and the
press to  help achieve this.
  We wish to offer lake managers, professionals, and
concerned individuals, two invitations  based on our
experience:
  First, endorse the concept that "the condition of a
lake is a reflection of the condition of its watershed."
Improving the watershed will usually require work with
many landowners and units of government, some of
whom may have little interest in the lake.
  Second, work  closely  with your soil  conservation
district. Start by  discussing the lake and  watershed
with the  district or SCS  staff and the district director
nearest the lake. These people work for you, represent
you, and know how to get action. Second, if it has not
already been done, work with the lake landowners to
secure some form of commitment to help solve the
problem. Before pointing fingers at farmers  in the
watershed, be certain the lake residents' "house" is in
order. Septic disposal, lawn fertilizers and pesticides,
detergents, and leaf disposal are problems that might
need to be addressed. If farmers living away from the
lake (many of whom  are fishermen, by the way) see a
concerted commitment from lake owners, they too will
likely cooperate.
  Attend district board meetings and report  on the
lake. Do not be discouraged  if you do  not get  im-
mediate attention; districts have established priorities
and work on an annual work plan. Ask the district to
give attention to the lake watershed in their  annual
work plan.
  Attend the district's annual planning meeting and
report on  progress in the watershed. Recognize that
district  directors are land managers,  that many are in
other community leadership roles, and that a pat on
the back goes much further than a kick in the shins.
As  in  many other endeavors, we find that credit for
work done is one of the few commodities which is  in-
finitely  divisible.  This  positive  approach  of working
with landowners is what districts have been doing for
close to 50 years.
BIBLIOGRAPHY

Illinois Department of Agriculture, Division of Natural Re-
  sources. 1983. Annual progress report. Springfield.
Illinois Department of Energy and Natural Resources. 1982.
  Illinois agricultural  water quality  programs—a  status
  report. Springfield.
Illinois  Environmental  Protection  Agency.  1979.  Water
  Quality Management Plan. Vol. 1-6. Springfield.
	. 1983. Water Quality Management Plan. Springfield.

Krone, James,  Jr.  1982.  Water Resources in Illinois: The
  Challenge of Abundance. Illinois Issues. Sangamon State
  Univ., Springfield.
                                                  557

-------
 WATERSHED MANAGEMENT: MODIFICATIONS  IN
 PROJECT APPROACH
 DONALD R. URBAN
 WALTER RITTALL
 Soil Conservation  Service
 U.S. Department of Agriculture
 Washington, D.C.



             ABSTRACT

             A number of pollutants acting either singly or in combination affect lakes and reservoirs. Attention
             is turning towards nonpoint sources and watershed management as offering a potential for more cost-
             effective control. Nutrient reductions below  1.0  mg/l from municipal treatment plants come with an
             extremely high reduction cost factor as a result of increased hardware, operations, and maintenance
             costs  Several nonpoint source control derronstration projects have been implemented since the
             passage of  PL 92-500. Some were directed at the unknowns surrounding the control of runoff from
             both urban  and agricultural sources. For agiiculture there were questions about the agricultural in-
             stitutional arrangement and the effectiveness of practices One restriction was the voluntary nature
             of participation, deemed the only feasible implementation method. Several projects have now been
             completed and conclusions and recommendations made. The project approach is emerging as an
             effective method of focusing resources to control identified natural resource problems Some pro-
             jects did not meet all objectives because the problem was not clearly defined and early implementa-
             tion was demanded. A strong relationship exists, suggesting that nutrient reductions may be obtain-
             ed from watershed treatment, but quantification is still lacking. Other factors influenced the implemen-
             tation and kept the focus from zeroing in or  critical sources and  areas, making project evaluation
             difficult. Modifications in project development and implementation  steps will be discussed based on
             these findings and suggestions made for more  efficient management of watershed projects.
 The results of water quality demonstration projecls,
 carried out under authority of Public Law 92-500, The
 Clean Water Act of 1972, suggest that some changes
 are needed in the traditional project approach. Natural
 resource agencies must reexamine their roles and le-
 sponsibilities  regarding planning and implementation
 of projects to improve water quality. This paper traces
 the evolution of the changes and some additional ad-
 justments  needed.  Key factors for project success,
 drawn from project evaluations, will also be discussed.
   P.L. 92-500 focused attention on the unhealthy con-
 dition of much of the Nation's water resources. Sub-
 stantial amounts of money and effort have gone into
 developing a process to  identify problems more pre-
 cisely and to provide remedial action as effectively as
 possible. The  Environmental Protection Agency and
 State  water quality agencies adopted a policy of
 treating the worst cases first. They focused attention
 and money on  large increments of pollution, large con-
 centrations  of people, and industrial sources.  Only
 after remedial  action was undertaken on these major
 problems did attention shift to small towns where pol-
 lution was not as great. Billions of dollars have  been
 spent in upgrading industrial and municipal treatment
 facilities.
  Much has been  accomplished;  however, the prob-
 lem has not been solved. For example, it was recog-
 nized by the mid-1970's that diffuse-source pollution
 needed to be  treated. More than  half the pollutants
entering the Nation's waterways come from nonpoht
sources. Attention directed to reducing this kind of
pollution has  slowly  increased.  The water quality
management plans developed have focused on redu-
cing sediment and nutrients attached to  sediments
using common soil and water conservation practices.
It has been generally agreed that if sediments are keDt
 out of water, its quality would improve. The planning
 in  the mid-1970's envisioned an  accelerated soil
 conservation program.
   Initial demonstration projects were implemented in
 watersheds identified as being critical from the stand-
 point of agricultural pollution. These first projects used
 the traditional project approach. Each agricultural
 agency assisted  individual  landowners in  much the
 same fashion that it provided services and assistance
 through its other programs. These early projects prov-
 ed  that it  is possible to address a water quality prob-
 lem through cooperative activities and  with voluntary
 programs. This was  a significant  result  as  some
 doubted if a voluntary approach could  succeed when
 the impact is offsite in a lake or reservoir miles away.
  Sound  resource  management planning considers
 soil, water, plants, and animals as part of an ecosys-
 tem. The  planning  process  must  incorporate the ef-
 fects  that modification of any  part has on  all  other
 parts  of the system.  Planning  for the protection of
 water resources cannot be done in isolation, nor can
 planning for protection of the soil. This is obvious now
 but was not reflected  in all of the first demonstration
 projects. We now recognize that there is only one
 natural resource planning process. When a  resource
 is impaired, the level  of treatment increases but the
 planning process  does not differ.
  This paper is based on a review of  reports from the
 Model Implementation Projects (MIP's): Black Creek,
 Ind. (Section 108,);  Skinner  Lake, Ind.  (Section 314);
 Lake  Erie Wastwater Management  Study, Honey
 Creek, Ohio (Section 108d); other Section 108 projects
 in the  Great Lakes; and some early comments from
the 21  experimental Rural Clean Water Program pro-
jects.
                                                 558

-------
                                                                               WATERSHED MANAGEMENT
  The reports of all these projects present similar con-
clusions and recommendations. These are of interest
to lake managers who may be called upon to devise re-
medial plans. The final  project reports indicate that a
watershed management project is a sound technique
to focus attention and resources on an identified prob-
lem. However, five general key factors for success
were identified in all  the final  project  reports.
   The first factor is adequate planning at  the initial
 stage of the project. Before any action is taken, all in-
 volved parties must agree on the definition of problem,
 the sources of the pollutant, the degree of pollutant re-
 duction needed, the  methods  for controlling the pollu-
 tant, and the division of responsibilities for  the imple-
 mentation. Many project  leaders reported that lack of
 consensus on the problem and the  methods and re-
 sponsibility for  correcting it  hindered their projects,
 although some  were  able to make adjustments to
 overcome these problems.
   Most  project  leaders reported that the  funding
 agencies encouraged  them to begin implementation
 before they had time to develop a strategy. As a result,
 the agencies just accelerated what they had been do-
 ing in their existing  programs.  In most cases the ex-
 isting programs were designed to help any landowner
 who asked for it; the relative severity of each farmer's
 resource problems and the degree to which each con-
 tributed to offsite problems  were not considered  in
 servicing requests for  assistance.
   A second but related factor is local  involvement  in
 the planning. Project leaders  agreed  that the projects
 need to be small enough  so that the agencies involved
 and the potential participants can reach a consensus
 on the critical questions. The local project leaders
 need to select a coordinator who can begin  the impor-
 tant task of  organization and detailed planning and
 determine the direction of the information and educa-
 tion program. The development of a local  focus and
 the involvement of local entities and individuals in the
 planning decisions permit the projects to concentrate
 on the specific local problem and solutions. The pro-
 ject leaders were all satisfied that they ultimately got
 the varied interests working together and that a large
 amount of work was accomplished. They said, how-
 ever, that given sufficient time to enlist local coopera-
 tion and refine their work plan, they  could have been
 more effective.
   The third key factor is closely related to the first
 two.  To be successful a water quality project  must
 control the major sources of pollution. Project leaders
 said that the lack of  adequate advance planning made
 it difficult for them to focus their  efforts on those
 areas that were delivering the greatest portion of the
 targeted  pollutant. They often were forced  to work
 with those people who voluntarily agreed to cooperate.
 The funding  agencies often hindered the process by
 encouraging the project leaders to begin implementa-
 tion.
   The fourth factor is the project length. Many project
 leaders said  that the funds and the time frame for the
 project ended before they had an opportunity to focus
 their efforts  on those areas  and sources  where the
 greatest reduction of the critical pollutant could be ob-
 tained. Most of the projects reviewed were funded for
 3 to 5 years. All project  leaders agreed that  a desig-
 nated period for implementation must be established.
 The short time allowed for the demonstrations did not
 permit  project  leaders to adjust  their strategy  to re-
 spond to new information. The project period for the
 experimental Rural Clean Water Program (RCVVP) pro-
 jects includes 5 years for development of the  plans
with the cooperators and 5 years to complete  the
installation. This resulted from some early indications
from the demonstration projects. Whether a 10-year
period  is  the  correct  length  with   good  pre-
implementation  planning  is not  known. A  definite
schedule does need to be established at the beginning
of the project to focus the effort and to avoid allowing
the implementation to drag on.
  The fifth factor is the project evaluation. Most of the
funding agencies did require an evaluation and the in-
clusion of the results in the final report. Most provided
a qualitative evaluation that dealt with agency cooper-
ation, level of  participation,  kinds of practices in-
stalled,  efficiency  in spending the money allocated,
and enthusiasm of the implementors and cooperators.
Most project leaders reported that they did not fulfull
the objectives that they had established and that the
objectives often shifted during the project. The fun-
ding  agencies   did  not  specifically  require  that
quantifiable goals  be set.  Few project leaders made
any attempt to  quantify the amount of the targeted
pollutant they controlled. Only two of the seven MIP's,
for  example, developed quantifiable goals. Three
MIP's developed operational   goals that translated
general  objectives into measurable terms. The project
leaders  said that in many cases they had little  idea
how much they had accomplished.  Many project
leaders  asked  for time extensions  and additional
funds as it became apparent that they would not meet
their objectives.
  At this point,  it would be easy to suggest that per-
haps we should forget watershed management as an
alternative for lake restoration. Treating the effects
rather than the  causes of in-lake problems may seem
an attractive  alternative in the short run. The rational
method, however,  is to attempt to reduce the cause.
This is not only sound logic, but the most realistic ap-
proach  in a period of limited financial resources. The
positive result from the problems encountered by the
demonstration projects  has been a quantum leap in
knowledge about how to develop and deliver a prob-
lem-solving project in rural areas.
  The final reports answer some questions about the
effectiveness of watershed management in reducing
targeted pollutants. In most cases complete elimina-
tion of  the problem may  not  have been necessary.
Many of the typical soil  and water conservation prac-
tices were found to be very effective in reducing the
detachment and transport of sediment and sediment-
bound pollutants. For example, Honey Creek reported
that no till reduced gross erosion by about 75 percent
and phosphorous transport by over 60 percent. In the
New  York  MIP project,  reducing  surface runoff
through barnyards reduced phosphorus loadings from
that source by 75 percent. The overriding reasons for
the lack of measurable reductions in the pollutants
reaching downstream water  bodies  have been the
short time of the projects, the relatively  small areas
treated,  the lack of precise problem  definition,  and
most importantly the lack of focus on the high delivery
sources.
  We have proved that certain  conservation practices
are very effective  in reducing  pollution in lakes  and
reservoirs. We have also shown that is is possible to
identify the relatively small areas of a watershed that
deliver the largest  part of the sediment and sediment-
bound nutrients. This has a great bearing on the cost
of the  projects and  makes watershed  treatment  a
feasible alternative for lake restoration.
  To take full advantage of the lessons learned from
the projects since P.L 92-500 was enacted, a number
                                                  559

-------
 LAKE AND RESERVOIR MANAGEMENT
of changes in how new projects are formulated must
be incorporated into our planning process. The most
important change that has evolved is in the emphasis
on pre-implementation planning. There is a gap in our
planning process between the selection of a critical
subbasin  and  the  detailed planning   done  with
individual landowners. A plan of work traditionally is
developed  after funds have  been  provided, and the
plan of work reflects how the funds will be allocated
for each work element.
   The major adjustment would be to  add a separate
and distinct planning effort that takes  place before al-
locating funds for implementation. We suggest the
term  "operations planning" stage that  produces an
operational plan. This  would  be a funded  effort,  with
the money  going to the local group  that would be
responsible for carrying out the plan. It would be their
responsibility to bring  together the people and agen-
cies who need to be involved,  refine the problem,
determine its sources, evaluate alternatives to solving
it, arrive at  a consensus on the solution, agree on who
will be  responsible for each of the tasks  involved  in
implementation, and agree on how progress will be
evaluated.  This permits tailoring the  project  to the
local situation rather than tailoring the situation to the
project  funding. Providing the local project sponsors
the opportunity to devise a cost-effective plan to solve
their problem without the pressure to begin implemen-
tation overcomes many of the problems that projects
identified. It also permits funding agencies an oppor-
tunity  to fund  those projects  for implementation
which have the best chance for reaching their goals.
  USDA's  programs are voluntary and  have tradi-
tionally been open to all landowners or operators.
Problem-solving projects break  that  tradition. It  is
possible that some potential  participants  will not be
given assistance because their lands do not deliver a
significant  amount of the targeted  pollutant.  The
operational  planning period is a  time  when the local
project  can make these determinations  and develop
the consensus on the critical areas to be treated. The
most important  result of the local determination is
that it becomes a local  project. The  funding source
loses  its  significance  when the  implementation
groups unite in solving the problem.
  The review of  the water quality demonstration pro-
jects has confirmed that sound planning is necessary.
The existing agricultural agencies can deliver a fo-
cused, problem-solving program. Such a program can
be effective when it is directed by a local organization
with clearly defined goals. A  few  minor changes by
funding agencies can be extremely effective in redu-
cing many of the concerns of watershed managers.
Sound planning with local leadership can deliver cost-
effective solutions for many lake management prob-
lems. The evolution which has and continues to take
place in how to organize and deliver a watershed man-
agement project suggests that many  of the fears re-
garding costs and effectiveness are not valid.
BIBLIOGRAPHY

International  Joint Commission, Water Quality Program
  Committee, Nonpoint Source Task Force. 1983. A General
  Survey of Governmental Programs to Plan and Manage
  Nonpoint Source Water Pollution Abatement in the United
  States Great Lakes Basin. Harbridge House, Inc.

Lake Erie Wastewater Management Study.  1982. Honey
  Creek Watershed Project. Final Prog. Eval. Rep.

Morrison, J. 1982. Environmental impact of land use on water
  quality. Executive summary, Black Creek Project.

Noble County Soil and Water Conservation District. 1982. Im-
  pact of land treatment on restoration of Skinner Lake, No-
  ble, County, Ind.

U.S. Department of Agriculture and U.S. Environmental Pro-
  tection Agency. 1983. The Model Implementation Program:
  Lessons Learned from Agricultural Water Quality Projects.
  Executive Summary. The National Water Quality Evalua-
  tion Project and Harbridge House,  Inc.
                                                 560

-------
WATERSHED MANAGEMENT: COOPERATION
AND COMPROMISE
WILLIAM  K. MORRIS
Office of Watershed Management
Charlottesville, Virginia



            ABSTRACT

            Watershed management activities in Albemarle County and the city of Charlottesville in the Piedmont
            Region of Virginia have been examined. In recent years these two localities have exhibited an unusual
            spirit of cooperation and compromise in protecting the area's water supplies. Creation of a unique
            position of Watershed Management Official, funded equally by both localities, was one of the first
            steps taken locally in recognizing that watershed management was a continuing process that required
            the coordination and integration of many diverse activities and recognizing that proper watershed
            management requires more than best management practices manuals and ordinances; it requires
            constant site investigations and surveillance of  all watershed activities Protection of existing water
            supplies is the major goal of watershed management; however, planning for additional and future
            needs has not been forgotten Charlottesville and Albemarle County have again taken that necessary
            first step in planning for the future by providing the mechanism whereby land for a supplemental water
            supply impoundment and associated buffer area protection zone will be acquired in the near future
            even though the actual impoundment will not be needed for 20 or 30 years. The aspects and the degree
            of the cooperation and compromise needed for a successful watershed management program are
            examined.
INTRODUCTION

Protection  of  an  area's  potable  water  supply
resources requires that an atmosphere of cooperation
and compromise exist between and among all of the
participants. The city of Charlottesville and the county
of Albemarle in Virginia have in recent years exhibited
the cooperation and compromise necessary to develop
and  maintain a  successful water supply protection
program for the area.
  This paper examines the efforts that have been un-
dertaken by these localities to protect the  local water
supply resources.
of 3.8  MLD  and a treatment  plant capacity of 0.95
MLD.
  The area's five water supply reservoirs and one river
intake  supply potable water to approximately 70,290
people, have a  combined drainage area  of 782.34
square kilometers, a combined source capacity of 83.6
MLD and a treatment capacity of 46.55 MLD.
  The  Rivanna Water and Sewer Authority manages
the water supply reservoirs and water filtration plants
while the city and the Albemarle County Service Au-
thority  maintains the distribution system.
BACKGROUND: SOURCES OF SUPPLY

The city of Charlottesville and the urban areas of Albe-
marle County receive their water from the Rivanna
Water and Sewer Authority. The  Rivanna Authority
supplies water directly to the city and to approximate-
ly 50 percent of the County's population  through the
Albemarle  County  Service Authority. The city of
Charlottesville and  the urban  surroundings  in the
county compromise a system which included approx-
imately  66,520 persons who depend on the combina-
tion of the Sugar Hollow, Ragged Mountain, and South
Fork Rivanna reservoirs and the North Rivanna River
intake  structure. These systems  have a combined
drainage area of 683.8  square  kilometers, a source
capacity of 72.2 million liters  per day (MLD)  and a
water filtration plant capacity of 41.8 MLD.
  Due west of Charlottesville, the Crozet area is de-
pendent upon water supplied by the Beaver Creek res-
ervoir with a drainage area of 24.44 square kilometers,
a source capacity of 7.6 MLD and a filtration capacity
of 3.8 MLD to serve a  population of approximately
3,150.
  Located South of Charlottesville, approximately 620
people in the town of Scottsville depend upon water
supplied from the Totier Creek reservoir which has a
drainage area of 74.10 kilometers, a  source capacity
THE PROBLEM

Four of the five water supply reservoirs utilized by the
Rivanna Authority are located in the South Fork Rivan-
na Watershed. This watershed area encompasses ap-
proximately one third of Albemarle County (62,000
hectares) ana contains the area's largest water supply
reservoir, the South  Fork Rivanna Reservoir, which
has a surface area of 156 hectares and a safe yield of
45.6  MLD.  This reservoir  along with  the others
available in the county are considered eutrophic and
have had both water quality and quantity problems,
with  numerous citizen complaints about fishkills and
taste and odor problems.
ACTIVITIES TO PROVIDE PROTECTION

At the request and urging of the Charlottesville City
Council, the Albemarle County Board of Supervisors
enacted a number  of  building moratoriums  in the
South Rivanna watershed that ranged from an all-out
ban on construction in the entire watershed to prohibi-
ting development within 15,000 centimeters of the res-
ervoir and tributaries and on  slopes  15 percent or
greater  within a 5 mile radius of the water intake (Mor-
ris, 1980).
                                                561

-------
 LAKE AND RESERVOIR MANAGEMENT
   In September 1975 the initial water quality manage-
 ment study of the South Fork Rivanna Reservoir and
 watershed was undertaken by the Rivanna Water arid
 Sewer Authority. This study recommended the imple-
 mentation of a comprehensive  watershed  manage-
 ment plan that included reservoir management, water
 treatment modifications, point and nonpoint source
 controls and routine watershed monitoring. The 1977
 report resulting from this initial study provided the
 basic guidelines  for the future measures that have
 been taken to provide water supply protection in the
 area. (Betz-1977)
   In recognition of the need for the potential restora-
 tion and  the need for a supplemental  water supply
 source, the Rivanna Water and Sewer Authority con-
 ducted  both a Rivanna Reservoir Restoration Project
 and an  alternative water supply source study for the
 Charlottesville/Albemarle area. The reports from these
 studies indicated that  the condition of the  reservoir
 could in fact be stabilized and possibly improved by
 reservoir aeration, agricultural grass waterways and
 residential sedimentation  ponds (Browne, 1979) but
 that there were few available sites of sufficient si.re
 left in the area that would merit consideration as an al-
 ternative or supplemental water supply source. (Camp
 Dresser & McKee, 1977)
   In September of 1977, the Albemarle County Boa-d
 of Supervisors in  response  to recommendations in
 The  1977 Water Quality Management Study of the
 South Rivanna Reservoir and Tributary Area adopted a
 Runoff Control Ordinance (Albemarle County Code).
 The purpose of this ordinance was  to protect against
 and minimize the pollution and eutrophication of the
 public drinking water supply impoundments  in the
 county resulting from land development in the water-
 shed  areas. The adoption  of this Runoff Control Or-
 dinance was the first step in implementing the recom-
 mended  watershed management plan. The Runoff
 Control  Ordinance is based on the environmental per-
 formance  standard  that  post-development  runoff
 characteristics should  not exceed  pre-developmeit
 conditions. The ordinance also requires that an upper
 limit  of  24.30  kilograms  per  hectare per year for
 suspended solids loading and .122 kilograms per hec-
 tare per year for total phosphorus loading be adhered
 to. A runoff control permit is required if the develop-
 ment  of  a parcel  results  in a total impervious lot
 coverage of more than 5 percent of the area, the
 establishment of more  than 45 square meters of im-
 pervious cover, or the  disturbance of more than  76
 cubic meters of earth.
  The main restriction of the current ordinance is th.at
 it prohibits the construction of any sewage disposal
 system on any part of which lies within 6,000 horizon-
 tal centimeters of the one hundred year flood plain of
 any  impoundment or within 3,000 horizontal cen-
 timeters  of  the  edge of  any   tributary   stream
 (Albemarle County, 1983).
  To use the data from  the initial water quality report
 and to implement it properly, the Albemarle County
 Board of Supervisors in December  of 1977 formed a
 Watershed Management Plan Committee. This com-
 mittee was made up of representatives from every
 agency and interest group involved with the water sup-
 plies of Albemarle and Charlottesville. The report gen-
 erated from this committee recommended that (1) a
 position of watershed management official be created
 to coordinate and review all watershed management
 activities; (2) the major point source discharger in the
watershed be eliminated; (3) the Virginia Department
of Highways and Transportation be requested to in-
 stall and maintain erosion and sedimentation control
 measures as specified in its Erosion  and Sedimen-
 tation  Control  Manual;  and (4)  specific  watershed
 management goals be  integrated  into agricultural,
 technical, and financial  assistance programs to em-
 phasize and give priority to problem areas and conser-
 vation measures (Browne, 1979).
   On July 28,1980, the position of watershed manage-
 ment official was  funded from the General Funds of
 the County and the City at 50 percent each. The crea-
 tion of the position of watershed management official
 was one of the first steps taken locally  in recognizing
 that watershed management is a continuing process
 that requires the coordination and integration of many
 diverse activities and in recognizing that proper water-
 shed management requires  more than  Best Manage-
 ment Practice Manuals  and ordinances;  it  requires
 constant site investigations and surveillance of all
 watershed activities.
   The  basic job responsibilities of the  watershed
 management official for the  county of Albemarle and
 the city of Charlottesville are (1) the coordination of
 ordinances regulating soil  erosion, sedimentation,
 runoff, and stormwater detention; (2) the integration of
 Federal, State,  city,  and county agency  programs
 relating  to watershed activities;  (3) participation in
 land use planning  directed towards improving water-
 shed management  programs;  (4)  development of
 educational programs  for land owners  to encourage
 best  management  practices  in  agricultural   and
 developmental   activities; (5)  review  of watershed
 management programs in other jurisdictions; (6) the
 dissemination of information; and (7) recommending
 improvements  and additions to  programs  and or-
 dinances relating to watershed management.
   In efforts to further protect the water supply water-
 shed areas of the city of Charlottesville  and the coun-
 ty of Albemarle, all the publicly owned properties in
 the  watershed  areas  were rezoned to  conservation
 district  classifications, the County's Comprehensive
 Plan was amended to delete the water  supply water-
 sheds from the urban area and a  comprehensive
 rezoning of the County was completed which included
 low density zoning  in the  watershed areas. Additional
 water quality studies, watershed management ac-
 tivities,  and implementation  projects are "continually
 underway in Albemarle County's water  supply water-
 sheds. The most recent  has been a Phase II Clean
 Lakes Project funded by  the Federal government to
 provide  for the  implementation of  agricultural  and
 highway best management projects in a  portion of the
 South Fork Rivanna Watershed. This program initiated
 in 1980, is due to continue through 1985 and provide
 $500,000 of Federal funds  for  locally implemented pro-
 jects to improve water quality (Browne,  1981).
FUTURE CONCERNS

