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
-------
|