Two areas of concern for future water supply protec-
tion and availability  in the Charlotesville/Albemarle
area have been the issues of the cost of water supply
protection measures and who should bear those costs
and the availability of adequate supplies for future
growth of the area.
  Since January 1978 the city of Charlottesville and
the county of Albemarle have struggled with the insti-
tutional arrangements for who should bear the cost of
water supply protection. Numerous ideas and sugges-
tions were presented during the past few years, how-
ever no agreement was reached until January 1983. In
                                                562

-------
                                                                               WATERSHED MANAGEMENT
a joint resolution the county of Albemarle, the city of
CharlottesviMe, the Albemarle County Service Authori-
ty, and the Rivanna Water and Sewer Authority agreed
to the following: (1) the periodic inspection and main-
tenance of the devices required by the County's Run-
off Control  Ordinance would be performed by the Ri-
vanna Water and Sewer Authority as the agent of Albe-
marle County and that the cost of the same shall be
borne by the  Rivanna Water  and  Sewer Authority
through its water rates. The cost of inspection and
maintenance, if not obtainable from the private owner
or developer pursuant to County ordinances, would be
charged to the water rates of the water system directly
served by the drainage area in which the device was
located; (2) the cost of the watershed management
official's off ice shall be paid by the Rivanna Water and
Sewer Authority through its water rates, but the water-
shed  management official shall  remain  administra-
tively within the County management structure. The
cost of  the  watershed management official's  office
shall be prorated over the various Rivanna Water rates
of the water system directly  serving the  drainage
areas in which the official spends his time (Albermarle
County, 1983).
   In resolving these cost issues the signatories to the
resolution also provided the mechanism by which the
community can acquire a supplemental water supply
site.
   Efforts to protect and preserve  the site of a  future
water supply impoundment started as early as 1979.
The Rivanna Water and Sewer Authority requested of
the county of Albemarle to place a moratorium on the
area of a future proposed water supply impoundment
in the Buck Mountain Sub Watershed of the  South
Fork Rivanna Watershed while feasibility studies were
being conducted. The Albemarle County Board  of Su-
pervisors enacted a 2-year building moratorium  in
August  1980  to   provide  protection  for both  the
impoundment and an area for flooding. During the
moratorium  the Rivanna Water and Sewer Authority
had a feasibility study conducted  to define the costs
of the project and the limits of the project area (Camp
Dresser & McKee, 1982).  Following  a more accurate
delineation of the area proposed for the impound-
ment, more time was needed to complete the feasibili-
ty study and arrange for the property acquisition. The
building moratorium providing  a  protection against
additional construction in the project area is now set
to expire on December 31, 1984.
   During the time that the Buck  Mountain area has
been under moratorium: (1) three of the four phases of
the feasibility study  have  been  completed; (2) a dam
site has been selected which would create a reservoir
with a storage capacity of 136.8 billion liters, a surface
area of 180 hectares, a drainage area of 54.88 square
kilometers and  a safe yield of 72.96 million liters per
day, and; (3) the amount of land to be acquired for the
future impoundment has been agreed to by the city
and the county (Camp Dresser & McKee, 1982).
   Even though  it has been projected that the supple-
mental water supply impoundment  would not  be re-
quired until  the year 2015, steps are currently under-
way to acquire the property.  The property to be ac-
quired will include: (1) the area of the pool and the dam
site and  a minimum buffer zone of 9,000 horizontal
centimeters measured from the normal pool level by
the fee simple method, and; (2) a 3,000 centimeter buf-
fer zone on either side of the tributary streams and all
land  lying  between  a  9,000  centimeter  buffer
measured from the normal pool level and a 9,000 cen-
timeter buffer measured from  the 100-year flood plain
by easement. It has been projected that the land ac-
quisition phase of the project will be completed  in 2
years at a cost of over $5 million.
CONCLUSIONS

Any  attempt at  watershed management and  water
supply protection has to be undertaken with a great
deal  of understanding and background information. If
the  Charlottesville/Albemarle  community  had  not
conducted   water  quality  studies  and  numerous
meetings with the various agencies and groups  in-
terested  and  involved in the area's  water  supply
resources, a successful watershed management pro-
gram could not have been initiated.
  Every water supply reservoir in the country should
be protected, but actions should not be taken hastily.
A firm basis in hard facts and an informed and com-
mitted public are prerequisites of a management pro-
gram. Successful watershed management, out of ne-
cessity, requires  both cooperation and compromise.
 REFERENCES

 Albemarle County,  Va.  1983.  Joint Resolution—Water.
  shed Management Costs. Jan. 5. Albemarle County Ser-
  vice Authority, City of Charlottesville and Rivanna Water
  and Sewer Auth.

 	. Code of the County, The General Ordinances. Pub-
  lished by order of the Board of Supervisors, Chapter 19.1
  Water and Sewers. Article II. Protection of Public Drinking
  Water Pages 196.2-196.7. The Michie Co. Charlottesville.

 Betz Environmental Engineers, Inc. 1977. Water Quality Man-
  agement Study of the South Rivanna Reservoir and Tribu-
  tary Area. Prepared for  Rivanna Water Sewer Author.

 Browne, F.X.  and Associates. 1979. Rivanna  Reservoir
  Restoration Project. Prepared for Rivanna Water Sewer
  Author.

 	. 1981. Phase II Implementation Project for the Rivan-
  na Reservoir and Watershed. Revised work plan prepared
  for Rivanna Water Sewer Author.
 	.  1982. 208 Watershed Management Study of the
  South Rivanna Reservoir. Prepared for the County of Albe-
  marle, Charlottesville, Va.

 Camp Dresser and McKee, Inc. 1977.  Report on Alternative
  Water Supply Sources for Rivanna Water Sewer Author.
 	1982. Buck Mountain Feasibility Study  Phase III.
  Final Rep. for Rivanna Water Sewer Author.

 Norris, W.K. 1980. South Rivanna Reservoir: A Brief History
  and an  Unsolved Problem.
                                                 563

-------
 A SCREENING METHODOLOGY FOR THE
 SELECTION OF URBAN  LAKES' ENHANCEMENT
 CARLA N. PALMER
 MARTIN  P. WANIELISTA
 Department of Civil Engineering and Environmental Sciences
 University of Central  Florida
 Orlando,  Florida

 RUSSELL  L. MILLS
 GILBERT NICHOLSON
 Dyer, Riddle,  Mills and Precourt, Inc.
 Orlando,  Florida

 ROBERT HAVEN
 Director of Public Works
 Orlando,  Florida


           ABSTRACT

           In the summer of 1983,106 lakes in the city of Orlando, Fla. were considered in a selection pro-
           cess to determine which most needed restoration in the City's Lake Enhancement Program. The
           existing lake systems were thoroughly investigated with regard to historical water quality data;
           existing water quality data; current public uses; visual, physical, and chemical condition; the
           drainage basins; the stormwater structures and management controls; as well as surface color
           aerial and false color infarared aerial photometric reconnaissance. These were analyzed to pro-
           vide an objective screening process to determine a discrete set of lakes, approximately four to
           six, which may be top candidates for a pilot lake project. This paper describes the screening pro-
           cess.
INTRODUCTION

The City of Orlando, Fla. has embarked upon a Water
Quality Enhancement Program for the degraded lake
systems in the City. For a water conscious communi-
ty, this study presents the first phase of a program;
assessing the lake systems and the providing a listing
of lakes from which an Enhancement Program  could
be initiated.
  The program includes four phases:
  Phase I: Lake Assessment Study
  Phase II: Design and Construction of the Pilot Lake
          Project
  Phase III: Monitoring, Evaluating, and Maintaining a
          Pilot Lake Project
  Phase IV: Designing, Constructing, and Enhancing
          Additional Lakes
  Most of Orlando's lakes result from sinkholes form-
ed by the dissolving of the limestone bedrock. Those
lakes are known as solution lakes. The depth and level
of water in the solution lakes is highly variable. Usual-
ly, the depressions are of sufficient depth to extend in-
to the groundwater table and permanently contain
water. Others fluctuate in water level in response to
seasonal  and  long-term   variations  in hydrologic
events.
  Urban development  acts   to  concentrate   soil,
nutrients, and some heavy metals in lakes from flood
control and stormwater management. The result is an
accelerated aging process for the lakes. Urban  struc-
tures and facilities for the city's inhabitants were sited
to take advantage of the natural beauty and serenity
of the lakes. Inadvertently the acts of living, working,
and driving close to the lakes has tended to wear them
toward premature old age.


SCREENING METHODOLOGY

The purpose of the Phase I Study is to determine the
four to six lakes which are the prime candidates for
the Pilot Lake Study in Phase II. The analysis of the
drainage basins in the city's Growth Management
Area indicated  a total of 106 lakes that may be con-
sidered for enhancement. It is important when selec-
ting only four to six from this large number of lakes to
be as objective as possible. It also should  be realized
that in the final analysis further considerations may
exclude some lakes and  include others. This paper
presents a method for objectively screening the lakes.
A flow chart of the screening process is shown in
Figure 1.
INITIAL SCREENING

The initial screening criteria included lakes in the city
of Orlando  for which available water quality  data
showed a trophic state index of 60 or more. The city
decided to consider those lakes that are almost entire-
ly within its boundaries, or whose drainage areas are
                                             564

-------
                                                                                 WATERSHED MANAGEMENT
within its boundaries. Lakes with large tributary areas
outside of the city require the joint participation of
other governmental bodies in their enhancement, and
may be considered at a later  date.
  Since the objective of lake enhancement is to im-
prove water quality, water quality data must be  con-
sidered as a  criteria in the  screening process.  The
Orange County Pollution  Control Department (OCP-
CD) collects data on a number of lakes within the  city,
but not all of them. Because of the time and expense
involved in collecting data on all of the lakes within
the city, it was determined that  lakes for which no
data were available would be  deleted from the screen-
ing analysis.
  The water quality data collected by the OCPCD in-
cluded physical data such as  temperature and Secchi
disk  reading  (a  measure of light  transparency);
chemical  data such as phosphorus,  nitrogen,  bio-
chemical oxygen demand; bacteriological data such
as fecal coliform bacteria and fecal strepococci,  and
biological data such as chlorophyll a and  counts of
particular species of algae. These data were examined
for completeness,  while trophic indices were examin-
ed to see which may be used in screening  Orlando's
lakes.  The Carlson Trophic  State  Index (Reckhow,
1979) was considered the most suitable index for in-
itial screening because it can utilize chlorophyll a, the
data most consistently available for all  lakes, and the
scale allows comparisons of one lake to another. The
relative ranking of  the lakes is important for a screen-
ing  process.
  The Carlson Trophic State Index (TSI) rates lakes on
a scale from 0 to 100, with 100 being the most trophic
state. The index  for chlorophyll a is  listed  below
(Reckhow, 1979).
 TSI Chlorophyll a (mg/m3)* TSI Chlorophyll a (mg/m3)*
0
10
20
30
40
50
0.04
0.12
0.34
0.94
2.6
6.4
60
70
80
90
100

20.
56.
154.
427.
1183.

   A TSI rating of 60 was selected as the cutoff for fur-
ther consideration. Lakes with a TSI of less than 60
were  considered to have a sufficiently  good water
quality that enhancement at this time would not be ap-
propriate. The Secchi disk reading corresponding to a
TSI of 60 is one meter.
   Lakes that remained in the analysis after the initial
screening received further analysis based on new field
and water quality information and other considera-
tions.
SECOND SCREENING

The objective of the second screening was to reduce
the numbers of lakes that remain  after the initial
screening to a discrete set of approximately four to six
lakes.
  A  comprehensive field investigation of the lakes
was  made prior to the  completion of the second
screening. This  investigation  included  additional
water quality data (collected by  the OCPCD), addi-
tional physical  measurements of the lakes, and the
determination of each lake's benefit to the public.
       MAJORITY OF
         LAKE AND
      DRAINAGE AREA
         W/l CITY
             YES
         WATER
       QUALITY DATA
        AVAILABLE
             YES
         CARLSON
           (TSI)
           >60
              YES
    NO
                            NO
   NO
    FIELD DATA COLLECTION
       WATER QUALITY
    PHYSICAL PARAMETERS
      FACILITIES SURVEY
Lakes with no water quality
                                                                                data and within city
                                                                                 NO
                            NO
                                                                  YES
  WEIGHTING PROCEDURE AND
     SENSITIVITY ANALYSIS
  CONSIDERING WATER QUALITY,
  BENEFICIAL USES, LIKELIHOOD
     OF POSITIVE RESULTS,
      POTENTIAL COSTS
Figure 1.—Screening Process.
                                                 565

-------
 LAKE AND RESERVOIR MANAGEMENT


 FLORIDA TROPHIC STATE INDEX

 In a comprehensive report to the Florida Department
 of  Environmental  Regulation,   Brezonik  (1982)
 evaluated  the Carlson Index as a model for ranking
 Florida lakes and recommended a modified Carlson
 Index for use in Florida. The procedure for calculating
 the  Trophic State  Index  (TSI), as recommended by
 Brezonik,  is listed in Table 1.  This  procedure  was
 followed, using newly obtained data from the OCPCD,
 for all  lakes remaining after  the  initial screening.
 Brezonik also suggested a TSI of 60 as a threshold for
 considering water quality as becoming  a problem in
 Florida lakes.
 THE BENEFICIAL USE INDEX

 To objectively evaluate the beneficial use that a lake
 may have to the public a Beneficial Use Field Form
 was developed. The Beneficial Use Field  Form was
 designed to objectively rate the amenities of each lake
 to the public as well as compare one lake's benefit to
 another's. The complete Beneficial Use Field Form is
 shown in Figure 2. The form was divided into five ma-
 jor categories  including:  (1) amenities,  (2) act ve
 recreation, (3) public access, (4) land value,  and  (5)
 visual access. The amenities listed included concert
 areas, sidewalks, children's facilities, adult facilities,
 structures for visual appeal, and concessions. If a la ke
 exhibited these amenities, a score of 5 points for each
 classification was  given.  Likewise,  if swimming,
 fishing, or boating were evident, points were awarded.
 Similarly, though on differing scales, the  other criteria
 were weighted. Therefore, a lake with a high Beneficial
 Use Index was considered to contribute more to the
 quality of life in Orlando than one with a lower value.
 Photographs were taken at each lake to document the
 field form.


POSITION OF LAKES

The position of lakes with respect to upstream lakes
and the TSI of upstream lakes is important in con-
sidering  which  lakes remain   as  candidates lor
enhancement. To try to improve the water quality of a
                                                   lake which is downstream of a lake with poorer water
                                                   quality would be "putting the cart before the horse." It
                                                   is better to improve the water quality in the upstream
                                                   lake first.  Therefore,  lakes which are downstream of
                                                   lakes  with a higher TSI  were eliminated from further
                                                   consideration.
                                                   AERIAL PHOTOGRAPHS

                                                   Since nearly hall of the lakes in the study area lacked
                                                   previous water quality data, the Assessment Team
                                                   sampled these lakes from a macroscopic view point.
                                                   To accomplish this purpose, aerial photos were taken
                                                   of all the City's lakes. The pictures were taken in both
                                                   false color infrared  and regular color.  Not  all lakes
                                                   have the same reflectance  with false color infrared
                                                   photometry.
                                                     The photos were analyzed to obtain a cursory view
                                                   of the relative trophic states of the lakes. This infor-
                                                   mation gained from the aerial photos supported  the
                                                   Florida TSI findings. One lake for  which no previous
                                                   water quality data had been available was indicated
                                                   as eutrophic and, therefore,  added to the list of lakes
                                                   for secondary screening. Subsequent calculation of
                                                   the Florida TSI indicated its score  was above 60.
                                                   IMPERVIOUS AREA TO LAKE SURFACE
                                                   AREA RATIO

                                                   Since the quality of many urban lakes is affected by
                                                   stormwater runoff, a ratio of impervious area in the
                                                   lake's drainage basin to lake surface area (IA/LA) was
                                                   calculated based on the assumption that 35 percent of
                                                   residential soils are covered over by impervious sur-
                                                   faces,  and 80 percent  of commercial and industrial
                                                   lands are impervious. The (IA/LA) ratio is to be used as
                                                   a general indication of the probability of success in
                                                   cleaning up a polluted lake. For lakes of similar shape
                                                   and depth there is a positive relationship between the
                                                   impervious area to lake area ratio and trophic state in-
                                                   dex. The idea behind this ratio is that the lakes receiv-
                                                   ing large amounts of stormwater runoff are the ones
                                                   most  likely  to  become polluted. In turn, these  im-
                                                   pacted  lakes may be the most difficult to clean up
                 Table 1.—Empirical procedure for calculating the Florida trophic state index.
          Phosphorus-Limited Lakes (TN/TP > 30)
          TSI (Avg) = 1/3 [TSI(Chl a) + TSI(SD) + TSI(TPI)]
          TSI (Chi a) =  16.8 + 14.4 In Chi a, (mg/m=)
          TSI (SD) = 60.0  - 30.0 In SD (mg)
          TSI (TP) = 23.6 In TP - 23.8 (
    where:
            Nitrogen-Limited Lakes (TN/TP < 10)
            TSI (Avg) = 1/3 [TSI(Cha a) + TSI(SD) + "SI(TN)]
    where:   TSI (TN) = 59.6 + 21.5 In  TN (mg/l)
  where:
            Nutrient-Balanced Lakes (10^
            TSI (Avg) = 1/3 [TSI(Chl a) + TSI(SD) + 0.5 (TSI(TP) + TSI(TN))]
            TSI (TN) = 56 + 19.8 In TN (mg/l)
            TSI (TP) = 18.6 In TP - 18.4
IV.
            TSI (Carlson) = 0.65 TSI (Florida) + 23.2
  Note TSI  = Trophic state index
      Chi a = Chlorophyll a
      SD  = Secchi Disk
      TP  = Total phosphorus (unfiltered)
      TN  = Total nitrogen
  "Source Brezonik (1982)
                                                566

-------
                                                                              WATERSHED MANAGEMENT
Lake Name'
Date:
A. AMENITIES AT LAKE
1. Concert area
2. Sidewalk along lakeside
3. Facilities for children
4. Facilities for adults
5. Structures for visual appeal
6. Food concessions
B. ACTIVE RECREATION IN LAKE
1. Swimming
2 Fishing
3 Boating
C. PUBLIC ACCESS
1. Percent of shoreline open to public
2. Boat ramps

D. LAND VALUE
1. Percent of shoreline with residential
property
E. VISUAL ACCESS
1. Public lakes
2. Private lake
SURVEYOR'S NAME:
NIIMRFR OF PHOTOGRAPHS:
ROI F NUMBER:

Lake Number:


No-0
No-0
No- 0
No-0
No-0
No-0
No-0
No-0
No-0


<5% 5-10% 15-40% 45-60%
036 9
No-0
<5% 5-10% 15-40% 45-60%
036 9

Low - 0 Good - 4
Low - 0 Good - 1

TOTAL POINTS:
EXPOSURE NUMBERS:




Yes -5
Yes -5
Yes -5
Yes -5
Yes -5
Yes -5
Yes -5
Yes -5
Yes -5


60-90% > 90%
12 15
Yes -5
60-90% > 90%
12 15

High -7
High -3




Figure 2.—Orlando Lake's beneficial use field form.
unless redirection and treatment of stormwater is ac-
complished.
  There are exceptions to this relationship. For in-
stance, some small, deep Orlando lakes do not adhere
to this general trend.  It is inferred  that pollutants
entering these types of lakes are carried to the bottom
and  remain in the sediments, unavailable to  the pro-
ductive zone at the surface. Therefore, the IA/LA ratio
should not stand alone as a criterion for the likelihood
of successful enhancement.
STORMWATER FLOWS

The  number of incoming stormwater flows to each
lake can be used as a measure of the relative cost to
install stormwater control devices. A lake having more
incoming flows than another lake is likely to have a
higher cost associated with controlling the inflows. Al-
though this may not always be the case, especially if
one lake has a large number of small inflow pipes and
another has a small number of large pipes,  it is a
useful  parameter at the level of detail  being con-
sidered  in the weighting procedure.
RANKING PROCEDURE

The next step in the second screening process was a
weighting procedure composed of four parameters: (1)
TSI, (2) beneficial use, (3) impervious area to lake area
ratio, and  (4)  the  number  of incoming stormwater
flows into the lake.
   TSI was described previously. The lakes were rank-
ed giving  the  highest TSI the highest  rank and the
lowest TSI the rank of one. Beneficial use was deter-
mined and quantified using the form described earlier.
This parameter is included in the analysis to account
for the differences in the public use of the lakes in the
city. The lake with the highest relative beneficial use
will be  ranked  the  highest, and the lake with the
lowest will be ranked as one. The IA/LA ratio was rank-
ed in  an inverse manner— the lake with the highest
ratio was ranked one and the lake with the lowest ratio
was ranked highest. A lake with  a large number of in-
flow pipes would be  ranked as one, while the lake with
the least number would be ranked the highest.
  The objective is to choose the lakes which are the
"best" candidates  for restoration.  Here,  the  term
"best" is defined by water quality, beneficial use, abili-
ty to control  the stormwaters, and cost. We have a
multiple objective function of rankings which must be
satisfied. In mathematical terms, one can write the
priority levels:

1st: Water Quality Ranking

         WWTSI,
where:    Ww = weight for water
                quality
          TSIj = TSI rank for lake
                "i"

2nd: Beneficial Use
where:    WB =
          BU, =
                weight for
                beneficial use
                beneficial use rank
                for lake "i"
                                                 567

-------
LAKE AND RESERVOIR MANAGEMENT

3rd: Ability to Control

         W|(IA/LA)i

where:     W, =  weight for chance
                 of success
       (IA/LA)| =  impervious area ratio
                 rank for lake "i"

4th: Cost
where:    Wc= weight for cost
            c _ number of inflows
            'i  ~ rank for lake "i"
     and: Ww  + WB + W, +  Wc =  1.0

so that:
     Maximize Ij (Ww TSI,  + WB BU| + W, (IA/LA), +
     Wcl|)

  The weighting procedure consists of ranking the
lakes with respect to each parameter, multiplying the
ranks by the weight and adding the products to obtain
the score. This procedure  is illustrated in Table 2 for
five lakes with assumed values as shown. In the exam-
ple shown, each of the parameters was given equal
weight (0.25). The sum of the  weights must equal 1.0.
The highest score any lake could have would be equal
to the  number of  lakes  being evaluated  in  the
weighting procedure. In the example, the  highest
possible score is five.
  The weighting procedure can be repeated by varying
the weights to determine how the ranking is changed
as the weights change. Such a process is called a sen-
sitivity analysis and enables a decisionmaker to deter-
mine  how  sensitive his choice  is to variation  in
weights.  Four repetitions of the weighting  procedure
are suggested  to  reflect various  preferences  in
weights. These are listed as follows:
Repetition
Parameter
TSI
Beneficial Use
IA/LA Ratio
Number of inflows
First
0.25
0.25
0.25
0.25
Second
0.40
0.20
0.20
0.20
Third
0.20
0.40
0.20
0.20
Fourth
0.20
0.20
0.40
0.20
  The first repetition, with all weights equal, indicates
no preference for any one parameter. In the second
repetition, a preference is indicated for the TSI; that is,
water quality is given the highest weight. The third
repetition shows a preference for beneficial use while
the last repetition shows a preference for confidence
in being able  to  control  stormwater  inflows.  The
number of inflows,  which  is  used as a measure of
cost,  is not given a higher weight because, at  this
stage, it is not considered more important than the
other three parameters. When performing sensitivity
analyses, it  is important to vary one parameter at a
time so  that results  can  be readily interpreted.
                                                     DISCUSSION

                                                     The methodology outlined  was the result of an  at-
                                                     tempt to objectively evaluate 106 Orlando city lakes to
                                                     determine those most in need of or that could benefit
                                                     from a pollution abatement,  lake enhancement pro-
                                                     cess.
                                                       The number of lakes entering the screening process
                                                     at the top of Figure 1 was reduced to 80 following  ap-
                                                     plication of the first criteria "majority of lake drainage
                                                     area within city." Forty-three lakes remained that had
                                                     historical "water quality data available." Twenty-five
                                                     lakes further qualified for the program because their
                                                     "Carlson TSI >60". Sixteen  lakes had a "Florida TSI >
                                                     60." Only two lakes were positioned such that their  up-
                                                     stream lakes in a chain of lakes did not "have a lower
                                                     TSI." Fourteen lakes  were entered into the weighting
                                                     procedure detailed in the preceding section. The sen-
                                                     sitivity analysis was performed and the  lakes were
                                                     ranked according to the highest scores receiving the
                                Table 2.—Illustration of weighting procedure.
Lake
Number
1
2
3
4
5

TSI (Rank)
65(2)
80(4)
74(3)
63(1)
82(5)
Beneficial
Use (Rank)
40 (4)
2- (1)
3-' (3)
50 (5)
2« (2)

(IA/LA) (Rank)
5.0 (2)
2.8 (4)
1.5 (5)
3.0 (3)
10.0 (1)
Number of
Inflows (Rank)
6(3)
10(1)
2(5)
8(2)
4(4)
                                       Product of weight and rank
Lake
Number
1
2
3
4
5

TSI
0.50
1.00
0.75
0.25
1.25
Beneficial
Use
1.00
0.25
0.75
1.25
0.50

IA/LA
0.50
1.00
1.25
0.75
1.00
Number
of Inflows
0.75
0.25
1.25
0.50
1.00
Total
Score**
2.75
2.50
4.00
2.75
3.75
 ' All weights are even m this example, i e, 0 25                                                   "
 ' Total score is the sum of products of weight and rank Lake with highest total score, Lake #3, ranks highest for enhancement
                                                568

-------
highest  rank. The methodology objectively selected
the top  six  ranking lakes as prime candidates for
enhancement.
CONCLUSION

The screening methodology represents a mixture of
scientific,  social,  and  economic  parameters  to
organize  a large number of lakes into a manageable
number of "prime" candidates for enhancement. The
remaining lakes can then be studied more fully, pollu-
tion abatement concepts may be designed more ac-
curately with the specific lake in mind, and conceptual
cost analysis may be estimated based on available
drainage  and water management data for each basin.
                          WATERSHED MANAGEMENT

The methodology objectively screened  106  Orlando
area lakes down to six. After conceptual cost analyses
are made for cleanup of each of the six lakes, the City
will make the final decision as to which lake becomes
the pilot for lake enhancement.
REFERENCES

Brezonik, P. In press. Development of a trophic state index
  scheme to rank Florida lakes.Chap. 3. Fla. Dep. Environ.
  Reg.
Reckhow, K.H. 1979. Quantitative Techniques for the Assess-
  ment of Lake Quality. EPA-440/5-79-015. U.S.Environ. Prot.
  Agency, Washington, D.C.
                                                569

-------
 COMPREHENSIVE MONITORING  AND EVALUATION
 OF THE BLUE CREEK WATERSHED
 THOMAS E. DAVENPORT
 Illinois Environmental Protection Agency
 Springfield, Illinois
             ABSTRACT

             Initial water quality planning efforts documented that agriculture activities are a major source of pollu-
             tion in Illinois. The most severe agricultural related problem is soil erosion resulting in sedimentation.
             In Illinois, estimated gross erosion exceeds 180 million tons annually, wilh 88 percent from cropland
             Illinois Environmental Protection Agency in cooperation with various U.S. Department of Agriculture
             agencies evaluated the water quality impacts of resource managment system implementation under
             the ACP Special Water Quality Project in the Blue Creek Watershed. The physical, biological, and
             chemical characteristics of the Blue Creek Watershed Lake have been studied since 1979 A com-
             prehensive monitoring network for the entire Blue Creek Watershed was established to document
             the basic hydrological, meterological, and water quality factors of the project area during 1980  The
             duration, timing and quantity of nonpoint source pollutants were evaluated within the watershed  The
             monitoring program was designed to link water quality to what  is happening on (he land. To unders-
             tand the processes of soil erosion and nutrient/sediment transport and to draw meaningful conclu-
             sions about land use effects, focus was givsn to  sources  and movement from sources to the  lake
             outlet. This evaluation  integrated reservoir sedimentation surveys,  lake water quality  monitoring,
             biological monitoring, water quality monitoring at stream gauging stations and small field sites, a channel
             dynamics study, and a computerized gross erosion estimate The stream gauging station and  field
             site monitoring were event-oriented  sampling to supplement the baseline monitoring on the stream
             network and in the lake. There were obvious seasonal and spatial trends exhibited by several physical
             and chemical parameters within the lake and watershed. In Illinois, erosion control is being used as
             a surrogate for sediment control because sediment control  is less amenable to quantitative analysis
             This reflects the current lack of knowledge concerning  sediment origin, transport, deposition, and
             control technology Integration of source and sediment budget studies with realistic concepts of storm
             runoff production results from this project will clarify some aspects of the interrelationships between
             gross erosion and water quality impacts.
Half of the land in the Nation is in  agricultural or
related uses. Agriculture is  the most widespread
cause of nonpoint source pollution. Initial 208 water
quality planning  efforts  demonstrated  that  agri-
cultural activities are a significant source of pollution
in many parts of the country. The resources required
to correct agricultural nonpoint source problems are
substantial  and will require  that both public and
private investment be used in the most cost-effective
manner.
   Nonpoint source pollution prevents the optimal use
of the water resources. This pollution  can originate
from runoff of agricultural land or from groundwater
return flow draining agricultural land. Surface runoff
from agricultural land transports nutrients, pesticides,
disease  causing  organisms,   and organic  and  in-
organic  particulate material  to water  resources.
Groundwater return  flows cause nutrient, pesticide,
and dissolved solids pollution problems. Eroded sedi-
ment is the most significant  pollutant in terms of
damage caused and total load carried by a stream. Ef-
ficient Resource Management Systems (RMS) must be
installed to correct these  problems.
  To successfully implement nonpoint  source control
programs one must evaluate past and ongoing control
projects. Watershed evaluation programs  must pro-
vide guidance to be of use for inputs to many adminis-
trative functions, such as: (1) investment of funds, (2)
justification for allocation of these funds, (3) verifica-
tion of overall program effectiveness, (4) evaluation of
regional effectiveness of RMS's, and  (5)  informing
local land owner operators as to the effectiveness of
their efforts to improve water quality. The Blue Creek
Watershed project will help provide this guidance.
  The  Blue Creek Watershed monitoring and evalua-
tion project is short term due to availability of funds.
Large incremental changes in the overall lake water
quality caused by RMS's implemented have not been
measurable within  the  4  year monitoring period
because of the high  degree  of inherent variability
within  the system and  the  long response time of
natural ecosystems to such subtle changes. Findings
have been reported for each component and phase of
the project.
 STUDY AREA

 The  Blue Creek Watershed  encompasses  slightly
 more than 7,000 acres in east central Pike County, III.
 Terrain is hillier than most areas of Illinois and has a
 high soil loss  potential  because of its steep slopes,
 fine-textured soils,  and agricultural land  use prac-
 tices. Over 80 percent of the soils within the watersh-
 ed have a soil eirodibility factor of 0.37. The Blue Creek
 Watershed drains into Pittsfield City Lake, which was
 constructed  in  1961 as  a multiple purpose reservoir.
 Recreational areas are concentrated around the lake.
 Average annual soil loss is estimated at 9.0 tons/acre/
 year (Davenport, 1983). Erosion from livestock opera-
 tions, primarily hogs, significantly contributes to the
 total basin sediment load. For a detailed description
                                                  570

-------
                                                                               WATERSHED MANAGEMENT
of the Blue Creek Watershed and the monitoring pro-
gram  refer to  Blue Creek Watershed Project  Pike
County, Illinois, May 1977-October  1980 (Davenport,
1981).
 BLUE CREEK WATERSHED PROJECT

 The Blue Creek Watershed in Pike County,  III., was
 selected to demonstrate the effects of erosion control
 practices upon water quality. The project area was
 designated  an  Agricultural  Conservation  Program
 (ACP) Special Water Quality Project Area, which pro-
 vides financial  assistance to land owners/operators
 for implementation of conservation and pollution con-
 trol practices. The  Illinois Environmental Protection
 Agency (IEPA) in cooperation with U.S. Department of
 Agriculture  agencies, Agricultural Stablization and
 Conservation Service  (ASCS) and Soil Conservation
 Service (SCS),  is evaluating the  impact  of RMS's
 implemented under the ACP Special Water Quality
 Project in the Blue Creek Watershed from May 1980 to
 September 1982. The monitoring and evaluation pro-
 ject's primary purposes  are to assess gross  erosion,
 determine the actual yield of sediment and nutrients
 from different agricultural practices to downstream
 receiving water, and to examine their effect on quality
 and use of the water resources.


 Institutional Aspects of the Blue Creek
 Watershed Project

 There was local interest and support for a water quali-
 ty program within the watershed. The area had been
 targeted  for various water quality programs in the
 past; though never actually funded and implemented,
 they had been supported by the local and State agri-
 cultural community (Davenport, 1981).
  A close working relationship developed between the
 project staff and local  agricultural agencies. Three
 primary methods were used to encourage adoption  of
 resource management systems in the project area. An
 intensive information/educational effort, spearheaded
 by the Cooperative Extension Service, was conducted
 using meetings, publications, and one-on-one con-
 tacts. The purpose was to inform the public about the
 project,  water quality problems,  resource manage-
 ment systems, and their potential involvement in the
 project.  SCS and Pike County Soil and  Water Con-
 servation District personnel went out into the field and
 helped eligible and interested farmers. ASCS provided
 cost-share incentives.
  The key to the success of obtaining the land treat-
 ment goal (soil erosion reduction) was the teamwork
of the local agency and organizations personnel, who
combined to provide assistance and education.
MONITORING STRATEGY

The comprehensive monitoring network for the entire
Blue Creek watershed was established by IEPA,  in
cooperation with Illinois State Water Survey, to docu-
ment the basic hydrological, meterological and water
quality factors of the project area (Fig. 1). The dura-
tion,  timing,  and  quantity   of  nonpoint  source
pollutants  are being evaluated  to  determine land
management  effects  on the  water quality/quantity
budget. The main stem of Blue Creek is monitored  at
two locations (Stations C & B) representing 50 and 70
percent of the drainage area, respectively. One direct
 tributary (Station D) to the lake is monitored to deter-
 mine the relative contribution of a major sub-basin
 directly entering the lake and to provide data for unit
 size contribution purposes. Two fields (Stations E & F)
 of 38 and 79 acres respectively are being monitored
 concurrent with  the  major  tributary  and sub-basin
 monitoring. Thus, spatial  patterns, land use, and
 physiographic  characteristics  can  be  evaluated
 relative to their actual and projected impacts on down-
 stream water quality and lake characteristics.
  There  were four biological monitoring  stations in
 the watershed to collect macroinvertebrates within  a
 given area. The lake was monitored at three locations,
 six times a year since 1979. The lake will be monitored
 until 1985, if funding permits.

 Monitoring Methods

 Methods used for field and laboratory procedures are
 those  accepted  for  agriculturally related hydrology
 research and  lake water quality sampling. Methods
 are documented  in  Water Resource  Data and Pre-
 liminary Trend Analysis  for the Blue Creek Watershed
 (Davenport, 1983).

 Results & Discussion

 Three periods based on  the amount  of surface cover,
 precipitation pattern, and land management activities,
 were selected for evaluating  seasonal variation  in
 stream flow constituent concentrations and transport.
 These periods are: PV.  fertilizer and  seedbed esta-
 blishment  period (April-June); P2:  reproduction and
 maturation period (July-November);  and P3: residues
 period (December-March).

 Suspended Sediment within Pittsfield
 City Lake
 Mean surface values  for samples taken during 1983
 were 17.4 and 9.3 mg/l for total suspended solids (TSS)
     stream sampling
       sites

     inlake sampling
       sites
Figure 1.—Map of water quality stations in the Blue Creek
watershed in Pike County, Illinois.
                                                571

-------
LAKE AND RESERVOIR MANAGEMENT
and total volatile solids (TVS), respectively. In  1979,
1980, and  1981  mean TSS concentrations were 41.1,
38.6 and 25.2 mg/l, respectively. On  the average  in
1981, TVS comprised 17 percent of TSS, in 1982 this
percentage was 25 percent.  In 1982 surface watsr
samples the percentage of TSS comprised by TV'S
ranged from 21 to 24 percent, whereas at both bottom
water sites TVS comprised 14 percent of TSS. This in-
dicates the suspended solids were probably inorganic
in  nature.  Inorganic suspended soils in  many Illinois
lakes result largely from soil erosion  and the subse-
quent runoff of soil particles into reservoir tributaries
(Sefton et al. 1980). Since the TSS/TVS ratio decreased
from 1981 to 1982, it indicates a decrease in inorganic
suspended solids  since  TSS decreased. Maximum
observed values for TSS and TVS were 60.0 mg/l and
10.0 mg/l,  while minimum values were 22.0 mg/l and
4.1, respectively. Maximum and minimum values for
both TSS and TVS were lower in 1982, 1980 and 1979,
in    comparison   to    1981    values.   An
"upstream—downstream" difference in  TSS and TVS
exhibited during 1979,  1980, 1981 also was present
during  1982.
  Seasonally, surface TSS and TVS values decrease
between P1 and P2 during 1979, 1980, and 1981. This
probably results from canopy establishment during P2
and the decrease  in potential source areas of tolal
suspended solids.
  Table 1.—Mean Lake surface water TSS concentration!!
            by period of analysis by year.
          1979
           1980
           1981
           19D2
P1
P2
P3
28.5
25.3
23.0
13.7
11.3
28.4
21.7
18.3
47.0
  Table 2.—Mean lake surface water TVS concentrations.
            by period of analysis by year.
          1979
           1980
           1981
           1932
6.0
5.3
*
4.4
3.7
3.0
7.8
6.0
*
5.6
6.1
<
P1
P2
P3
  Secchi transparency within the lake was looked at
to determine spatial variation in TSS loadings. There
is a definite upstream-downstream  effect: the lowest
transparencies were  measured upstream in cross-
sectional measurments after rainfall events  showed
the western portion of the lake had  less transparency
than the eastern portion. The land adjacent to  the
eastern portion of the lake is managed park land. In
comparison,  the  western portion  has considerable
drainage of cropland and pasture areas. The decreas-
ed transparency probably results from heavy sediment
contributions in the runoff from the western side. High
sediment concentrations measured at Station D (Fig.
2) in comparison to Station  B's concentrations,  in-
dicates that on a per  acre basis the western side of
the lake is delivering more  sediment than the  re-
mainder of the lake's drainage area.
  To  relate trends in lake quality  to  RMS imple-
mentation, a series of relationships will have to  be
demonstrated: first, relationships between field sites
and stream sites, and then relationships from stream
to lake. These relationships are based on precipita-
tion-related effects   on sediment  transport.  The
greatest potential  for  documenting  improvement
related to RMS's is P1 when flow into the lake is ex-
tremely high, and the lake essentially becomes an ex-
tension of the stream. At this time, because flows are
high and water is moving through the lake at a high
rate (detention time is probably measured in days or
weeks), well defined relationships are assumed to ex-
ist between stream and lake sites.
  Stream monitoring results from 1980 to 1982 docu-
mented a change in the sediment transport function
within the watershed (Davenport, 1983). Sediment con-
centrations decreased on a per unit basis for precip-
itation-related transport during P1. Since sediment
concentrations decreased in the stream during P1, a
similar decrease was  expected and was measured
within the lake.
  The TSS concentration decreases within the stream
are assumed to  be related to the implementation of
RMS's. Therefore it is assumed the decreases in TSS
concentrations within  the lake are related to RMS's
implementation within the watershed.
  The extent to which other factors such as lake turn-
over affect water quality is not fully known. For exam-
ple, during late 1982 a significant increase in TSS con-
centrations unrelated  to runoff and  apparently  at-
tributable to mixing  resulting  from  destratification
(fall   overturn)  was   measured.  The lake  will  be
monitored for a number of years to document the full
impact of RMS implementation and other factors such
as lake turnover  on water quality.
  A detailed discussion by parameter sampled is pro-
vided for the lake and watershed monitoring sites in
the Phase III report (Davenport,  1983).


Project Conclusions

1. The institutional working arrangements employed
by the involved  agencies and  organizations at the
local level were the key to the success of reaching the
project's soil reduction goal.
  2. Potential gross erosion for the Blue Creek Water-
shed Basin is estimated to be 63,313 tons/year. Sheet
and  rill erosion  contribute 62,867 tons. Gully and
streambank contribute 446 tons. Bottom channel ero-
sion is not  included in this estimate.
  3. The gross erosion rate from the watershed was
compared with  lake sedimentation data from Pitts-
field Lake to determine the sediment delivery ratio.
The sediment ratio from 1961-79 was calculated to be
66 percent. The sediment delivery ratio for the period
of 1974 to 1979 is 38 percent.
  4. Currently (1974-9), 0.64 percent per year of lake
capacity is lost to sedimentation. Theoretically, under
these conditions the lake will  be completely  filled in
92 years. It should be noted that the lake would never
completely fill up and that the useful life of the lake is
significantly less than the estimated time it would
take to fill it up.
  5. Any reduction in  soil erosion resulting from im-
proved land management practices would lead to a
reduction in material delivered to the  lake.
  6.  Lake sedimentation survey indicates the finer
material (silt, clay) is remaining in suspension, there-
fore practices which control the soil erosion detach-
ment and transport mechanisms for the finer particles
should be  recommended to  land  owner/operators
within the Blue Creek Watershed. Conservation tillage
and contour farming are two such practices.
  7.  The extent  of variation in sediment concentra-
tions  from  time period to time period, illustrates the
need for long-term studies to document the impact of
RMS's on loading of nutrients and sediments.
                                                 572

-------
  8. Lake water quality reflected the interaction  of
land mangement and precipitation. Lake water quality
was at its poorest during the P1 management period
and excess rainfall. Lake water quality was the best
during P3 with no rainfall.
  9.  An  overall   improvement  was  measured   by
decreased suspended solids concentration from 1979
to 1982.

ACKNOWLEDGEMENTS: This project was financed  in part
with funds from the U.S. Environmental Protection Agency,
Region V, Chicago, III., under Section 208 of the Clean Water
Act (P.L 95-217). The contents of this report do not necessari-
ly reflect the views of the U.S. EPA. John Little, Bill Ettinger,
Jill Hardin and Rodney Mutz of Illinois Environmental Protec-
tion Agency contributed significantly to this project. The Il-
linois State Water Survey  was contracted to perform com-
ponents of the Blue Creek Watershed Monitoring and Evalua-
                             WATERSHED MANAGEMENT


tion Project; Dr. Ming Lee's assistance and contributions
were greatly appreciated.
REFERENCES

Davenport, T.E.  1981. Blue Creek Watershed Project, May
  1979-October 1980. Plann. Section Div. Water Pollut. Con-
  trol, III. Environ. Prot. Agency, Springfield.
	.  1983. Water resource data and preliminary trend
  analysis for the Blue Creek Watershed project, Pike Coun-
  ty, III.  Phase III IEPA/WPC/83-004.  Plann. Section, Div.
  Water  Pollut.  Control,  III.  Environ.  Prot. Agency,  Spring-
  field.

Sefton, D.F., M.H. Kelly, and M. Meyer. 1980. Limnology of
  63 Illinois  Lakes, 1979. Monitor Unit, Plann. Section, Div.
  Water Pollut.  Control,  III.  Environ.  Prot. Agency,  Spring-
  field.
                                                     573

-------
                               Sediment  Problems  &
                       Management Techniques
CAN A MICROCOMPUTER HELP THE MANAGER OF A MULTIPURPOSE
RESERVOIR? THE EXPERIENCE OF LAKE COMO
G. GUARISO
S. RINALDI
R. SONCINI-SESSA
Politecnico di Milano
Milano, Italy


          ABSTRACT

          A synthesis of a 5-year study on the efficient regulation of Lake Como in Northern Italy is
          presented. Its main result was the proposal of a new operating rule for the lake, which has been
          implemented on a microcomputer. This has been operated daily for 1 year and a first evaluation
          of the impact of this innovation on the lake manager's attitude is briefly outlined.
INTRODUCTION

A synthesis of a 5-year research effort on the efficient
regulation of  Lake como in Northern Italy and the ex-
perience gained in the first years of implementation of
the results are  reported in this paper. The research
was jointly supported by the Italian National Research
Council, the  International Institute for Applied Sys-
tems  Analysis (Guariso,  1982a),  and the  Consorizio
dell'Adda,  the agency responsible for  the manage-
ment  of Lake Como dam. Intermediate results have
already  been  presented  by  the  authors  (see
references).
  The main purposes of the study were to:
  1. Analyze  the impact of recent different structural
and institutional proposals to alleviate flood damages
on the lake shores (particularly in Como town, where
the commercial center has been progressively sinking
since the sixties, probably because of overpumping
from the underground aquifer);
  2. Test the possibility of designing a computerized
system, which  could  be really  used  by the lake
manager in his daily release decision. One can, in fact,
notice that the  great majority of the studies of this
type presented in the literature, failed to generate real
decisionmaking  systems since the final solution they
suggested was never implemented. This fact can
often be imputed to the  opposition  of the lake
manager, who considered the proposed solution unac-
ceptable  for practical implementation.  Thus,  in the
study described here, the manager's experience has
been strongly emphasized and an optimal operating
rule has been searched "not far" from the one he was
actually using. This condition proved to be really help-
ful in overcoming the manager's psychological diffi-
culties and led to the final implementation and daily
operation of the system.

LAKE COMO  MANAGEMENT

Lake Como receives water from a catchment of 4,508
km2, at a mean elevation of 1,500 m in the central part
of the Alps. The  outflow rate of the lake can be varied
from day to day  by operating the dam built at the end
of World War II. The inflow rate averages 160 m3/sec
and has the typical annual pattern of alpine rivers with
two peak flow periods, one in early summer from snow
melt, and another in autumn from rainfall.
  Water from Lake Como supplies a group of down-
stream users before reaching the Po river, some 140
                                        575

-------
LAKE AND RESERVOIR MANAGEMENT
 km south of the lake. More precisely, six agricultural
 districts and  seven run-of-river  hydroelectric power
 plants are located along the course of the Adda river
 and are served by a complex of canals. The production
 functions of all these users are not well-known  and
 economic data on agriculture are scarce and quite
 unreliable. It was thus agreed  with the manager to
 characterize the performance of the lake operation by
 using simple  physical indicators affecting the costs
 and benefits of all parties involved. In particular, the
 objectives of the downstream users were assumed to
 be  the minimization of the expected value A of the
 total  annual  deficit  in the  agricultural  sector
 (evaluated with respect to the nominal requiremenls)
 expressed in 106m3, and the minimization of the ex-
 pected value E of the annual hydroelectric power loss
 (evaluated with the respect to the installed capacity)
 in GWh. The objective of Como  town was considered
 to be the minimization of the expected  number F of
 days of flood per year, (which has been about 10.2 in
 the last 15 years), an  indicator commonly used to
 represent indirect damages,  but which proved, in  this
 case, to be highly correlated with all the other main
 aspects of the flood events (Zielinski et al. 1981).


IMPROVEMENT OF THE OPERATING RULE

The license act, issued by the Ministry of Public Works
at the beginning of the regulation, specifies that  the
manager can freely decide the release rt of each day t,
whenever the lake level x, at the beginning of the day
falls in a specified interval (x, x), called control range.
This control range, measured relative to the elevation
of the Fortilizio hydrometer,  was, until very recently,
 -0.50  m to 1.20  m,  which corresponds  to an active
storage of about 250 million cubic meters.
  When the level  xt of the lake is outside the control
range the manager must follow some pre-specified
rules. In particular, when xt >x the manager must pro-
gressively open all the gates of the dam in order to dis-
charge as much as possible, thus avoiding too much
flooding on the lake shores. Furthermore,  the level
cannot decrease below x in order to meet navigation
and sanitary requirements. In other words, outside the
control range there is no freedom for the manager in
making the decision.
  The manager's behavior in the control range was
not  specified in the license act and thus it was mainly
based on  past experience. The  help of the manager
himself led to the following formal interpretation (see
Garofalo et al. 1980) which proved  to be quite sat s-
factory. The release r, of each day t of the year mainly
depends upon the amount of available resource,  I.3.,
rt =  r(xt,  t), and this  operating  rule  is periodic with
respect  to t because of the  yearly periodicity of in-
flows and water requirements. In a particular day t, the
shape of this rule can be represented as in Figure 1,
where xt. and  rt* represent the schedule of levels and
releases followed by the manager in mean hydrologi-
cal  conditions (rule curve), wt is the agricultural de-
mand and the slopes (at, ft) can be considered as
representing the manager's sensitivity to deviations of
the  level from  standard  conditions. For instance, the
values of ft are  particularly high during traditional
flood  periods.  A computer  simulation of the lake,
using this operating rule, replicated very well what
really happened between 1965 and 1978.
  The approach used to improve the operating rule
(see Zielinski et al. 1981; Guariso et al. 1981a, 1982a)
was mainly aimed at preserving as  much as possible
the  basic structure of the manager's operating rule, in
 such a way that the proposed changes could be easily
 accepted  and  adopted  by him. According  to the
 manager's recommendations,  only  the  values  of at
 and ft were considered to be modifiable, provided that
 their seasonal variations were  preserved. Consistent-
 ly, the operating rules from which the optimum was
 searched are still of the type shown in Figure 1, but
 with slopes proportional, through two unknown cons-
 tant parameters a and b to be determined through op-
 timization, to  the  values of at and ft used  by the
 manager.
   Consistently, the stochastic multiobjective program
 that  allows  the  determination of the  efficient
 operating  rules (see, for example, Cohon and Marks,
 1975), can be formulated in the following way
     min [A E F]
                                  (D
 (previously defined), subject to a set of mass balance
 equations  representing   the  network  of canals
 downstream of the  lake  and the actual  rules  of
 distribution among the users, and to the continuity
 equation
          =x, + a, - r(x,,t,a,b)   t = 1,2,...
                                  (2)
 where  the  inflow a,  is  a 1-year cyclostationary
 stochastic  process and  r(xt,t,a,b)  is the  family  of
 operating rules considered as candidates for optimali-
 ty. As is well known, the solution of this problem is not
 unique,  but is represented by  a set  of  efficient
 operating rules. Each one of them  is identified by a
 particular pair (ao,b°) of parameters and  has the pro-
 perty that any variation of such parameters weakens
 at least one of the three objectives of the problem.
   This set of efficient operating rules (i.e., the set of
 pairs (a°,b°) has been determined by simulating the
 daily behavior of  the system  in  a series of years
 (1965-1979) for various pairs (a, b) of the parameters,
 thus estimating the corresponding values of the objec-
 tives. The efficient solutions can be represented by a
 set in the two-dimensional space of the parameters (a,
 b) or by a surface in the three-dimensional space (A, E,
 F) of the objectives.  In  Figure 2,  such a surface is
 represented by the contour lines E = constant. In the
 same  figure,  point  H  represents the  "historical
UJ
LU
cr
>

<
o
open gates
stage-discharge    /
function        /
                                    reference level
                                       and release

control range
i I

x xt x* x x,
LAKE LEVEL
Figure 1.—The operating rule of Lake Como.
                                                 576

-------
                                                            SEDIMENT PROBLEMS & MANAGEMENT TECHNIQUES
values" of the objectives, nameiy the values that
would have been obtained under nominal conditions
by systematically  applying the operating rule with
a = b = 1 (i.e., the manager's previous operating rule);
while the Utopia point U represents the (independent,
and hence infeasible) absolute minima of A and F.
This showed that consistent improvements of  the
management performance were possible. Selecting,
for example, the operating rule corresponding to point
P in Figure 2, maintains the hydropower deficit at its
historical  value and  decreases the mean number of
days of flood from 10.2 to 6.3 and the average agri-
cultural deficit from 201 to  about 98 million cubic
meters.
  The  behavior of  the lake when using the operating
rule corresponding to point P in Figure 2 was carefully
inspected (point P on the segment HU, has the proper-
ty of sharing the surplus of benefit in equal propor-
tions to agricultural users and lake shore inhabitants,
without affecting power production which is anyway
the marginal user in the system). Comparing the peak
and duration of each flood and the peak, duration, and
volume of each deficit episode that would have occur-
red  in  the  period  1946-1979 with  the proposed
operating rule with those generated  by the  previous
rule  showed clear advantages in  using the first  one
(see Guariso et al. 1982b).
THE IMPACT OF OTHER ACTIONS

In this section we will analyze the impact of other pro-
posals which  were raised  in  the same period to
alleviate flood problems. Since hydroelectric  power
production turned out to be relatively insensitive to
reservoir operation (see Figure 2), the analysis  in this
section will disregard this objective. Thus only the
results relative to  an  average yearly hydroelectric
power deficit of 200 GWh will be shown.
   The first  proposal, raised by the municipality of
Como, was a change in  the institutional framework of
the management, consisting of a reduction of the up-
per limit x of the control range defined in the license
act.This proposal, which is particularly attractive be-
cause it can be simply realized by a formal revision of
that document, can  be analyzed by solving the multi-
objective problem of Section 3 for different values of x.
In this way a set of efficient solutions is generated for
F
[days]
                                    H (historical)
      40      80      120     160     200    AHO6 m3]

      MEAN VOLUME OF WATER DEPICTS IN AGRICULTURE


Figure 2.—Efficient solutions in the space of the objectives.
each value of x. The study showed (Guariso  et al.
1981 b) that the sensitivity of floods to a reduction of x
was only  moderate,  but, on the other hand, the in-
crease in agricultural deficits induced by this change
was negligible. As a consequence,  the Ministry  of
Public Works reduced 5c to 0.90 m in June 1982.
  Another proposal,  advanced  by the downstream
users, was to partially restore the pre-sinking condi-
tions of Como center, through  structural works  to
heighten or protect it. Again, this problem was dealt
with  by  solving  the  multiobjective problem  of
parameterizing the value xc of the lake level at which
Como center  is flooded. This means that a set of effi-
cient operating rules  and the corresponding values of
the objectives are computed for each  value of xc. This
study showed (Guariso et al. 1982c) that even an eleva-
tion of only 20 cm of the central main square would en-
tail consistent benefits, reducing to less than half the
number of days of flood for all  agricultural deficits
higher than 80 million cubic meters per year. On the
other hand, if sinking should continue, the number of
days of flood  may increase dramatically so that struc-
tural protection becomes almost mandatory. Indeed,
after presentation of these results to the municipality
of Como,  the elevation of  the  sunken part  of the
square was  immediately considered and  should be
completed very soon.
  The third and final proposal  was  to implement  a
hydrometeorological  data collection  network to sup-
port the manager's decision with information on the
values of three variables characterizing the status of
the catchment: snow cover, depth of  the aquifer, and
rainfall over the catchment. The  possibility of a real-
time use of this information, which was known in the
past only at the end of each month, was thus analyzed
(see Guariso et al. 1982d).
   According  to the manager, the main effect of these
variables was similar to a change in the pattern xf and
rt* of reference levels and releases. This can be easily
understood, if one thinks, for example, of snow cover.
Whenever in  fact, snow cover indicates an abundant
availability of water, the level xt* can be  decreased,
meaning  that less storage is required in  the  lake if
more resource is known to be available in the catch-
ment. On the contrary, if it  indicates that less water
than usual is present upstream, the  release rt*  has  to
be reduced in order to store more water than in stan-
dard  conditions and  still  satisfy  the  water  re-
quirements during the following dry season. The ef-
fects of aquifer depth and rainfall are similar, except
for the fact that the last one has obviously no meaning
for future deficit conditions. A variation of  xj or rt* pro-
portional to the differences between the actual values
of  the three  hydrometeorological variables and their
mean  values (or a threshold value for rainfall), was
thus imposed, and the daily release became a func-
tion r(xt, t, yt, p) also  of the three hydrometeorological
variables  represented by the vector yt, and of a vector
p of parameters (two for each new variable). Again the
management policy r(xt, t, yt, p) can be inserted in the
continuity equation of a multiobjective stochastic pro-
gramming problem analogous to the one previously
presented and efficient management policies can be
determined.
   The results are presented in Figure 3 in the plane (A,
F). The three  curves denoted by (1), (2) and (3) represent
the efficient  solutions that can be  reached by con-
sidering only one hydrometeo variable at a time. The
improvements obtained with respect to point P are not
irrelevant, particularly if compared with point U. When
all the information   is used, the  set of solutions
                                                 577

-------
LAKE AND RESERVOIR MANAGEMENT


Fldays]
    6-
 Q
 O
 O
    5-
  <
  O
                                 10 use of extra-mformatior}
 [2] (aquifer depth)
 [1] (snow cover)


[3] (rainfall)


[1,2,3] (all data)
       ,U (utopia)
  <      50       70       90       110   —130 A[106m3]
  5     MEAN VOLUME OF WATER DEFICITS IN AGRICULTURE

Figure 3.—Efficient solutions with direct use of hydromete-
orological information.


denoted by (1, 2, 3) is obtained. This shows that the
use of  hydrometeorological  information can  con-
siderably improve the management of the lake. For ex-
ample, point X on the (1,2,3) efficient set, represents a
20  percent reduction of agricultural deficits and a 55
percent reduction of floods with respect to the  max-
imum possible improvements. On the basis  of these
results the Consorzio dell'Adda is presently installing
an  automatic  raingauge network.
 IMPLEMENTATION ON A MICROCOMPUTER

 The practical implementation of the operating rule
 corresponding to point P of Figure 2, would simply
 correspond,  under  normal   circumstances,  to  the
 reading of the release rt corresponding to day t and
 level  x, on a table where the values rt = r(xt, t) cor-
 responding to all possible pairs (xt, t) have been pie-
 computed. However,  real world  circumstances may
 differ from theoretical ones. It is therefore compulsory
 to give clear  suggestions on how they must be faced.
 For instance, the actual water requirement wt can be
 different from the nominal  value considered in  the
 study, in which case, the operating rule (see Fig. 1) has
 obviously to  be used with the actual value. Moreover,
 as  shown, the operating rule has to be changed  as
 soon  as the elevation of the square of Como is com-
 pleted or if the active storage is revised again. Finally,
 the presentation of the operating rule by  means of a
 table  becomes practically impossible if all the hydro-
 meteorological information becomes available.
  For these reasons, the operating rule corresponding
 to point P of figure  2, as well as  all its  considered
 modifications, was implemented on a very cheap per-
 sonal computer based on a Z80 microprocessor. With
 its  64 K bytes memory, two floppy-disk drivers for a
 total of 800 K bytes and a 1,520 characters screen, this
 computer constitutes a fast  and easy tool  to deter-
 mine each day the suggested release. It has also bean
 equipped with other dedicated software for 1 and 3
 days  inflow forecasting, basic statistics, lake simula-
 tion, data collection,  and printing of summary tables.
  In practice, each morning the manager types  the
 date,  the  lake level, and the actual  agricultural  de-
 mand and gets  the proposed  release.  If  he has
 hydrometeo  information on  the  catchment  (some-
 times snow cover is  available through the National
 Power Agency, and  some other data through other
 private agencies or municipalities), he has the optian
 of  using the management  policy. When willing  to
 check the effects of a certain rule on a longer term, he
 may use the program  for inflow forecast and then
 simulate the lake behavior for 2 or 3 days.'ln the mean-
 time he has statistics on inflows, levels, and releases
 in order to verify how far they are from standard condi-
 tions at that date.  He may also check the pattern of
 any hydrometeorological  variable including inflows,
 levels, and releases by looking at the appropriate sum-
 mary table, which also  provides current statistics for
 each 10 day period and each month. In the present
 structure, 10 years of daily data can be stored on a in-
 expensive microfloppy.


CONCLUDING REMARKS

  A table representing the operating rule was given to
the manager at the beginning of 1981, while the imple-
mentation described in the previous section was com-
pleted at the end of the  year.
  Table  1 shows  the  performances  of  the  new
operating rule versus the actual ones in the years
1980-1982, which were  not  previously used in this
study. None of these  years presents  particularly
severe drought situations, while there have been few
remarkable flood episodes. The performances of the
new rule confirm this fact, by exhibiting an average
agricultural deficit,  A, strongly lower than the mean
value of the preceding years, and a slightly reduced
number of days of flood  F.  On the contrary, the perfor-
mances obtained by the manager in 1980 and 1981
correspond quite closely to the average values of the
preceding years.  From this emerges that the manager
was  not  fully using the table  of  the proposed
operating rule,  which  would have  given him  the
possibility of avoiding the 6-day flood at the end of
May 1981. Nevertheless,  in the second half of the year
the pattern of levels was closer to the one that would
have been obtained using the new rule. The introduc-
tion  of  the  microcomputer, which  allowed   the
manager to deeply explore the effects of alternative
decisions, pushed him to follow more closely the new
operating rule. The third column of Table 1 shows in
fact that the number of  days  of flood actually occur-
ring almost equals those of the new operating rule and
the reduction of the agricultural deficit with respect to
values of the preceding years is stronger than the cor-
responding  reduction  obtainable with  the   new
operating rule. However,  the absolute value of the agri-
cultural deficit reveals that the suggested releases are
not yet fully accepted.
  Several reasons  may explain this  fact. First,  in
some cases the manager turned out to be more risk-
adverse than the proposed operating rule. In fact, it
seems that at very low lake levels, the primary objec-
tive of the regulation becomes the avoidance of high
peaks of deficit,  so that the manager always prefers
not to fully exploit the  lake capacity. However, a com-
parison of past peak deficits with those of the propos-
ed  rule  proved that this attitude is very often  un-
justified. Second, real-world circumstances were often
different from those assumed in the study. The shut-
                 Table 1.— Manager's and new operating rule performances in
                                  the years 1980-82.
                 ^  New rule
                 „  Manager

                     New rule
                 u.
                     Manager
1980
53
191
4
1C)
1981
48
217
8
15
1982
1
113
3
4
                                                 578

-------
down of a turbine for maintenance or a repair to the in-
take of a downstream channel may have induced the
manager to  significantly deviate from the suggested
release.
   Finally, though we  have strongly emphasized the
differences between the manager's behavior and the
proposed operating  rule, we want to stress that the
purpose of this research cannot be the substitution of
the manager with an automatic procedure, but to give
him a cheap and easy tool to have a systematic in-
sight  over his options and  to help him  in  better
evaluating the impact of his choice.
REFERENCES
Cohon, J.L., and D.H. Marks. 1975. A review and evaluation of
  multi-objective  programming  techniques.  Water Resou.
  Res. 11 (2): 208-19.
Garofalo, F., U. Raffa, R. Soncini-Sessa. 1980.  Identificazione
  della politica di gestione del lago di Como.  Proc. XVII Con-
  vegno di Idraulica  e Costruzioni  Idrauliche, B9, Palermo,

Guariso, G.S. Rinaldi, R. Soncini-Sessa. 1981a. La regolazione
  ottimale  del  lago di  Como: analisi  a  molti  obiettivi.
  L'Energia Elettrica, 58(7): 281-6.
         SEDIMENT PROBLEMS & MANAGEMENT TECHNIQUES


	.  1981b. La regolazione ottimale del lago di Como:
  conseguenze di una riduzione dell'invaso. L'Energia Elet-
  trica 58(9): 363-8.
	1982a. The management of Lake Como. Work. Pap.
  WP-82-130. Int.  Inst. Applied  Sys.  Analysis,  Laxenburg,
  Austria.

	.  1982b. Analisi dell'affidabilita di una proposta di
  regolazione del  Lago di Como. Proc. XVIII Congresso di
  Idraulica e Costruzioni  Idrauliche, Bologna, Sept.

	1982c. La regolazione del lago di Como: effetti della
  subsidenza della Piazza Cavour. L'Energia Elettrica, 59(2):
  61-6.

Guariso G., S. Rinaldi and P. Zielinski. 1982d.The value of the
  information in  reservoir  management. Work.  Pap.
  WP-82-129.  Int.  Inst. Applied  Sys.  Analysis, Luxenburg,
  Austria.

Zielinski P., G. Guariso and S. Rinaldi. 1981. A heuristic ap-
  proach for  improving reservoir management: application
  to Lake Como. Proc. Int. Symp. on Real-time Operation of
  Hydrosystems, Waterloo, Ontario, Canada, June 24-26.
                                                    579

-------
 VANCOUVER LAKE:  DREDGED MATERIAL DISPOSAL AND RETURN
 FLOW MANAGEMENT IN A  LARGE LAKE DREDGING PROJECT
 RICHARD RAYMOND
 FRED  COOPER
 Cooper Consultants,  Inc.
 Portland, Oregon


             ABSTRACT

             The restoration of Vancouver Lake required the dredging of 6.5 x io6 m:i of material from the Lake,
             the construction of 17 km of land based retaining dikes to enclose 180 ha of disposal area, and the
             disposal of nearly 3 x 1 o6 m3 of material in the lake to form an island. The requirement that all dredge
             return flow be returned to the lake necessitated careful control of dredging activity and the imposition
             of several design and operation features to CDntrol the quality of the return flow water. Some of the
             measures used included multiple, or settling basins, extended wier length to reduce crest height, silt
             curtain enclosures around dredge disposal site outfalls, rapid alteration of dredge disposal sites, and
             careful monitoring of dredging activity and return flow quality. These measures enabled the project
             to be completed with minimum delay, ahead of schedule, and with no serious violation of water quali-
             ty standards. Observations and data on water quality conditions during construction and the efficacy
             of specific dredging and sediment containment methods will be described.
 INTRODUCTION

 Vancouver  Lake is  located in the Columbia River
 floodplain, adjacent to the city of Vancouver in south-
 western Clark County, Wash., within the greater Port-
 land, Ore., metropolitan area (Fig. 1). The predominant
 land use adjacent to the lake is agriculture, although
 the south and west shorelines are included in a count/
 park. Industrial activity related to the Port of Van-
 couver occurs south of the lake and includes a large
   VICINITY MAP
Figure 1.—Vicinity map of Vancouver Lake.
aluminum smelter. The primary residential use close
to the lake is in conjunction with farming. Additional
residential areas are located on lowlands southeast of
the lake and along the top of the east shore bluff.
  The low lying lands to the north, west and south are
subject to seasonal flooding from the Columbia River
which flows  within  1.6 km of the southwest shore of
the lake. These lowlands have an elevation of from 3
to 6 meters above mean sea level (msl). The northeast
shore of the lake is formed by bluffs rising  to an
average elevation of 60 meters msl.
  The climate of the region is maritime Mediterranean
with moderately warm, dry summers and mild, wet
winters. Seventy-five percent of the annual precipita-
tion occurs between October and March. Annual total
precipitation is approximately 100 cm.
  Vancouver Lake has a surface area of 1,100 ha, is 4
km across from east to west, and has a mean shore-
line length of about 12 km. The depth varies seasonal-
ly,  ranging from a depth  of less  than  1 meter in
September and October (maximum  depths as low as
0.6 m have been recorded) to about  4 m in early June.
Prior to restoration the lake had a virtually flat bottom
except for higher areas at the northern end caused by
sedimentation of materials  carried into  the lake by
Lake River (Fig. 2). These shallower areas  are exposed
during lowest waiter.
Hydrology

The hydraulic regime of Vancouver Lake is complex
and involves Burnt Bridge Creek, Lake River, Salmon
Creek, and the Columbia River (Fig. 3).
  Burnt Bridge Creek flows east through commercial
and suburban sections  of Vancouver and drains an
area of approximately 70 square km. Mean  annual
flow is about 0.6 mis, but flows are quite variable. Con-
siderably higher  flows are observed during  rainfall
events. Burnt Bridge Creek flows are usually between
0.1 and 0.3 m/s (Dames and Moore, 1977b).
                                               580

-------
                                                            SEDIMENT PROBLEMS & MANAGEMENT TECHNIQUES
  Lake River joins Vancouver Lake at its northern ex-
tremity and connects to the Columbia River approx-
imately 20 km to the north. The volume and direction
of flow in Lake River is directly related to seasonal and
tidal changes in the stage of the Columbia River. Dur-
ing April  and  May, while the Columbia River is rising
due to spring runoff, Lake River flows south into Van-
couver Lake with flows that  reach 5.7 cu m/s. In late
June the stage of the Columbia River drops rapidly
and Lake River reverses to flow north out of Vancouver
Lake. Flows during this period may reach 4.25 cu m/s.
  During much of the year, August through March, the
flow in Lake River is variable. It may reverse daily in
response to tidal changes in the Columbia River, or it
may flow north or south for several days at a time be-
cause of longer-term changes in the Columbia River
resulting from weather or power generation.
  Salmon Creek is a tributary of Lake River which
drains a large rural and agricultural area north of Van-
couver. It empties into Lake River about 3 km north of
Vancouver Lake. Water from Salmon Creek can enter
Vancouver Lake during southward flow of Lake River.
  The Columbia River has  an  annual mean flow at
Vancouver of 5,714 cu m/s. Flow is distinctly seasonal
with low  flows in the fall as low as 1,398 cu m/s and
flows during the spring snow melt as high as 18,400 cu
m/s.
  Because of the great variation in flow, there is con-
siderable seasonal change in the level of the Colum-
bia River at Vancouver. Mean maximum river stage is
greater than 4.6 meters msl, while during August river
stage may fall as low as 0.6 meters msl.
   During low water in late summer, there is a daily
tidal fluctuation in the Columbia River stage near Van-
couver of approximately 0.6 meters. The hydrology of
Vancouver  Lake is directly controlled by the stage of
the Columbia River. The lake is at  its lowest level in
                           Station 1
                                        Mud Flats
late October. As the Columbia River and flow in Burnt
Bridge Creek begin to rise from the winter rains, the
lake level  rises accordingly to an intermediate level
with perhaps a mid-winter peak following particularly
heavy  rain. Flow in  Burnt Bridge Creek  increases
rapidly as the rainy season begins. Between  Novem-
ber and February the flow  in Lake River is frequently
reversed and may flow into Vancouver Lake for several
days at a time. The effects of tidal fluctuations are
reduced as the river stage  rises.
  In late spring (April-May) the  Columbia River rises
in response to snowmelt runoff. During this time, until
the highest water in  mid-June, flow in Lake  River is
south into Vancouver Lake. The volume of water in the
lake may increase fourfold during November to June
(8.5 to 35 million cubic m).  From mid-June to mid-July
the lake level drops  rapidly  from its high mark  of
around 3.7 meters msl to near the annual  low of 1.0
meters msl. During this time, the flow in Lake River is
north carrying water away  from the Lake.
  During the period of July to December the lake re-
mains at a low level, fluctuating around 1 to 1.7 meters
msl  elevation.  While the  flow in Lake  River  may
reverse daily in  response to Columbia River tidal fluc-
tutation, it is not clear whether water from the Colum-
bia River  is actually  reaching the Lake during this
period.
  Prior to restoration, Vancouver Lake was extremely
shallow during  the summer, and  suffered  from high
concentrations of algal nutrients (nitrogen and phos-
phorus) and coliform bacteria.  Use of the lake was
restricted by the shallow  depths  and the  low  trans-
parency resulting from high algal  densities and wind
resuspension of bottom sediment.
  To restore the lake, a three-faceted plan was devel-
oped. This plan proposed to introduce water  of lower
nutrient content from the nearby Columbia River via a
3.6 km long excavated channel; to dredge the lake to
promote the circulation of Columbia  River water and
restrict the circulation of  nutrient laden  water from
                                                                                 VANCOUVER LAKE
                                                                                     COMPLEX
                                                    Columbia River
                                                           \
                                    Salmon Creek
                                         /
                                                                                      Burnt Bridge Creek
                                                                                            /
                       ^

 Figure 2.—Vancouver Lake prior to restoration.               Figure 3.—Vancouver Lake and Associated Waterways.
                                                 581

-------
 LAKE AND RESERVOIR MANAGEMENT
 Burnt Bridge Creek; and to impose controls in  tha
 Burnt Bridge Creek watershed to reduce the nutrient
 load to the lake.
 PROJECT DESIGN

 Flushing Channel

 The flushing channel is an open channel, 30 m wide at
 the bottom with side slopes of 3:1 and bottom eleva-
 tion of -2.5 m msl. At the west end it is open to the
 Columbia River and on the east it terminates in two
 2.13 m diameter reinforced concrete culverts. Near the
 lake end of the culverts is a tide box containing flap
 gates to prevent backflow from the lake into the river,
 and sluice gates which can close the culverts to pre>-
 vent all flow. The flushing channel design was chosen
 to obtain a minimum flow of 8.5 cu m/s and a dilution
 ratio (percent per day replacement) of approximately 5
 percent  during low flow conditions  (Dames & Moore',
 1980).

 Dredging Plan

 The dredging plan developed by Dames and Moon?
 (1980) had five  major components:
   1. A channel approximately 300 m wide dredged to
 -0.3 m  msl along the west side of the lake.
   2. A channel approximately 300 m to 950 m wide
 dredged to - 1.2m msl along the east side of the lake.
   3. A channel approximately 300 m wide dredged to
 -0.3 m  msl along the south side of the lake.
   4. Sediment traps in the lake dredged to -2.5 m
 msl and -1.5 m msl  at  the discharge  from the
 flushing channel and the entry to Lake River, respec-
 tively.
   5. Extension of the east channel at a width of 150 m
 dredged to -6.5 m msl into the lake area at the north
 of Burnt Bridge Creek.
   These  components were deemed the  minimum
 necessary to accomplish adequate  water quality im-
 provement in the lake. This dredging plan required the
 removal  of approximately 6.5 million cubic meters of
 sediment from  the lake bottom.


 Disposal Sites

 To dispose of the sediment removed from the lake bot-
 tom channel, eight disposal  sites near the lake were
 identified. Six of these sites were on land, and two
 were in the lake itself. Factors considered in selecting
 dredging sites included the proximity of the disposal
 area to the dredging site, the suitability of the dredged
 material  for  the intended future use of the site,
 minimizing or mitigating loss  or damage to wildlife
 habitat, suitability of on-site material for dike con-
 struction, and potential damage or disturbance to ex-
 isting structures or archeological sites.
   Because of the anticipated  increase in volume ol
the sediments due to disturbance during dredging, the
total volume of disposal site was greater than the total
volume of dredged material. The design bulking factor
was 1.3 based on a pilot dredging study  (Dames and
 Moore,  1977a).  Approximately  7.6 million cubic
yards/meters of disposal area  were available for the
project. Dredging and disposal areas are shown in
Figure 4.
   In  the design phases, the in-water disposal areas
were to be contained by silt curtains, rather than by in-
water dikes (Fig. 5). During the  project, the contractor
determined that the design method of island construc-
 tion would not meet his needs and so elected to build
 in-water dikes to contain the dredged material to be
 disposed on the inlake disposal areas.

 Water Quality Requirements

 All the return flow from dredging disposal areas had to
 be returned to the lake. This requirement necessitated
 careful consideration of water quality. The focus was
 on three main areas: (1) the quality of the return flow
 leaving a retention facility; (2) the extent of the allow-
 able dilution zone at the point of return to the lake; (3)
 the impact of the return flow on the lake as a whole.
    The purpose of the dilution zone was to allow an
 area of transition between the point of return flow and
 the remainder of the lake (Fig. 6). Different water quali-
 ty requirements were set for each of the three stages
 of return flow: the disposal area discharge, the dilu-
 tion zone, and the lake as a whole.
   The major consideration for establishment of water
 quality requirements was to strike a balance between
 what could be reasonably achieved at the dredge sites
 and what would be likely to result in adverse  impact
 on the fishery resource in the  lake. The standards
 established for each area in Table  1. While the only
 water quality  parameters specified which had an  in-
 fluence on the conduct of the dredging operation were
    DREDGING &
    DISPOSAL AREAS
 Figure 4.—Dredging and disposal areas for the Vancouver
 Lake Restoration Project.
                           N NX—200'
                            A

                | NORTHWEST   \\_.	SILT CURTAIN,
                                       LENGTH - 7,700'
                                   TOE OF  ISLAND FILL
               NO SCALE

Figure 5.—Diagram of island construction.
                                                 582

-------
                                                             SEDIMENT PROBLEMS & MANAGEMENT TECHNIQUES
dissolved oxygen and suspended solids,  many more
parameters were measured during the restoration pro-
ject. These are listed in Table 2. Water samples were
taken  biweekly  before dredging  began, daily during
the  dredging, and weekly  after  the completion of
dredging  in  March 1983. Post-project monitoring is
planned to continue through the summer of 1984.

Water Quality Control Measures

Several measures  were used to attain  compliance
with the water quality criteria established for the con-
struction  activity.  Provision  was  made to route the
return flow through several cells before it was return-
ed to the lake in order to achieve a longer settling time
for the dredge spoils. In addition, weir crest lengths in
the cells were designed so that the crest of return flow
over the weir would not exceed 5 cm. To provide more
settling time, especially as the spoils areas began to
be filled,  silt  curtain  enclosures  were erected  in the
lake surrounding the dredge return flow outfall. Final-
ly, activities were monitored  daily to ensure the early
detection of any problem to permit remedial action to
be taken.
  The combination of multiple retention ponds  and
silt curtains  was used to provide a design  retention
Diagram of Dilution Zones


     CASE I  LS1000 FT
                             CASE H : L>1000 FT
     CASE HI : L<1000 FT
                             CASE H : L>1000 FT



                        SILT CURTAIN
        D  IS THE LOCATION OF THE DREDGE


        0  IS THE LOCATION OF THE OUTFALL


        r  IS EQUAL TO 500 FT


        L  IS DISTANCE BETWEEN THE DREDGE AND RETURN FLOWS ENTRY POINT
           TO THE LAKE
time of approximately 8 days. This design criterion
was met by all of the disposal sites and substantially
exceeded by some, during the initial stages of filling.
During  the later  stages of  filling  several of  the
disposal sites fell below this design criterion. During
this period the use of a silt curtain around the outfall
served to provide the necessary retention time for pro-
per settling of the dredged material.
  Within the land disposal areas, the design criteria
were to provide a maximum crest height over the weir
of 5 cm while maintaining a ponding depth in the site
of 1 m.

RESULTS DURING CONSTRUCTION

Initial Problems

All of the design parameters for the dredged material
disposal areas,  water quality control measures, and
retention  time  calculations  were  based  on  the
assumption of  a 20-inch  (50  cm)  hydraulic dredge
operating 18 hours per day for six days per week. The
contractor chose  to do the work using a 26-inch (66
cm) hydraulic dredge operating 24 hours per day, five
days per week. This resulted in more than a threefold
increase in the  average amount of material per day
pumped into the disposal areas (Table 3).
   While this change in equipment put a greater strain
on the water quality control measures, it also signifi-
cantly reduced the total time for the dredging thus per-
haps reducing the overall impact of the project. Table
4 summarizes the operating conditions of the equip-
ment used in dredging Vancouver Lake.

Weirs

The initial dredging demonstrated that changes were
required. The weir length as designed was inadequate
for the volume  of material produced by the dredge,
and little  settling  took  place before dredge spoils

Table 2.—Water quality parameters measured during Van-
                couver Lake restoration.
Figure 6.—Diagram of dilution zones.
 Physical

   Depth
    Temperature
    Conductivity
    Turbidity

 Chemical
    Dissolved oxygen
    Conductivity
    NO3
    NH3
    Kjeldahl N

 Biological
    Chlorophyll a
    Algal species

 Toxics
    Mercury (in water and fish)
    Pesticides (in fish)
                                                                                 Suspended solids
                                                                                 Transparency (Secchi disk)
pH
Alkalinity
PO4
Total phosphorus
Agal cell counts
Zooplankton
                        Table 1.—Water quality requirements for Vancouver Lake dredging.
Return flow leaving a disposal area
Inside a silt curtain perimeter
Within a dilution zone
In the lake as a whole
Dissolved
Oxygen
4.0 mg/l
4.0 mg/l
4.0 mg/l
5.0 mg/l
Suspended
Solids
2000 mg/l
2000 mg/l
2000 mg/l
500 mg/l
 Toxic conditions resulting in dead or dying fish are not followed
                                                   583

-------
  LAKE AND RESERVOIR MANAGEMENT
  passed directly over the weir into the secondary diis-
  posal cell. The secondary cell had insufficient volume
  to accommodate the extra burden and by the third day
  of dredging the return flow leaving the disposal site
  was exceeding the established limit of 2,000 mg/l
  suspended solids. There was a silt curtain in place in
  the lake surrounding the outfall which contained the
  silt and prevented any problem in the lake itself. Dur-
  ing this  period the suspended solids in the dilution
  zone never exceeded 100 mg/l, well within the  allow-
  able 500 mg/l  limit.
   After the initial dredging the contractor built an ad-
  ditional weir in the end cells of each disposal area,
  thereby doubling the weir length in the cell just prior to
  returning to the lake. With this modification, weir crest
  heights and return flow suspended solids were kept
  within project limits for the rest of the project with the
  exception of the final days when disposal areas were
  severely limited.

  Multiple Retention  Ponds

  The design retention period of 8 days was never  realiz-
  ed because the equipment used by the contractor was
  so much larger than the design equipment. This com-
  bined with the very low density and silty nature  of the
  material being dredged to reduce the effectiveness of

 Table  3.—Comparison  of  design  and  actual  dredging
                    parameters.
                                   Design
  Actual
 Dredge size pipe dia.
 Horsepower
 Mrs/day
 Days/week
 Cubic meter/hr
 Cubic meter/day
 Work time to complete
 Pumping rate cu m/s
 Table 4.—Vancouver Lake Project dredge McCurdy— Sum-
     mary of Operations May 12,1982-March 24,1983.
50 cm
2000
18
6
658
11,835
2.0 yr
1.22
66 cm
5000
24
5
1,898
39,562
0.83
4.05
 Total cubic m
 Total effective hours
 Average effective hour/day
 Average cu m/effective hour
 Total working days
 Average cu m/working day
 Total working hours
 Average cu m/working hour

              Averages Per Working Day
 Crew size
 Bank
 Width cut
 Advance
 Pontoon line
 Shoreline
 Tofal pipeline
 Lift

             Noneffective Time Distribution
 Pipelines
 Clean pump & pipeline
 Spud & swing wires
 Clean suction & cutter
 Moving & anchors
 Elements
 Repairs
 Electric cable
Shore delay
Secure
Miscellaneous
Total
6,685,990
 3,521:35
   20:50
   1,898
     169
  39,5(52
 4,055:00
   1,648
     6.2
     526
     560
   2,3!52
   2,425
   4,777
    17.3


   00::>3
   00:18
   00:06
   00:01
   00:!50
   00:00
   01:03
   00:04
   00:12
   00:05
   00:12
   03.10
  the settling ponds. Only the final weir, prior to return
  to the lake, was enlarged sufficiently to reduce the
  crest height, further reducing the effectiveness of the
  system.
    Nevertheless, there was a noticeable improvement
  in quality of  the material  leaving each successive
  disposal area. At one time,  the effluent from the first
  pond had 1,000 rng/l  suspended solids, effluent from
  the second pond had 820 mg/l, and flow crossing the
  last weir into the lake had 580 mg/l suspended solids.

  Silt Curtains

  The initial use of the silt curtain around the return flow
  outfall was very effective.  Suspended  solids in the
  dilution zone outside the silt curtain  never exceeded
  the project limits,  even when measured within 0.5
  meter of the curtain.  For one 15-day period the mean
  suspended solids measured at the weir was 754 mg/l
  while the mean  suspended solids measured just out-
  side the silt curtain was 153 mg/l.
   During  this initial  use, the curtain was extended
  completely to the bottom.  As a result, the settling
  material fell on the curtain,  pulling it under the water
  surface and trapping it. This  curtain could not be
  retrieved by the contractor. As a result, in subsequent
  applications the contractor was reluctant to lower the
 curtain completely to the bottom. When not in contact
 with the bottom, the silt curtain was not as effective.
   The contractor elected not to construct the island
 disposal site using silt curtain containment, but chose
 to build in-water berms to contain the sediment. Con-
 sequently, the application of silt curtains at this site
 was quite similar to  that at  the land based  disposal
 areas. The curtains were used around the return flow
 outfall to provide a larger area for settling. Only when
 the water rose to cover a portion of the in-water berm
 were the curtains used as envisioned to provide the
 primary retaining device for the dredged material.
   In this application,  with the silt curtain not in  con-
 tact with the bottom of the lake, sediment-laden water,
 or perhaps more accurately watery sediment, could
 flow  as a  turbidity  plume  beneath the curtain. In
 general, this was not a problem except in instances
 when the water was mixed by wind. The period during
 which this problem existed was short in relation to the
 total length of the project, and project water quality re-
 quirements were maintained for the lake as  a whole,
 even in instances of excessive suspended solids in
 localized areas.
  To minimize the adverse effects of escaping silt, the
 contractor modified operations. The northeast shore
 disposal  site  and  the  island site,  both  in-water
 disposal areas, were in close proximity. By pumping
 for short periods (1 or 2 days) to each site alternately it
 was possible to minimize the water quality impacts.
SUMMARY

The restoration of Vancouver Lake required the dredg-
ing of 6.5  x 106 m3 of material from the lake, the con-
struction  of 17  km of land-based retaining dikes to
enclose 180 ha of  disposal area, and the disposal of
nearly 3 x 106 rn3 of material in the  lake to form an
island.
  The  requirement that all  dredge  return flow  be
returned to the  lake necessitated careful  control of
dredging activity and the imposition of several design
and operation features to control the quality of the
return flow water.
                                                  584

-------
  Some of the measures used included multiple set-
ting  basins,  extended weir length to reduce  crest
height, silt curtain enclosures around dredge disposal
site  outfalls, rapid  alternation  of dredge disposal
sites, and careful monitoring of dredging activity and
return flow quality. These measures enabled the pro-
ject to be completed with minimum delay, ahead of
schedule, and with no serious violation of water quali-
ty standards.
  Vancouver Lake is now recognized as a prime recre-
ational  resource for the Portland-Vancouver  metro-
politan  area. The success of this project can be at-
tributed to the willingness of the contractor and the
project  management and design team  to work to-
gether to find effective solutions to problems,  and to
        SEDIMENT PROBLEMS & MANAGEMENT TECHNIQUES


the skill and  imagination of the  contractor to  im-
provise solutions to meet project requirements while
still maintaining a productive work effort.
REFERENCES

Dames and Moore. 1977a. Vancouver Lake Pilot  Dredge
  Program. Prepared for the Port of Vancouver.
	.  1977D.  Master Plan: Rehabilitation of Vancouver
  Lake,  Vancouver, Washington.  Prepared  for Regional
  Plann. Counc. Clark County.
        1980. Operations Plan: Rehabilitation of Vancouver
  Lake, for the Port of Vancouver.
                                                   585

-------
  DREDGING AND  DREDGED MATERIAL  DISPOSAL

  TECHNIQUES FOR  CONTAMINATED SEDIMENTS


 RAYMOND  L MONTGOMERY
 U.S. Army Corps of Engineers

 Waterways Experiment Station
 Vicksburg, Mississippi


             ABSTRACT

             Significant advances have been made in recent years on dredging equipment, operating tech-
             niques, and disposal methods for contaminated sediments. Preliminary guidance is available for
             selecting dredges and operational techniques to minimize resuspension  of contaminated
             sediments during dredging. This is important since research results have indicated that most
             contaminants are attached to the clay-sized particles and natural organic solids found in lakes.
             Detailed guidance is available for designing dredged material containment areas based on the
             settling and physical properties of sediments. A modified elutriate test has been developed to
             predict levels of both dissolved and particui ate associated contaminants in containment area ef-
             fluents.  This test  can be used to determine the need for effluent treatment. Guidelines are
             available for designing treatment systems to clarify the effluent from containment  areas. This
             paper provides information on selecting dredging equipment and operational techniques to
             minimize sediment resuspension, design of containment facilities, and chemical treatment of
             containment area effluent to improve suspended solids removal.
 INTRODUCTION

 During recent years, the sediments  of the Nation's
 lakes have  increasingly become repositories  for a
 variety of  contaminants. This contamination is a
 result of man's activities including industrial expan-
 sion, widespread use of pesticides in  agriculture, and
 intentional  or  inadvertent  dumping  of  pollutants.
 Many of  the contaminants have an affinity for clay-
 size particles and natural organic solids found in most
 lake sediments. Although conventional dredges  are
 not specifically designed or intended for use in dredg-
 ing highly contaminated sediments,  many  feel that
 dredges are the logical, and perhaps only, means of
 removing contaminated  sediments  that  have beein
 found in the Nation's lakes. The Japanese have beem
 successful in developing special equipment for remov-
 ing contaminated  sediments from lakes and harbors
 (Kaneko et al. 1982).
  The Waterways Experiment Station has been con-
 ducting studies to determine the relative effectiveness
 of various methods of dredging contaminated sedi-
 ments. The specific environmental concerns address-
 ed include resuspension  of contaminated sediments
 and the possibility of contaminant release during the
 dredging  operation. A paper by Montgomery and Ray-
 mond (1982) presents an overview of the Corps  re-
 search program on dredging contaminated sediments.
  Potential problems associated with dredging con-
 taminated sediments include water quality impacts
 from releases during dredging, water quality impacts
 from effluent discharge during disposal, surface run-
 off and leachate following disposal, and uptake of
 contaminants by plants  and animals  inhabiting the
 area following disposal operations. Each of these pro-
 blems can be offset by one or more management prac-
tices.
  Since the nature and level of contamination in lake
sediments vary greatly on a lake-to-lake basis, ap-
 propriate  methods of dredging and disposal may in-
volve any of several available dredging and  disposal
alternatives. This strategy must provide a framework
for decisionmaking to select the best  possible dredg-
 ing  and disposal  alternatives and to identify ap-
 propriate  control  measures  to   offset  problems
 associated with the presence of contaminants.
   This paper provides information on selecting dredg-
 ing  equipment  and  operational  techniques  to
 minimize sediment resuspension, design of contain-
 ment facilities, and chemical treatment of contain-
 ment area  effluent  to  improve  suspended  solids
 removal.
DREDGING PRACTICE AND EQUIPMENT

Dredging  equipment  and  methods  have  been
developed over the years to enhance one of the two
basic uses of dredging, namely:
  •  Underwater  excavation to provide  or  maintain
navigable water depths in harbors and channels.
  •  Underwater mining and sand and gravel produc-
tion.
  Dredging practices in the United States have evolv-
ed to achieve the greatest possible economic returns
through maximizing production with only secondary
consideration given to environmental or aesthetic im-
pacts. The type of equipment and methods used in a
given job have been traditionally based  on  practical
considerations:
    Type and amount of sediment to be  dredged.
    Physical  and hydrologic characteristics of the
dredging site.
    Water depths in the area to be dredged.
    Dredged material disposal considerations.
    Availability of dredging equipment.
  Conventional dredges are not specifically designed
or  intended  for  use  in  removing  contaminated
materials resting  on  the lake bottom,  but  are the
logical and perhaps only feasible means to this end.
The considerations listed above are important in plan-
ning a normal dredging operation; however, additional
factors must  be  considered  because contaminated
materials may be involved, among them.
                                               586

-------
                                                            SEDIMENT PROBLEMS & MANAGEMENT TECHNIQUES
  •  Need for  precise determination and marking of
boundaries of  area to be dredged.
  •  Need for very precise lateral and vertical control
of dredging head.
  •  Requirement for special  precautions tailored to
specific chemicals.
  •  Requirement for special  treatment  during dis-
posal of dredged material.
  •  Need to predict the likely damage to aquatic and
benthic organisms  to be caused by  the  dredging
operation  and  its effect on resuspension of any con-
taminant.
  Sediment Resuspension During Dredging. Investi-
gations by  Fulk, Gruber, and Wullschleger (1975)
showed that, for suspended solids concentrations of
less than 100 g/l, the amount of pesticides and PCBs
that are dissolved or desorbed into the water column
from the resuspended sediment  is negligible. They
determined that contaminants were basically trans-
ferred to the water column attached to solids. They
also reported that the reduction of suspended solids
concentration due to settling resulted in a decrease in
contaminant  concentrations.  The  spread  of  con-
taminants during dredging operations then is linked to
the resuspension of sediments, particularly clay-sized
and organic particles.
  It has been demonstrated that elevated suspended
solids concentrations are generally limited to the im-
mediate vicinity of the dredge  and  dissipate rapidly at
the  completion of  the  operation. However, when
dredging  contaminated  sediment,  equipment  and
operational techniques must be selected to minimize
sediment resuspension. Sediment  resuspension char-
acteristics of selected dredges and operational tech-
niques have been evaluated by Raymond (1983). He
concluded  that sediment resuspension  caused by
many  dredges  could  be lessened  by  controlling
operating techniques or by modifying the equipment.
Many  researchers suggest  that  controlling  cutter
revolutions per minute, swing speed, and depth of cut
of a cutterhead dredge can reduce sediment resus-
pension.  In  fact, any operating technique  that im-
proves the production value of the cutterhead dredge
will probably reduce resuspension. Bucket dredges
will probably require some equipment modification to
achieve a meaningful reduction in sediment resuspen-
sion.  Finally,   a wide  variety of special-purpose
dredges are available that appear to substantially
reduce the resuspension of sediment. However, most
of these dredges have low production rates, and more
research is needed to evaluate their areas of applica-
tion and their limitations.
  Hydraulic Cutterhead  Dredge. The  cutterhead
dredge is basically a hydraulic suction pipe combined
with a cutter  to loosen material that  is too  con-
solidated to be removed by suction alone (Fig. 1).  This
combination of  mechanical  and  hydraulic  systems
makes the cutterhead one of  the  most versatile and
widely used dredging systems; however, its use  also
increases  the  potential for sediment  resuspension.
While a properly designed cutter will cut and guide the
bottom material toward the suction efficiently, the
cutting action and the turbulence associated with the
rotation of the cutter resuspend a portion of the bot-
tom material. The level of sediment resuspension is
directly related to the type and quantity of material cut
but  not picked up by the suction.
  While  little   experimental  work on  cutterhead
resuspension has been done, several  field studies
have attempted to identify the extent of cutterhead
resuspension. Barnard (1978), reporting on the field in-
vestigations of Huston and Huston (1976) and Yagi et
al. (1975), stated that, based on the limited field data
collected  under  low-current  speed  conditions,
elevated levels of suspended  material appear to be
localized in the immediate vicinity of the cutter as the
dredge swings back and forth across the dredging
site. Within 10 ft of the cutter,  suspended solids con-
centrations are highly variable, but may be as high as
a few 10's of grams per liter; these concentrations de-
crease exponentially with depth from the cutter to the
water surface. Near-bottom suspended solids concen-
trations may be elevated to levels of  a  few  hundred
milligrams per  liter at  distances of 1,000 ft from the
cutter.
   Recent advances  in  dredging  technology  have
transformed  conventional  dredges  into  highly
specialized  excavation  equipment.  One major
advancement in this field has been in the design of
portable  dredges.  In   the  past, mobilizing   and
demobilizing a conventional dredge often required a
large portion of total job time. Now, with smaller hulls,
modular  construction,  and  even  amphibious
capabilities, many dreges can be transported overland
from one job to the next with minimal effort. Clark
(1983)  conducted a comprehensive survey of  portable
hydraulic dredges available in the  United States and
summarized  the performance capabilities  of  each
dredge.
   Bucket Dredge. The  bucket  dredge consists of
various types of buckets operated from a crane or der-
rick mounted on a barge or on land. It is used exten-
sively  for removing  relatively small  volumes  of
material, particularly around docks and piers or within
restricted areas. The sediment  removed is at nearly in
situ density; however,  the production  rates are quite
low compared  to that for  a  cutterhead  dredge,
especially in  consolidated material. The  dredging
depth  is practically unlimited, but the production rate
 DISCHARGE LINE
                                           CUTTERHEAD
Figure 1.—Hydraulic cutterhead dredge.
Figure 2.—Spillage from conventional bucket.
                                                 587

-------
 LAKE AND RESERVOIR MANAGEMENT
 drops with increase in depth.  The bucket  dredge
 usually leaves an irregular, cratered bottom.
   The  resuspension of  sediments during  bucket
 dredging is caused primarily by  the impact, penetra-
 tion, and withdrawal of the bucket from the  bottom
 sediments. Secondary causes are loss  of material
 from the bucket as it is pulled through the wateir,
 spillage of turbid water from the top and through the
 jaws of  the bucket as it breaks  the surface,  and in-
 advertent spillage while dumping (Fig. 2). Limited field
 measurements on sediment resuspension caused by
 bucket dredges showed that the maximum suspended
 sediment concentration in the immediate vicinity of
 the  dredging operation  was less than  500 mg/l and
 decreased rapidly with distance from the operation
 due to settling and mixing effects (Raymond, 1983).
 The major source of turbidity in  the lower water col-
 umn is sediment resuspended at the impact point of
 the clamshell.
   Although experimenters have reported some reduc-
 tion in sediment resuspension with the variation of
 hoist speed and depth of cut, the greatest reduction in
 resuspension with clamshell dredging came from the
 use  of a so-called "watertight" or enclosed clamshell
 bucket.  The Port and  Harbor  Institute of  Japan
 developed a watertight bucket in  which  the top is on-
 closed  so that  the  dredged material  is contained
 within the bucket. A direct comparison of a 1 cu m
 standard open clamshell  bucket with  a watertight
 clamshell bucket indicates  that  watertight buckets
 generate 30 to 70 percent less resuspension in ihe
 water column than the open buckets (Fig. 3).
   Raymond (1982) conducted a field test to compare
 the effectiveness of enclosed clamshell buckets. The
 resuspension  produced by  an  enclosed 13  cu yd
 bucket was compared to a  12 cu yd standard open
 bucket during dredging of the St.  Johns River near
 Jacksonville, Fla. The results of this test are given in
 Table 1.
   This test reveals a marked reduction (greater than 50
 percent)  in sediment  resuspension in the upper water
 column with the enclosed  bucket. However, some
 drawbacks were also revealed. The enclosed bucket
 increased resuspension near the  bottom, probably
 resulting from a shock wave of water that precedes
 the watertight  bucket because of  the enclosed toD.
   Special Purpose Dredges. Special purpose dredging
 systems  have been  developed during  the last  few
 years in  the United  States  and  overseas to pump
 dredged material slurry with a high  solids content or
 to minimize the resuspension of  sediments. Most of
 these systems are not intended  for use on typical
 maintenance operations; however, they provide alter-
 native methods for dredging contaminated sediments
 from  lakes  when  the  capabilities  of  a particular
                                                     system provide some advantage over conventional
                                                     dredging  equipment. The  special purpose  dredges
                                                     that  appear to  have the  most potential in limiting
                                                     resuspension are shown in Table 2, which was taken
                                                     from Herbich and Brahme (in press).


                                                     CONTAINMENT AREA DESIGN AND
                                                     OPERATION

                                                     Containment Area Design.  Diked containment areas
                                                     are  used  to retain  dredged material  solids while
                                                     allowing the carrier water to be  released from the
                                                     containment area. The two purposes of containment
                                                     areas are: (1) to provide adequate storage capacity to
                                                     meet dredging requirements, and (2) to attain the
                                                     highest possible efficiency in retaining solids during
                                                     the dredging operation in  order to meet  effluent
                                                     suspended  solids requirements.  These  considera-
                                                     tions are  interrelated and  depend upon effective
                                                     design, operation, and management of the  contain-
                                                     ment area.  Basic guidelines for design, operation,
                                                     and  management of containment  areas are  pre-
                                                     sented by  Palermo et al. (1978) and Montgomery et al.
                                                     (1983). Confined disposal of contaminated sediments
                                                     must be planned to  contain potentially toxic mate-
                                                     rials to control or minimize potential environmental
                                                     impacts (Fig. 4). There are four major mechanisms for
                                                     transport  of contaminants from upland disposal
                                                     areas.
                                                      1.  Release of contaminants in the effluent during
                                                     disposal operations.
                                                      2.  Leaching into ground water.
                                                      3.  Surface  runoff  of  contaminants  in either
                                                    dissolved  or suspended particulate form following
                                                    disposal.
                                                   Figure 3.—Enclosed 13-cu yd bucket.
                 Table 1.—Resuspension rates for isediments using open and watertight buckets.
                                                                  Average Suspended Sediment
                                                                         Levels, mg/*
    Type of
Clamshell Bucket
                                 Sampling
                                  Radial
   Upper
Water Column
  Near
Bottom*
   Watertight
                                    1
                                    2
                                    3
                                    1
                                    2
                                    3
 "Averages adjusted for background suspended solids levels
"Measurements made with 5 ft of bottom
     Open
     27.0
     35.6
     80.6

    233.0
    300.0
    N/A
 123.25
   61.0
  133.3

  146.6
  122.0
  N/A
                                               588

-------
                                                            SEDIMENT PROBLEMS & MANAGEMENT TECHNIQUES
  4. Plant uptake directly from sediments, followed
by indirect animal uptake from feeding on vegeta-
tion.
  The major components of a dredged material con-
tainment are  shown  schematically in Figure 5. A
tract of land  is surrounded by dikes to form a con-
fined surface area into which dredged sediments are
placed. Dredged material  is usually placed in these
sites by pipeline dredges or barge pumpout. In some
instances material may be placed  directly into the
sites by a bucket dredge. When placed hydraulically,
the coarse material rapidly falls out  of suspension
and forms a mound near  the dredge  inlet pipe. The
fine-grained material (silt and clay) continues to flow
through the containment  area  where most of the
solids settle out of suspension and thereby occupy a
given  storage volume.  The  fine-grained dredged
material is usually rather homogeneous and is easily
characterized.
  The clarified water is discharged  from the contain-
ment area over a weir. This effluent can be char-
acterized by its suspended solids concentration and
rate of outflow. Effluent flow rate  is approximately
equal to influent flow rate for continuously operating
disposal areas. To promote effective  sedimentation,
ponded water is maintained in the area; the depth of
water is controlled by the elevation of the weir crest.
The thickness of the dredged material  layer increases
with time until the dredging operation is completed.
  HYDRAULIC DISCHARGE OF
  POLLUTED DREDGED MATERIAL
Figure 4.—Confined disposal area effluent and leachate con-
trol.
                    CROSS SECTION
 Figure 5.—Schematic diagram of a dredged material con-
 tainment area.
                             Supernatant waters from confined disposal sites
                           are discharged after a retention time of up to several
                           days. Procedures have been developed to predict con-
                           centrations of suspended solids in disposal area ef-
                           fluents, taking into account settling behavior of the
                           sediment in question (Montgomery et al. 1983). Several
                           factors influence the concentration of suspended par-
                           ticulates  present in  supernatant  waters. Fine par-
                           ticulates  become suspended  in the disposal  area
                           water column at the point of entry due to turbulence
                           and mixing. The suspended particulates are partially
                           removed from the water column by sedimentation.
                             A modified elutriate test procedure developed by
                           Palermo (1983)  can be used to predict  both the dis-
                           solved  and particulate-associated concentrations of
                           contaminants  in confined  disposal  area effluents
                           (water discharged during active disposal operations).
                           The laboratory  test simulates contaminant release
                           under confined  disposal conditions and reflects sedi-
                           mentation  behavior of dredged material, retention
                           time of the containment, and chemical environment in
                           ponded water during active disposal.
                             Effluent  Controls.  A well-designed  containment
                           area will reduce the effluent suspended solids to a low
                           level. The modified elutriate test can be performed to
                           evaluate  the  particulate-associated contaminants.  If
                           the effluent suspended solids level is not acceptable,
                           further treatment is necessary. The suspended solids
                           and turbidity  in the effluent from a containment area
                           are colloidal  materials that do not readily settle by
                           gravity. Chemical clarification  is one  method  to
                           remove these fine particles from water. Many different
                           chemicals are  available which, when added to the
                           water,  will aggregate  the particles into  a dense floe
                           that settles quickly. Efficient clarification requires op-
                           timum  chemical doses and mixing conditions  for floe
                           formation.  Chemical   clarifictaion  is   a treatment
                           method primarily for the removal of  suspended, not
                           soluble or dissolved, material from water.
                             The  chemical treatment  process can be  used to
                           polish the effluent from the primary containment area
                           as shown in  Figure 6. This system would require a
                           small secondary cell to collect the remaining solids in
                           the effluent and a structure to mix the flocculant with
                           the effluent; normally, the weir structure is used. The
                           required  chemical doses would  be typical of water
                           treatment systems. This system  would  ensure that
                           discharges sent to the receiving waters would be of
                           good quality.
                             Procedures  and  design   guidelines  have  been
                           developed by Schroeder (1983) for designing treatment
                           systems to clarify the effluent from a dredged material
                           containment  area. Guidelines are also presented on
                           operating the systems and estimating  the treatment
                           costs.  Treatment cost, excluding the costs  of con-
Table 2.
Name of Dredge
                          Suspended Sediment Level
 Mudcat Dredge


 Pneuma Pump


 Clean-Up System*


 Oozer Pump*

 Refresher System*

 'Japanese dredges
5 ft from auger, 1000 mg/l near bottom (background level 500 mg/l)
5 to 12 ft in front of auger, 200 mg/l surface and mid depth (background level 40 to 65 mg/l)

48 mg/l 3 ft above bottom
4 mg/l 23 ft above bottom (16 ft in front of pump)

1.1 to 7.0 mg/l 10 ft above suction
1.7 to 3.5 mg/l at surface

6 mg/l (background level) 10 ft from head

4 to 23 mg/l at 10 ft from head
                                                  589

-------
 LAKE AND RESERVOIR MANAGEMENT
 structuring a secondary cell, would range from $0.08 to
 $0.25/yd of in situ material dredged. Additional con-
 trols can be used to remove fine particulates that will
 not settle or to remove soluble contaminants from the
 effluent.  Examples of these  technologies are filtia-
 tion, adsorption, selective ion exchange, chemical ox-
 idation, and biological treatment processes. Beyond
 chemical clarification, only limited  data exists  lor
 treatment of dredged material (Gambrell et al. 19761).
   Leachate Controls.  Subsurface drainage from con-
 fined disposal sites in an  upland environment may
 reach adjacent  aquifers.  Fine-grained dredged mate-
 rial tends to form its own disposal area liner as par-
 ticles settle with percolation drainage water, but the
 settlement  process may  require some time for solf-
 sealing to develop. Since most contaminants poten-
 tially present in dredged material are closely adsorbed
 to particles, only the dissolved fraction will be present
 in leachates.  The site-specific nature of subsurface
 conditions is the major factor in determining  possible
 impact (Chen  et al. 1978).
   Leachate controls consist of measures to minimize
 groundwater pollution  by preventing mobilization of
 soluble contaminants. Control measures include pro-
 per site selection as described earlier, dewaterinci to
 minimize leachate production, chemical admixing to
 prevent or retard leaching, lining the bottom to prevent
 leakage and seepage, capping the surface to minimize
 infiltration and thereby leachate production, vegeta-
 tion to stabilize  contaminants and to increase dry ng,
 and leachate collection, treatment, or recycling (Gam-
 brell  et al. 1978). A rule-of-thumb for cost of  liners is
 $0.01 per mil thickness per ft2 (installed).
   Runoff Controls. After  dredged  material has been
 placed in a confined disposal site and the dewatelng
 process has been  initiated, contaminant mobility in
 rainfall-induced  runoff is considered in the overall en-
 vironmental impact of the dredged material being
 placed in a confined disposal site. The quality of the
 runoff  water  can  vary depending on  the  physico-
 chemical process and the  contaminants present ir the
 dredged material. Drying  and oxidation will promote
 microbiological   activity,  which breaks down  the
 organic component of the dredged material  and ox-
 idizes sulfide compounds to more soluble sulfate
 compounds. Concurrently reduced iron  compounds
 will become oxidized and iron oxides will be formed
 that can act as  metal scavengers to adsorb soluble
 metals and render them less soluble.
   The pH of the dredged  material will be affected by
 the amount of acid-forming  compounds present as
 well as the amount of basic compounds that can buf-
 fer acid formation. Generally, large amounts of sulfur,
 organic matter, and pyrite material will generate acid
 conditions.  Basic components of dredged  material
 such  as calcium  carbonate will tend to neutralize
 acidity produced.  The resulting pH of the  dredged
 material will depend on the relative amounts of acid-
 formed and basic compounds present.
   Runoff controls at conventional sites consists of
 measures to prevent  the erosion  of  contaminated
 dredged material and the dissolution and discharge of
 oxidized contaminants from the surface. Control op-
 tions  include maintaining ponded conditions, planting
 vegetation to stabilize the surface, liming the surface
 to prevent acidification and to reduce  dissolution,
 covering the surface with synthetic geomembranes,
 and/or placing  a lift of clean material to cover the con-
taminated dredged material (Gambrell et al. 1978).
   Control of Contaminant Uptake. Plant and animal
uptake controls are measures to prevent mobilization
  of  contaminants into  the food chain.  Control
  measures  include selective vegetation  to  minimize
  contaminant uptake,  liming or chemical  treatment to
  minimize or prevent release of contaminants from the
  material to the plants,  and capping with clean sedi-
  ment or excavated material (Gambrell et al. 1978).  A
  test protocol has been developed for evaluating poten-
  tial plant uptake. This procedure has been applied to
  testing a number of contaminated dredged  materials
  and has given appropriate results and information to
  predict the potential for plant uptake of contaminants
  from dredged material (Folsom and Lee, 1981; Lee et
  al.  1982; Folsom et al.  1981).


  SUMMARY

  Significant advances  have been made in  recent years
  on  dredging equipment,  operating techniques, and
  disposal methods for contaminated sediment. Preli-
  minary guidance  is available for  selecting dredges
  and operational techniques to minimize resuspension
  of contaminated sediments during dredging. This is
  important since  research  results have indicated that
  most contaminants are  attached to the clay-size and
  natural organic  solids  found  in  lakes.  Detailed
 guidance is available  for designing dredged material
 containment areas based on the settling and physical
 properties of sediments.
   A modified elutriate  test has been developed to
 predict  levels  of   both dissolved and  particulate
 associated  contaminants  in  containment  area ef-
 fluents. This test can  be used to determine  the need
 for  effluent treatment.  Guidelines  are available for
 designing treatment systems to clarify  the effluent
 from  containment  areas.  Treatment   costs  are
 estimated  to range  from $0.08  to $0.25/yd3  (1981
 dollars) of in situ material  dredged.
REFERENCES

Barnard, W.D. 1978. Prediction and control of dredged mate-
  rial dispersion around dredging and open-water pipeline
  disposal operation. Tech. Rep. DS-78-13. U.S. Army Eng.
  Waterways Exp. Sta., Vicksburg, Miss.

Chen, K.Y., D.  Eichenberger, Mang, J.L, and R.E. Hoeppel.
  1978. Confined disposal area effluent and leachate control
  (laboratory and field investigations). Tech. Rep. DS-78-7.
  U.S. Army Eng Waterways Exp. Sta., Vicksburg, Miss.

Clark, G.R. 1983. Survey of portable hydraulic dredges. Tech.
  Rep.  HL-83-4.  U.S. Army  Eng. Waterways Exp.  Sta.,
  Vicksburg, Miss.

Fulk, R, D. Gruber, R. Wullschleger. 1975. Laboratory study
  of the release of pesticides and PCB materials to the water
  column during dredging operation. Contract Rep. D-75-6.
  U.S. Army Eng. Waterways Exp. Sta., Vicksburg, Miss.

Folsom, B.L, Jr., and C.R. Lee. 1981. Zinc and cadmium up-
  take  by the  freshwater marsh plant Cyperus escu/entus
  grown in contaminated sediments under reduced (flooded)
  and oxidized (upland) disposal conditions. J. Plant Nutr. 3:
  233-44.

Folsom, B.L, Jr.,  C.R. Lee, and K.M. Preston. 1981. Plant
  bioassay of materials from  the Blue  River dredging pro-
  ject.  Misc. Pap. EL-81-6. U.S. Army Eng. Waterways Exp.
  Sta.,  Vicksburg, Miss.

Gambrell, R.P., R.A. Khalid, and W.H. Patrick. 1978. Disposal
  alternatives  lor  contaminated dredged material as  a
  management tool to minimize adverse environmental ef-
  fects. Tech. Rep. DS-78-8. U.S. Army Eng. Waterways Exp.
  Sta., Vicksburg, Miss.
                                                 590

-------
Herbich, J.B., and S.B. Brahme. In press. A literature review
  and technical evaluation of sediment resuspension during
  dredging. Tech. Rep. U.S. Army Eng. Waterways Exp. Sta.,
  Vicksburg, Miss.

Huston, J.W.,  and W.C. Huston.  1976. Techniques for re-
  ducing turbidity associated with  present  dredging  pro-
  cedures and operations. Contract  Rep. D-76-4. U.S. Army
  Eng. Waterways Exp. Sta., Vicksburg, Miss.
Kaneko, A., Y. Watari, and  N. Aritomi. In press. The special-
  ized  dredges designed for the bottom sediment dredging.
  In Proc. 8th U.S./Japanese Experts  Conf.  Toxic  Bottom
  Sediments, Tokyo, Japan, November  1982. Tech. Rep. U.S.
  Army Eng. Waterways Exp. Sta., Vicksburg, Miss.
Lee, C.R., B.L. Folsom Jr., and R.M. Engler. 1982. Availability
  and  plant uptake  of  heavy metals from  contaminated
  dredged material placed in  flooded and upland disposal
  environments.  Environ. Int. 7:65-71.
Montgomery,  R.L, and G.L. Raymond.  1982. Overview of
  Corps research program on dredging contaminated sedi-
  ments. In Proc. 8th U.S./Japanese Experts Conf. Toxic Bot-
  tom Sediments, Tokyo, Japan, November 1982. U.S. Army
  Eng. Waterways Exp. Sta., Vicksburg, Miss.
Montgomery, R.L, E.L. Thackston, and  F.L.  Parker. 1983.
  Dredged material sedimentation basin design. Am. Soc.
  Civil  Eng. J. Environ. Eng. 109(2).
         SEDIMENT PROBLEMS & MANAGEMENT TECHNIQUES

Palermo, M.R.  1983. Interim guidance for conducting modi-
  fied elutriate tests for use in evaluating discharges from
  confined dredged material disposal sites. Draft Eng. Tech.
  Letter. Office, Chief Eng., Washington, D.C.
Palermo, M.R., R.L. Montgomery, and  M. Poindexter.  1978.
  Guidelines for designing, operating, and managing dredg-
  ed material containment areas. Tech. Rep. DS-78-10. U.S.
  Army Eng. Waterways Exp. Sta., Vicksburg, Miss.
Raymond, G.L. 1983. Sediment resuspension characteristics
  of selected dredges. Draft Eng. Tech. Letter. Office, Chief
  Eng., Washington, D.C.

Raymond, G.L. In  press: Field study of the  sediment resus-
  pension characteristics of selected dredges. In Proc. 15th
  Annu. Texas A&M Dredging Seminar, New Orleans. Texas
  A&M Univ. College Station.
Schroeder, P.A. 1983. Chemical  clarification  methods  for
  confined dredged material  disposal. Tech.  Rep. D-83-2.
  U.S. Army Eng. Waterways Exp. Sta., Vicksburg, Miss.
Yagi, T.,  et al. 1975.  Effect of operating conditions of  hy-
  draulic dredges on dredging capacity and turbidity. Tech.
  Note  228.  Port  Harbor Res. Inst.,  Ministry  Transport,
  Yokosuka, Japan.
                                                     591

-------
 DREDGING  FOR CONTROLLING E-UTROPHICATION OF
 LAKE KASUMIGAURA, JAPAN
 KEN  MURAKAMI
 Public Works Research Institute
 Ministry of Construction
 Tsukuba Science City
 Ibaraki-ken  305, Japan
            ABSTRACT

            Lake Kasumigaura, the second largest lake n Japan, is an extremely eutrophicated lake, yet is one
            of the most important water resources around the Tokyo metropolitan area. It is a shallow lake with
            an average depth of 4 m. Nutrient exchange between the lake water and sediments was found to
            be the major factor affecting the nutrient balance within the lake. Therefore, it was decided to carry
            out dredging of sediments along with other measures to control eutrophication of the lake. To dredge
            fluffy sediments on the top layer effectively and to avoid excessive disturbance of the sediments, a
            special purpose dredge equipped with the Oozer system was built, and later, a modified version was
            constructed. The engineering aspects of drsdge efficiency are summarized in the paper based on
            the 5-year experience.
INTRODUCTION

Lake Kasumigaura, the second largest lake in Japan,
is located about 70 km northeast of Tokyo (Fig. 1, 2).
It  consists of three portions, Nishiura, which  is the
largest portion, and is often called "Kasumigaura,"
Kitaura, and Sotogasakaura. The lake is connected
with the Pacific Ocean through the Tone River estuary,
and is slightly saline. A gate was constructed  at the
mouth of the connecting river to the Tone in 1963 for
flood and saline water control. Since then the lake has
become completely freshwater.
  The morphological characteristics of the lake and
its basin are shown in Table 1. It is a very shallow Lake
with an average depth of 4 m. The retention time of the
lake water is about 7 months. There are 47 munic-
ipalities within the basin with a  total population of
720,000. The industries in the basin are mainly prirrary
industries, among which rice farming, hog raising, and
carp cultivation in  the lake are significant sources of
nutrient loadings. The inventory of the nutrient loading
to the lake is summarized in Table 2.

CURRENT STATUS OF
LAKE KASUMIGAURA

Lake Kasumigaura is one of the few remaining  large-
scale water resources which can be further developed
around the metropolitan Tokyo area. However, it is ex-
tremely eutrophicated, and restoration of the lake is
one of the most important targets in the water pollu-
tion control program in Japan.
  Figure  3 shows  the  changes of  annual average
chemical  oxygen demand (permanganate method),
total nitrogen, and total phosphorus of the lake water.
The change of annual average  transparency is also
shown in  Figure 4. In summer, a heavy algal bloom of
Microcystis develops, sometimes covering the whole
lake surface. The variation of chlorophyll a is shown in
Figure 5.
  In  1982, the prefectural government established an
ordinance to regulate the nutrient loadings to the lake,
that  includes effluent  standards for nitrogen  and
phosphorus and a ban on the sale and use of phos-
phorus-containing detergents. The major part of the
effluent standards is shown in Table 3. The ordinance
also  restricts  the  nutrient  runoff  from  nonpoint
sources to some extent by promoting the use of hog
manure as a  fertilizer and  proper  application of
chemical  fertilizers. The  nitrogen and phosphorus
standards for the lake water are to be established by
the central government in the very near future.
  It has been considered, however, that the nutrient
                   Table 1.—Morphological characteristics of Lake Kasumigaura and its basin.
Water Surface (km2)
Nishiura
Kitaura
Sotogasakaura
Others
Depth (m)
Maximum
Average
220
171
34
6
9

7
4
Lake volume (million mj)
Annual inflow (billion m3)
Drainage area (km2)
Annual precipitation (mm)




800
1.4
2,157
1,370




Table 2.— Inventory of nutrient loadings to the lake.

Nitrogen
Phosphorus
Natural
2.50
0.17
Farming
4.31
0.08
Domestic
:\.OA
0.42
Industrial
0.61
0.17
Livestock
3.35
0.28
Fish
Cultivation
1.45
0.27
Total
15.26
1.39
                                                                                     Unit: ton/d
                                               592

-------
                                                            SEDIMENT PROBLEMS & MANAGEMENT TECHNIQUES
loading control is not sufficient for restoring the lake
since the nutrient release from the bottom sediments
contributes greatly to the nutrient budget within the
lake.

SEDIMENT QUALITY

Sediment quality  was surveyed horizontally and ver-
tically throughout the whole lake. Water content, igni-
tion loss, Kjeldahl nitrogen (KN) and total phosphorus
(TP) of the surface sediments were in the ranges of
100-600 percent,  7-20 percent, 1-10 mg/g (dry), and
0.4-2.4 mg/g (dry), respectively, except the area with
sandy  sediments. The horizontal distribution of TP
content of the sediments is shown in Figure 6. TP con-
tent is  high at the northern bay of Nishiura (Takasaki-
iri) and Kitaura. The horizontal distribution of KN con-
tent is a little different from that of TP, and KN content
is high at the central part of Nishiura.
  The  vertical distribution  of sediment quality was
also examined at  many locations. Some examples are
shown in Figure 7. Both phosphorus and nitrogen con-
tents are high at  the surface layer—particularly, the
phosphorus  content of the top 20-30 cm layer, which
                                     Pacific
                                      Ocean
                    is quite high compared with that of the lower layer in
                    most cases. Using these kinds of data, the relation-
                    ship between the sediment volume and TP  content
                    was calculated, and is shown in Figure 8. The figure
                    indicates, for example, that the volume of sediments
                    containing more than 1.4 mg/g of phosphorus is about
                    10 million m3 and that containing more than 1 mg/g of
                    phosphorus is about 40 million m3.
                    NUTRIENT RELEASE FROM SEDIMENTS

                    Extensive experiments and field observations have
                    been made to evaluate the nutrient release from the
                                                     Figure 2.—Lake Kasumigaura.



                                                       15

                                                     S. 10
                                                     Q
                                                     O
                                                     O
                                                       15


                                                      ,1.0

                                                       05


                                                        0
                             OI5
                             0.05
                                                            TN
                                                          TP
                                                       _L_J	L_
 Figure 1.—Location of Lake Kasumigaura.
                               I960      1965      1970     1975
                    Figure 3.—Variation of annually averaged COD (Mn), TN and
                    TP.
                        Table 3.—Effluent standards provided by the Prefectual Ordinance.
         Category
                                                            Nitrogen
Daily Flow (m°/d)
Existing
                                                                     New
                                                     Phosphorous
Existing
                                                                New
Sewage Treatment Plant

Night Soil Treatment Plant
Food Processing Industry



Metal Processing Industry


Other Industries

20-100,000
more than 100,000
more than 20
20-50
50-500
more than 500
20-50
50-500
more than 500
20-50
50-500
more than 500
20
15
20
25
20
15
30
20
15
15
12
10
20
15
10
20
15
10
20
15
10
12
10
8
1
0.5
2
4
3
2
3
2
1
1.5
1.2
1
1
0.5
1
2
1.5
1
2
1
0.5
1
0.5
0.5
                                                                                         Unit: mg/l
                                                  593

-------
  LAKE AND RESERVOIR MANAGEMENT
                    Table 4.—Comparison of loadings from sediments and basin, Nishiura, in 1977.
Loadings from Sediments

Nitrogen
Phosphorus
TOG
'Static
Release
661
69.8
2,810
'Dynamic
Release
547
90.9
3,190
Subtotal
1,208
160
6,000
Loadings
from
Basin
4,340
418
12,700
Total
5,548
578
18,700
  'by diffusion through the water sediment interface
  'rapid transit from sediment caused by disturbance

 sediments  of  Lake Kasumigaura.  The result of the
 survey is summarized in Table 4.
    It  is clear from the Table that  the effect of the
 sediments  is quite significant in the nutrient budcjet
 within the  lake. Moreover, both static and dynamic
 releases are the function of temperature, and become
 large in the hot season,  supporting  the heavy algal
 bloom during summer.


 DREDGING

 Dredging  by Kasumi

 Dredging in Lake Kasumigaura started in FY1975 with
 a planned dredging volume  of 300,000 m3 during  7
 years from FY 1975 to FY  1981. At that time, 'the
 knowledge  of  sediments  with  respect  to Lake
 Kasumigaura was much less than now, and the total
 volume of sediments to be dredged was considered to
 be 1.2 million m3.
   The dredge selected for the project was choser (1)
 to be capable of dredging fluffy sediments at the sur-
 face, (2) to minimize resuspension of  sediments, and
 (3) to keep the solid content of the dredged material as
 high  as possible to reduce the site area for disposal
 and also to reduce the quantity of supernatant from
 the disposal site.
   It was considered that  a pneumatic pump dredge
 was  the most appropriate for the project. An  experi-
 mental dredge with  pneumatic pumps was available
 at that time; it had been built for dredging sediments
 in the tidal  estuary of the Tsurumi River near Tokyo.
 This dredge, later named "Kasumi," had a pneumatic
 pump system which was a  modified version  of the
 "Pneuma Dredge" developed by SIRSI of Italy. The ma-
 jor specifications of Kasumi were as shown in Table 5.

          Table 5.—Specifications of Kasumi.

 Vessel

    Displacement Tonnage: 78 tons
    Dimension: 16 m(L) * 5 m(W) * 1.7 m(D)
 Dredge

    Type: Pneumatic Pump 0.5 m3 * 2 barrels
    Nominal  Capacity: 60 m3/h

  The schematic diagram of the pump is shown in
 Figure 9. By  the vacuum pressure applied to the barrel
(-400 mmHg), the sediments  are sucked into the* bar-
rel until the dredged material  in it reaches a predeter-
mined level. Then, the compressed air (7 kg/cm?) is
supplied to the barrel in order to discharge the dredg-
ed material into a barge. The  pump consisted o: two
barrels operated alternatively  to give semicontinuous
suction and delivery. One cycle of the pump operation
took 60 seconds. The operation, however, was essen-
tially spot dredging with a batch mode of operation.
                                                                                        Unit: ton/year
   The first series of dredging was carried out from
 January to  March  1976. Since  the dredge Kasumi
 could not transport the dredged  material by pipeline,
 barges were used to carry the dredged material from
 the dredge to the unloading station at the shore. The
 dredged material then was transported to the disposal
 site by a pipeline. Prior to the disposal, a solidifying
 agent was dosed to increase the bearing strength of
 the dredged material.
   The result of the first dredging series is shown in
 Figure 10. The volume  of the dredged sediments per
 day was about 100 m3 on  the average. The rate of
       1911 1950
                     I960
                           1965   1970    1975
 Figure 4.—Variation of transparency.
      3 8 M 2 5 8 Jl 2 5 S III 2  5 8 II1 2 5  8 II 2 58 II 258 II 25U
      1972 -i- 1973  --1974 ^ -1975 -<  1976- '-1977- ' -1978  ' 1979
Figure 5.—Variation of chlorophyll.
Figure 6.—Horizontal distribution of phosphorus content.
                                                 594

-------
                                                            SEDIMENT PROBLEMS & MANAGEMENT TECHNIQUES
 dredging, calculated  as the volume of the dredged
 sediments per 1 hour of actual operation, was about
 17 m3/h. This was much smaller than the nominal
 capacity of the dredge.
   The solid content of  the sediments was from 13.4
 percent to 80.2 percent  with a  typical value of 30 per-
 cent. The difference between the water contents of the
 sediment and the dredged material was very small.
 Judging from the  increase in  the water content, the
 ratio of water to sediments being pumped was about 1
 to 4. No supernatant to  be disposed of was produced
 at the disposal site.
   With the addition of a solidifying agent of 5 percent
 (w/w) prior to discharging into the disposal site, the
 bearing strength of the dredged  material became 4
 tons/m2 after 40 days from the disposal.
   The dredging operation by Kasumi was carried out
 for 3 years. The quantity of the dredged sediments is
 shown in Table 6.

   Table 6.—Quantity of sediments dredged by Kasumi.
   Year
                      Quantity (m3)
FY 1975
FY 1976
FY 1977
Total
4,000
18,000
13,000
35,000
                                      Although Kasumi had many advantages, the effi-
                                    ciency of dredging was too low to dredge the planned
                                    volume of sediments for the following reasons:
                                      1.  The design capacity of the dredge itself was too
                                    small,
                                      2.  Clogging of the pump valves stopped the opera-
                                    tion fairly frequently, and
                                      3.  Operation of barges in a shallow water body with
                                    fishery activity often limited the total efficiency.
                                      Moreover, defects of the spot dredging, such as
                                    difficulty in controlling the sediment thickness to be
                                    dredged and difficulty in getting clear sediment sur-
                                    face  after dredging, were evident. Therefore,  it was
                                    decided to build a modified version of the dredge with
                                    a larger capacity.
                  Dredging by Koryu

                  The design of the new dredge "Koryu" started in 1976,
                  and the new dredge was completed in March 1978.
                  The basic concepts of the design were as follows:
                    1. To have the capacity of more than 100 m3/h when
                  dredging sediments with the  water content (water-
                  solid ratio) of 200 percent,
                    2. To employ  the swing arm system for moving the
                  suction head, and
      o
  GL  0

     20


 —   40
 o
 ~"   60
 _c

 f   80
 o

     100


     120
           WC
           TP
250        500
 ,  (mg/g)  ,
           KN
25         5
 .  (mq/g)
                           to
    No. 2
 GL  0


    20


—   40

o
~"   60
-c:

S"  80
O

   100


   120
                            WC
                                                                             WC
                                                                       /
                                                                       I
          TP
          KN
250        500
 ,   (mq/g I  ,
2.5         5
 ,   (mg/g)   ,
                                                            10
0        250        500
    TP   ,   (mg/g)   ,
0         2.5         5
,    KN   ,   (mg/g)   ,
                                                                          o
                   GL. 0

                      20

                  _   40

                  I   60


                  s-  80
                  a
                      100

                      120
                                                                      No. 4
  GL.  0

     20


 —   40
 o
 -   60
 f  80
    100

    120
           WC
           TP
250        500
  i  (mg/g)   ,
           KN
 2.5         5
  ,  (mg/g)  ,
                           10
        o  Water Content  (WC)
        •  Total Phosohorus  (TP)
        A  Kjeldahl Nitrogen  (KN)
                                                            JO 2
   No. 6
                                                                        NO. 6
Figure 7.—Vertical distribution of sediment quality.
                                                 595

-------
LAKE AND RESERVOIR MANAGEMENT
  3. To have a booster pump for pipeline transporta-
tion of the dredged material.
  The major specifications of  Koryu are shown in
Table 7.
          Table 7.—Specifications of Koryu.
Vessel
    Displacement tonnage: 260 ton
    Dimension: 25 m(L) * 8 m(W) * 2.4 m(D)
Dredge
    Type: Pneumatic Pump 0.85 m3 * 2 barrels
    Nominal capacity: 100 m3/h
Pipeline Transportation Facility
    Maximum transportation distance: 2 km
    Delivery head: 400 m

  The schematic diagram of the pump and the suction
head is shown in Figure 11. The pump and the suction
head are supported by a ladder which swings laterally.
This configuration enables continuous dredging of a
thin layer. The mechanism of the pump itself is essen-
tially the same  as that  of Kasumi. Each barrel of the
pump is operated  alternatively to make semicon-
tinuous suction and delivery. The time needed for one
cycle of pump operation is reduced to 30 seconds by
increasing the vacuum  pressure to  -500 mmHg. The
dredged material  is discharged into a small storage
tank installed in the vessel. After being  screened to
remove  large debris which might  clog  the  booster
pump, it is pumped to the disposal  site by a 150 mm
diameter pipeline. The booster pump for this purpose,
which is also mounted in  the vessel, is of an oil
pressure type with a delivery head of 400 m.
  The shape of the suction head greatly affects the
performance of the dredge. A couple types of the suc-
                    12   14   16

                    TP Content (mq/g)
                                        20
                                                      tion heads were tested, and the one shown in Figure
                                                      12 is currently used when the thickness of the sedi-
                                                      ment layer to be dredged is 30-50 cm. The relationship
                                                      between the solid content of the dredged material and
                                                      the swing velocity of the  suction head  is shown in
                                                      Figure 13 when the suction head shown in Figure 12 is
                                                      used. As the  swing velocity of  the suction head in-
                                                      creases, the solid  content of the dredged material in-
                                                      creases. However, the percentage of the sediments
                                                      actually dredged among the sediments intended to be
                                                      dredged decreases as shown in Figure 14. Moreover,
                                                      resuspension  of the sediments  becomes significant
                                                      when the swing velocity is high. Therefore, it was con-
                                                      cluded that the swing velocity of a range from 5 to 7
                                                      m/min would be appropriate under normal conditions.
                                                        Inevitably, this way of dredging pumps up a con-
                                                      siderable amount of water with the sediments. Figure
                                                      15 shows the  ratio of the dredged material volume to
                                                      the dredged sediment volume. The ratio varied widely
                                                                                      Vacuum  Pump
                                                             Delivery Pipe
                                                       Level Switch
                                                                                           Compressor
                                                                                 VaLve

                                                                                 Diffuser

                                                                                 Delivery Valve

                                                                                 Inlet Valve
                                                                Screen
                                                    Figure  9.—Schematic  diagram  of pneumatic  pump
                                                    "Kasumi."
                                                                                                20
                                                                                                <0 A
                                                       ISO
                                                    JE

                                                    ~  too

                                                     E

                                                    co
                                                    -o
                                                     CT1
                                                    -o
                                                    a  50 -
                                                          15          30 I
                                                               Jon
10      20
  Feb
                                                                                             Mar
Figure 8.—Relationship between sediment volume and TP
content.
                                                   Figure 10.—Dredging performance of "Kasumi."
                                                596

-------
                                                              SEDIMENT PROBLEMS & MANAGEMENT TECHNIQUES
 from 1.4 to 2.4 with an average value of about 1.8,
 which is still much too low compared to a cutterless
 pump dredge. The supernatant at the disposal site is
 currently being discharged into the lake. A treatment
 facility to remove phosphorus from the supernatant is
 to be installed in the near future.
  The volume of the dredged material and the area
 dredged in a day  (10  hours) obtained from  actual
 operation  are plotted  on  Figure  16. The  average
 volume of  the dredged  material was  about 1,000 m3,
                                        Vpcuum Pump
                                        Air Compressor
     Sediments
    ~~7
Suction Mouth

Figure 11 .—Schematic diagram of pneumatic pump "Koryu."
                                    Mouth
                                     960    /A
                                  1.080
Figure 12.—Schematic diagram of suction head.
   20
   15
   10
T3
i  5
                                    0-,30cm
                   6      8      10     12
                   Swing Velocity  (m/mm )
14
Figure 13.—Relationship between solid content of the dredg-
ed material and the swing velocity of the suction head  D in
the figure is the depth from the sediment surface to the bot-
tom of the suction head.
       which is equivalent to about 550 m3 of the dredged
       sediments, and the average area dredged was about
       1,400m2.
         The quantity of the dredged sediments from 1978 to
       1981 is shown in Table 8. Although the efficiency of
       the dredge was somewhat  lower than originally ex-
       pected,  dredging of  the  planned  300,000 m3  of
       sediments within 7 years could be performed success-
       fully.
                                                         Table 8.—Quantity of sediments dredged by Koryu.
         Year
                                          Quantity
                                                       FY 1978
                                                       FY 1979
                                                       FY 1980
                                                       FY 1981
                                                        Total
                                                  34,000
                                                  40,000
                                                  85,000
                                                 106,000
                                                 265,000
                                                        ~ 90

                                                         o 80
       -t* 60
          50
       •5   40
       •2   30
                     2      4     6      8      (0     12
                       Swing  Velocity  (m/min.)

       Figure 14.—Relationship between swing velocity and percen-
       tage of sediments actually dredged among the sediments in-
       tended to be dredged : D in the figure is the depth from the
       sediment surface to the  bottom of the suction head.
                                                        20 r
                                                         10
       1.4   15   16  17   18  19  20  21  22  23  24
       ">   5   $    (    5    5    i    5    5   5   S
       15   16   17  18   19  20  21  22  23  24  25

       Ratio of  drdged  material volum to dredged
        sediment volume


Figure 15.—Frequency distribution  of the ratio  of dredge
material volume to dredged sediment volume.
                                                  597

-------
LAKE AND RESERVOIR MANAGEMENT


   2,000r
   1,500
 1
 <
O O
  •  •
  COM
                         o   o
                            o
                 O  O


                  O


oxr
o 
-------
GIBRALTAR  LAKE  RESTORATION  PROJECT - A RESEARCH AND
DEVELOPMENT PROGRAM FOR EVALUATION OF THE
TRANSPORTATION (DREDGING) OF CONTAMINATED SEDIMENTS
RAYMOND E. SPENCER
Spencer Engineering
A Division of Martin, Northart & Spencer, Inc.
Santa Barbara, California
           ABSTRACT

           Santa Barbara has had a leadership role in the field of environmental concerns for over 50 years
           and for that reason had a concern of the total environmental considerations involved in the
           transportation (dredging) of contaminated sediments from their main water supply reservoir,
           Gibraltar Lake. Silt from the adjacent watershed has been reducing the City's available water
           supply (equal to 360,000 cu yds) annually with the total of approximately 22,000,000 cu yds of silt
           currently contained within the reservoir. This R&D Program was sponsored in part by the U.S. En-
           vironmental Protection Agency (EPA) to evaluate the removal of mercury contaminated silt using
           an air pump to a containment area and to assist other responsible agencies in evaluating this
           method when reviewing other similar projects in the future.
 INTRODUCTION

 Siltation of Gibraltar Lake since its construction in
 1920 has been occurring annually at an average rate of
 275,250 m3 (360,000 cu yds) for 60  plus years and in
 1975 reached a point such that additional siltation will
 diminish a valuable water supply in an already water-
 short area. Since the silt was contaminated (cinnabar
 mining operations), the city of Santa Barbara (90 miles
 northwest of Los Angeles) applied  in 1977 to the En-
 vironmental Protection Agency's Clean Lakes Pro-
 gram to propose  a pilot program  for the use  of a
 recently developed air pump manufactured in  Italy.
 This air pump (trade name "Pneuma") came to be the
 heart of the dredging system adopted by the city as
 the most feasible in terms of environmental efficiency,
 cost effectiveness, and manageability. As of  June
 1983, the city of Santa Barbara has contracted to pur-
 chase a Pneuma pump (revised to specifications) and
 to operate it annually similar to the pilot dredging of
 the harbor. This will maintain the current water supply
 reservoir, but does not provide for any long-term  solu-
 tion regarding the removal of the remaining 22,000,000
 cu yds currently contained within Gibraltar Lake.  Also,
 environmental  concerns  surrounding  canyons  limit
 the relocation of contaminated silt even for the annual
 dredging program.
  This paper presents first hand information on incor-
 porating a  piece of foreign equipment with current
 dredging technologies and evaluates the environmen-
 tal  results  of relocating  contaminated silts from a
 potable water supply.

 LIMNOLOGICAL STUDIES AND  WATER
 QUALITY MONITORING PROGRAM
 (PRE-DREDGING)

 For the Gibraltar Lake restoration project to serve as a
 meaningful model to predict environmental impacts in
 similar lake desiltation  projects,  the research and
 development study included a carefully designed and
 executed limnological monitoring program conducted
by an independent research organization along with
additional monitoring by the city of Santa Barbara.
Certain specific aspects of this program depended on
meteorological  conditions,  time  and  duration  of
dredging, and choice of disposal site, which had to be
addressed before the lake restoration project began.
  Baseline data on the limnology of Gibraltar Lake
was available from the water quality analyses con-
ducted weekly by the city and from its consultant,
Ecological Research Associates of Davis, Calif. How-
ever, since lake characteristics are variable from year
to year, it was essential to obtain additional informa-
tion immediately prior to the initiation of dredging and
during dredging. Also, since any dredged water must
be regulated and  monitored through  the State of
California Regional Water Quality Control Board, two
coordinated lake monitoring programs  were set up
between ERA and the city of Santa  Barbara to meet
the requirements established for the project.
Figure 1. — Location map.
                                              599

-------
LAKE AND RESERVOIR MANAGEMENT
   In the first, three intensive field surveys were con-
 ducted by ERA. These investigations concentrated on
 defining the complete limnological status of the lake
 before and during the dredging. (Table 1 lists the para-
 meters measured.)
   A second  monitoring  program was conducted by
 the city of Santa Barbara throughout the period  of
 dredging. Weekly water quality  analyses were per-
 formed on samples taken from near the water supply
 intake (near the dam), near the site of dredging and at
 the point where the spoil supernatant flowed back into
 the  lake. This would allow the  city to adjust  its
 domestic supply intake in case noxious algal blooms
 or elevated levels of mercury were observed.
   In concert, the two programs provide valuable data
 for predicting the environmental impacts of similar
 desiltation projects. The information also assures the
 city a  high quality water supply and a balanced and
 aesthetically appealing lake ecosystem.
       Table 1.—Parameters measured in intensive
                limnological surveys.
Dissolved O2
Temperature
PH
Secchi depth
Nitrate - N
Ammonia - N
Kjeldahl - N
Turbidity
Alkalinity
Ortho P
Total P
Hg (water)
Hg (sediments)
Hg (fish)
Light penetration
Specific conductivity
Chlorophyll a
Algal biomass
Primary productivity
Suspended solids
Particulate - C and N
  The relationship between low dissolved oxygen and
the potential effects of dredging is especially impor-
tant when  considering the release of mercury from
suspended sediments. Through bacterial metabolism
in anaerobic environments, mercury in the sediments
can be transformed to methyl mercury, which is highly
toxic and  readily taken  up by aquatic  organisms
(Jernelov, 1970; Wood et al. 1968).
  The first intensive field investigation was the pre-
dredging survey conducted in October 1980, approxi-
mately 3 months prior to the initiation of dredging. The
locations of the sampling stations for  the lake survey
are shown on Figure 2.
  Mercury  content of sediments sampled  in mid-
October ranged from 0.08 ppm at station F to 0.29,0.27
and 0.47 ppm at stations A, B and D, respectively, the
latter three stations being the future sites of dredging
operations. Predredging levels of mercury in the water
column were quite low, < 0.0006 ppm at all  stations
and all depths tested, except for values of 0.0013 and
0.0018 ppm at  station A (Om) and station C  (7.5 m).
respectively.  Total  mercury concentrations  in  fish
have also been relatively high (0.2-0.8 ppm) at times,
pointing to bioconcentration of  this toxic metal.
  The Food and Drug Administration has established
an action level for mercury in the edible portions of
fish, shellfish, crustaceans, and  other aquatic animals
of 1.0 ppm.  Although levels of fish  and sediment
averaged  only half this  amount,  future dredging
should continue to be monitored especially as dredg-
                          OESIGNATED DISPOSAL AREA
                          NORTHSIDE  WEST
                       IOO' DOWNSTREAM FROM  DRE03E

                       300' DOWNSTREAM FROM  DREDGE

                       WATER INTAKE

                       AT  DREDCE SITE

                       LAKE , 15 '  FROM SHORE

                       LAKE , 40' FROM SHORE
       \,
                                                FIGURE  2

                                        LAKC  SAMPLING  STATIONS
                                                                                  SCALE  IN  MILES
Figure 2 —Lake sampling stations.
                                                 600

-------
                                                           SEDIMENT PROBLEMS & MANAGEMENT TECHNIQUES
ing approaches the old cinnabar tailings located far-
ther up the lake.


GEOTECHNICAL DATA

Prior to any construction plans and specifications, a
technical report was requested of Geotechnical Con-
sultants, Inc., of Ventura, Calif., on their findings from
the silt in  the lake bottom and soil tests of adjacent
canyons they conducted in the summer of 1979.
  Siltation of the lake occurs seasonally as flood run-
off, primarily from the Santa Ynez River and secondar-
ily, from several minor  watersheds  which enter the
lake from  the north and south. Major siltation has
resulted from forest fires that have consumed over
135,700 acres of the valuable watershed area (Fig. 3).
Approximately 80 percent of the watershed has been
burned at  one time or another over the past 50 years.
   Earth materials recovered from  the  silt  bottom
generally consist of interbedded mixtures of sand, silt,
and clay with local accumulations of organic debris
and then  layers of gravelly sand and  silt. (Later we
discovered that the debris had a deleterious effect on
the pumping  characteristics and rubberfaced valves
of the Pneuma system.)
   No discernible trends in sediment sorting were ap-
parent from the cores other than an overall westerly
direction grain size across the lake toward the dam
and the obvious contribution of angular gravel mine
tailings extending  roughly 200 feet into the lake from
the presently inactive Mercury Mine.
   The  technical findings of the geotechnical report
were evaluated by the various dredging contractors to
determine the ability and optimum  design of equip-
ment for use  in bidding the desilting pilot program.
                                                     Phase  One—Filling the  pump: Each cylinder is
                                                   rapidly filled with liquid, either by a counterpressure
                                                   from the  hydrostatic head (in the case of immersed
                                                   plants) or by gravity (in the case of stationary convey-
                                                   ing and booster plants). As soon as one cylinder is fill-
                                                   ed, the inlet valve automatically closes by its own
                                                   weight.
                                                     Phase  Two—Emptying the  pump  and reflowing:
                                                   When the cylinder has been  filled, compressed air,
                                                   supplied by a compressor through the distributor and
                                                   air hose, acts as  a piston and the  liquid is thus  forced
                                                   out through the delivery valve.
                                                     Phase Three—Discharging compressed air and pre-
                                                   paration for Phase One: When the cylinder has been
                                                   almost emptied, the distributor discharges the air into
                                                   the atmosphere.  Once the internal pressure is  releas-
                                                   ed the cylinder once again becomes filled with liquid,
                                                   as described in phase one.
                                                     The Pneuma system consists of a pump body (com-
                                                   posed of three cylinders), compressors, shovels, and a
                                                   distributor  system that  automatically controls the
                                                   supply of compressed air to the  cylinders. When the
                                                   pump is submerged, sediment and water are forced in-
                                                   to one of the empty cylinders through an inlet valve,
                                                   simultaneously forcing the material  out of an outlet
                                                   valve and into the discharge line. When the cylinder is
                                                   empty, the air pressure  is reduced to atmospheric
                                                   pressure, the outlet valve closes and  the inlet valve
                                                   opens. The  two  stroke  cycle is  then  repeated. The
                                                   distributor system controls the cycling phases of all
                                                   three  cylinders  so  there  is  always  one  cylinder
                                                   operating in the  discharge mode.
                                                      Using two 1,400 CFM and one 1,600 CFM air com-
                                                   pressors  by Ingersoll Rand,  and a  14 = inch steel
                                                   discharge pipe, the contractor, using a Pneuma plant
                                                   Model 450/80 pumped an average of 700-800 cu yd of
                                                   slurry per hour at an average velocity of 7-9 feet per
 DREDGING

 The Pneuma system, developed  by Sirsi (Societa
 Italiana Recerche Struttamenti Idrice), Florence, Italy,
 was the first dredging system to use compressed air
 instead of centrifugal motion to pump slurry through a
 pipeline. Although it  has been used extensively on
 European  and  Japanese  dredging  projects, the
 Pneuma system has only been available in the United
 States since 1975. (According to the literature pub-
 lished by the manufacturer,  this system can pump a
 slurry with a relatively high solids content with little
 generation of turbidity.)
   A patented system, the work cycle of the Pneuma
 pump can be divided into three  phases:
                      *  i  5 ?
                   CALENDAR  YEAR

Figure 3 —Sedimentation at Gibraltar Lake (May 1979)
                                                    Figure 4 -Dredging
                                                 601

-------
 LAKE AND RESERVOIR MANAGEMENT

 second with  an average of 40-50 percent  solids, a
 distance of 2,300 feet and  170 feet of vertical  head.
 (The 170 feet  of head was from the bottom of the lake
 at the feet of pump up to the discharge line and with
 the depth of  water being 60 feet ± to the bottom of
 the trench  at the  pump head.) At times it actue.lly
 pumped a slurry with the consistency of a milk shake
 (sometimes with solid chunks of clay). During most of
 the pumping,  only the two 1,400 CFM air compressors
 were used  to push  the material  up to the canyon
 disposal site.
    Using  imported  equipment presented  a  few pro-
 blems.  First, a court hearing was held in Florence, Italy,
 to determine  the  legalities of shipping the Pneuma
 plant to the contractor. This hearing caused a 60-day
 delay. Also, while the pumping literature was written
 in English,  other communications and the timingi of
 telephoning were difficult.
    When the pump arrived and was ready to be f itted to
 the barge, it was found that the wrong barge dimen-
 sions had been sent from  the factory in Italy. This
 caused a delay of approximately 5 days. (In this case,
 it was very  difficult to check the shop drawings).
    Then when  pumping actually got under way on Nov.
 19,1980, the Italian engineer stated that the discharge
 lines were not fabricated properly and were causing
 restrictions in the discharge of  air and slurry. Three
 weeks were spent rectifying these problems through a
 process of trial and error. (During this time the Italian
 engineer had returned to Italy.) After the discharge line
 problems had been solved, it was found that the pump
 itself was part of the problem right from the beginning.
 This time two engineers were dispatched from Italy
 along with  the president of the company. After a 2
 week period  of trial and error of changing internal
 parts and adjustments, the entire system started per-
 forming as  promised.
    After a few  months of good production, the contrac-
 tor experienced major problems with internal  parts
 breaking and malfunctioning. Welding of the metal
 material inside the pump was failing, causing the con-
 tractor to patch and mend  continuously. The rubber
 fittings on the valves in each cylinder did  not erdure
 debris in the silt such as rocks and wood  well; these
 tore the fittings so that the valves would not clos« pro-
 perly, and lost air pressure.
   At times a major shut down was required. The con-
 tractor  completely stripped  the internal  parts of the
 pump  and  refabricated  the parts  out  of superior
 metals  and different compounds of  rubber   and
 urethane. Certified welders were used to perform all
 welding  to ensure  proper  welding technology.  (On
 future projects using a pump of this nature, the tanks
 of the pump should be tested for proper welding as per
 ASME code, Section VIII, Division  I. This will ensure
 the tanks will hold designed air pressure safely).
    Once this service was performed, the  pump  per-
 formed at its maximum ouput again. But, this same
 problem  existed throughout the pumping phase, of
 parts breaking and the debris eating away at the  rub-
 ber valves. The contractor ultimately devised a proper
 mixture of rubber and  urethane to solve the valve
 deterioration problem.
    When the pump was first used the so  called "pot
 holing" feet were used in lieu of trailing or pulling the
 barge thorugh the silt. The pump was simply dropped
 into the silt, letting the silt flow into the hole dug by it.
 This method worked for the first month  in getting the
 loose material off the top of  the lake bottom. After
 this, the material stopped flowing to the pump and the
 efficiency of the  pump dropped  drastically. The
 straight extended  feet were  taken off and trailing
 shovels installed, requiring pulling the barge and forc-
 ing the  silt into the three  shovel  feet. The  frontal
 shovels are normally equipped with cutting grills that
 facilitate penetration into the compact bottom sedi-
 ment.
    Depending on the size of the particular Pneuma
 pump used, production  rates can  range  from 40 to
 2000 cu m/hr. The pump can be deployed from a land-
 based floating crane, pulled through the sediment  in a
 trailing position, or attached to a dredging ladder.
Figure 5.—Gibraltar Lake barge
Figure 6.—Disposal site ponds before and after decanting.
                                                 602

-------
                                                             SEDIMENT PROBLEMS & MANAGEMENT TECHNIQUES
POST DREDGE LIMNOLOGY

Dredging  began  on a  nearly  continuous basis  in
January  1981  and extended into August 1981 with
several inoperative periods (Feb. 2-9; April 30-May 19;
and June 17-July 6) while repairing the pump.
   In late April  1981, the lake was just  beginning  to
stratify at  a depth of 6-8 m. By July the lake had
developed a stable stratification with the thermocline
at 9.5 m. Levels of dissolved oxygen were severely
depleted in the near bottom waters. A comparison  of
oxygen profiles at Station C (control station near the
dam) and Station  A (100 ft from dredge) shows no
significant difference between  them and in point  of
fact, concentrations in the deeper waters at Station C
were slightly lower than  at Station A. These data in-
dicate  that sediment resuspension during dredging
was not causing increased oxygen depletion even  in
bottom waters close to the site of dredging.
   In the July 1981 survey, which was carried out while
the dredge was in actual operation, turbidities were
extremely low at all Stations A, C and F (<1  NTU). The
level of suspended solids at all  stations at all depths
was less than 3.2 mg/l, and there was no significant
difference (P >0.1) in the values between Stations A, C
or F. Secchi depth was slightly higher, however,  at Sta-
tion C (4.9 m) as compared with Station A (4.3 m) and F
(4.3 m). Such data provide for a sensitive measure  of
the extent to which air pump dredging leads to sedi-
ment resuspension and increased turbidity. Using this
air driven dredge system the effect is indeed insignifi-
cant, except for minor increases in  solids content  in
the nearbottom waters within 100 ft of actual dredging
and  from  possible return  flows   from  sediment
disposal  site.
   Considering the above, it is not surprising to note
that profiles of dissolved oxygen, as measured in the
city's weekly sampling  program, did not reveal any
enhanced  oxygen  depletion at  stations  near the
dredge as compared to control station C. In addition,
vertical profiles of primary productivity (using the sen-
sitive 14C method) and paniculate carbon and nitrogen
which  were determined  during the three  intensive
studies, did not show significant differences between
Stations  A, C, and F. Enumerations  of algal cell den-
sities also failed to show major differences between
stations during dredge operation. Taken together with
the  chlorophyll  and  turbidity  values described
previously, these data clearly indicate that the  dredg-
ing operation  did not significantly increase  either
algal biomass or rates of production (growth).
  This limnological study of  the  Gibraltar  Lake
desiltation  program provides important data  which
allow the environmental evaluation of air pump  dredg-
ing and its future application to similar desiltation pro-
jects. Upon detailed evaluation of  all  of  the  para-
meters, the results clearly reveal that when properly
operated, the air pump dredge has little or no adverse
environmental impact on  the aquatic system. Specific
data supporting this conclusion are the following:
   1.  Elevated levels of mercury, which could pose a
health  hazard  to domestic supplies,  were  never
detected  in the lake water at any station, at any  depth.
  2.  Increased levels of nutrients (N or P) were never
found at stations  near  the dredging operation, but
were restricted to near-shore waters close to the point
of spoil supernatant return flow.
  3.  There was no significant  increase in  turbidity,
suspended solids concentration, or color in surface
waters at any station, even adjacent to  the dredging
operation (Table 2). For the most part, during normal
operation sediment  resuspension was  restricted to
the bottom waters near the area of dredging. (Tran-
sient events did occur, however,  which  led to short-
lived  resuspension near the dredge or in waters off-
shore of the point of spoil supernatant discharge).
   4. There  was  no  pattern  of  oxygen  depletion
associated with any aspects of the desiltation pro-
gram.
   5. There were no noxious algal blooms which could
impart taste, odor, color or turbidity to the water.
Table 2.—Levels of turbidity, suspended solids, Secchi depth
 and color from the city of Santa Barbara weekly sampling
  program (X/Y denotes surface/mid-depth; all other data
               from surface samples).
Date/Station
(1981)
4 March



11 March



18 March



25 March



1 April



8 April



15 April



22 April



29 April




27 May




3 June




10 June




A
B
C
E
A
B
C
E
A
B
C
E
A
B
C
E
A
B
C
E
A
B
C
E
A
B
C
E
A
B
C
E
A
B
C
E
F
A
B
C
E
F
A
B
C
E
F
A
B
C
E
F
Turbidity
(NTU)
5.1/2.6
3.4/1.4
—
78.0
1.9/2.7
1.2/1.8
—
93.0
1.0/1.0
1.0/1.0
—
0.9
1.8/1.9
1.1/1.0
—
5.2
1.0/1.3
0.8/0.98
—
20
0.64/0.63
0.83/0.81
—
0.82
0.99/1.8
1.0/0.93
—
39
0.87/1.00
0.85/0.88
—
1.20
0.86/0.99
0.84/0.91
—
76
1.1/12
0.87/0.79
0.42/0.45
—
0.99
0.73/0.84
0.67/0.84
0.71/0.54
—
25.0
1.90/3.70
0.56/0.61
0.58/0.57
—
22
0.62/0.85
Susp. Solids
(mg/l)
5/5
—
—
189
1.8/1.6
—
—
240
—
—
—
—
1/4
—
—
—
2.6/4.20
—
—
17.6
3.0/3.2
—
—
1.6
1.8/1.4
—
—
15.0
0.2/0.2
—
—
0.6
1.2/2.4
—
—
304
1.2/19.6
2/0.2
—
—
1.8
0.8/1.4
1.8/1.6
—
—
49.2
3/29.6
1.4/1.6
—
—
530
1.6/1.0
Secchi
(m)
—
—
—
—
—
—
—
—
4.2
4.2
4.2
3.5
2.6
3.0
3.0
1.1
3.5
3.5
3.5
0.4
3.8
3.8
3.5
3.5
3.8
4.0
3.8
0.2
4.0
4.0
4.0
4.0
3.0
3.0
3.0
—
—
4.5
5.0
5.0
1.5
4.0
4.0
4.0
4.0
2.0
2.0
3.0
3.0
3.0
2.0
—
Color
28
12
14
—
17
17
18
—
13
15
12
-^
18
15
13
—
12
16
12
—
12
11
9
—
10
10
10
—
7
7
10
—
9
9
10
818
10
—
—
—
—
—
6
8
8
112
14
4
6
4
140
4
                                                 603

-------
 LAKE AND RESERVOIR MANAGEMENT
                                    Table 3.—Cost: breakdown of grant funds.
  Federal funds (EPA - Clean Lakes Program)
  City of Santa Barbara (matching funds)
  Categories
  1. Planning and environmental studies
  2. Engineering, design, construction plans and specs
  3. Construction supervision, administration, inspection, testing and reports
  4. Construction and pumping contract
  Construction and Pumping Cost Breakdown
  1. Mobilization and demobilization
  2. Disposal site development costs
  3. Research & development (changes to pump & disposal methods)
  4. Pumping (including downtime and employee travel time)
                                                                      Total Funds
                                                                            Total
                                    $1,150,000(50%)
                                    $1,150,000 (50%)
                                    $2,300,000
                                    $ 184,000 ( 8%)
                                    $ 184,000 (  8%)
                                    $ 161,QUO ( 7%)
                                    $1,771,000(77%)
                                    $2,300,000
                                    $
                                                                            Total
  442,750 (25%)
  637,550 (36%)
$ 124,000 ( 7%)
$ 566,700(32%)
$1,771,000
SUMMARY AND CONCLUSIONS
Gibraltar  Lake is a source  of  high  quality drinking
water for the city of Santa Barbara that should be pro-
tected from degradation. Desiltation  will assure both
an adequate water storage capacity  and prevent p'o-
gressive eutrophication  as the  lake  fills in with silt.
The air pump dredge is  indeed  a unique technology,
with little or no adverse environmental impact. Such a
system is recommended for similar  lake  restoration
projects  which  require the  environmentally  sale
transport  of aquatic sediment.
  In conclusion, the Pneuma system  has the capabili-
ty to pump a high percentage of solids (40 percent to
50  percent)  that  are  noticeable above a standard
dredge system. Also, the pump has limited distur-
bance of the lake bottom and minimal turbidity of mer-
cury contaminated  silt  within  the adjacent  waters.
Based  on  our  findings,  the  system  (with  some
modifications) was a very satisfactory dredging tool
and the city of Santa Barbara in June 1983 entered in-
to a contract to purchase the equipment for a continu-
ing dredging operation.
 REFERENCES

 Ecological Research Associates. 1977.  In Application for
  Gibraltar Lake Restoration Project to the U.S. EPA, City of
  Santa Barbara, Calif.
         1979. Pre-Dredging Baseline Limnological Survey
  of Gibraltar Lake.

Goldman, C.R. 1963. The Measurement of Primary Produc-
  tivity and Limiting Factors in Freshwater with Carbon-14.
  Tech. Inf. Rep. TID-7633: 103. U.S. Atomic Energy Comm.
Jernelov, A. 1970. Release of methyl mercury from sediments
  containing inorganic mercury at different depths. Limnol.
  Oceanogr. 15:958.
Standard Methods for the Examination of Water and Waste
  Water. 1975. 14th ed. Am. Pub. Health Ass., New York.
Strickland, J.D.H.,  and T.R  Parsons. 1968. A practical hand-
  book of seawater analysis. Bull. Fish. Res. Board Can. 167.
U.S.  Environmental Protection Agency. 1979. Methods for
  Chemical Analysis of  Water and Wastewater. U.S. Gov.
  Printing Off., Washington, D.C.
Wood, J., S. Kennedy, and C.G. Rosen. 1968. The synthesis
  of methyl mercury by extracts of methanogenic bacterium
  Nature 220:173.
Wetzel, R.G. 1975. Limnology. W.B.  Saunders,  Philadelphia.
                                                   604

-------