JMunjcjpai Environmental Research EPA-600/9-80-056
, ^.Laboratory ' ;/.; '" December 1980
* ; : ?i Agency Cincinnati OH 45268
;::\a;^ iiS'/iii"::-:;':1 „.,,/'':l",-.
'SteS^ • - • • • ••• •
Urban Stormwater and
i^K:-::!.'ftV!i' frM'W';:':• -, • ' ,„, •
Sewer Overflow
Hf » 'h
* \ i 1
on Receiving
l^fcl'IT'tt'^rtr'''/^"^'"':1' '.'.'IVki^f'.^'''''.^^*-''1'1,.1 "fv V:,1 i"i', *hi«t-'" ',",'l,.» 'is,..' >','.!',
ii^^
•T'T''T r!*!/'':!It1iiter Bodies
,^iT.:;,.^j:tVJ'>^.i-::^ft.t^^ $M$i$N!^
>• . , ; ,'' '•;,,;;;?: V ^i^'S'^ .,','..;•:;;: •;',*;"...:,":;,
^Q£S£i^I^National
inference
'".'•ins' "i VSiMSra^ra.ilKilSrtT'n1:1- WS-'-nr":-!!1".' l;!*?1^-"™?!.!;1!,11- v
i'M1 ..... 1 ..... Si'' t\'i!jfif:.x>i. ........ K'f'iii
'^'''i^^^-^f
..... iji !!!i,i,.,;:|"iil(;iii:ij;|i!iiiiir:;», ..... J.,*',"* Si1;;!,*1"'1'"!!!".!: ii .'Jii/li" ..... ' ..... iji
^
SikifeiJ.HEivMilidiCi,yi»3ii;&l*i,,4i• '?iHI.B...•.;«,4";.» •'!'...-•,;..'"-'• -••. , i.
;'"",!'''!!'!'», iltf:;:!!'?*,*,*! ^ ifwlt i'4i f1'^!h^ r iSfSW S.'frhS'Vi'iEJJ "S^ff*1:' OS'S1* 11,^^ 1= ;"-, rf ? ,..'."'. I."! .,,''.,„- ^' ,!:• : *,
t ,-! i "J:!' rtr 'I iihtS it, I'll', US1,;!'' If III1 it' "i1 ftn
Jiiil'i''.'!!'^'^'!'!:!!'''^!',^ !r!,, i«i ''vf'* B', -'Wilh',!, I IIS'iW
*:^
~
SS|S||||*!^||f !lfl':i'§||¥ ^Sfe^^l ;'M^^^' ftj.l^'^ '•?•. "i'l *":':::' •' •::
""" '""' ' ™ Bl'll'ljillMi'''^^:;^^^ '"^" X ''!'!, "'"''"";' iii'1''''' *':'' i »" ,'",,.''!.„ ,'"';;'' H'i.' ""
;1:" ""! " '•' "B'S'Sji'iSS *'.'«!!•'! Ill'Si"..';:!¥'',«;;'': ;,,: ','' I'l" :<''
([w*A!a i'» i*!^isi,;»i s iii'Sy PI^ ^.^'^i1 W1'^' f»:::
., "W «'li '«l«l«l»'»npl".f »iri;|,|ll"iHl.n|"ui,l.» W"p| .II^Hi'.nflll.n.l.Hi'l^.riium.J'llllpplw » Jf:«,y ,„'.„:,..-•!,, .!'.., :( ,, , , ; . „ ',i's , - \ ' ' . _ i •, . " .
fjiklJII'IIWIp WByrB^ ''^I'li1*!*!1'*11!^ l''.^!.^!!!: li'iHSI "i "l-ll'l1:1' Wl* <>« ff " |II:"IHM :T; sil "I' .K ;:'",';''"''.; :"', ".
, ,,,, ,,,,,,,,,
,,,,1!;„•; !,;:,.ti«i,,ii,fjij.'.M «..friPf.,!,rffusii,tiiniutv«Mifira3W
t;;E,;i:y,,;;;J|v,i;;|= ,4^
J 'J... i* iij,,,, : ,.,wi,r,*. ri; .iiinrt*.' '««i ,"|M:i!' |i|"«i." i': ill. uii«' • ft, i'i||:ii*!.. 'JNiii'ii',', ji^i'll.'.'.^^^'^ rjiiint!;™^ jiiir ill '"|l1*!|;pi
i,?:^
,l!'i,;'iijil4n!i: t»!ji*Si>i(4J» i',i jft^i^i'' '44^!i!^^ ' ii'Skfe ^'/,!(ili(fft>/^iS'.1: """S'i'i"'^'""! ii1'!":" :i .'.(•..' >.. ",';, ri. :', "' •
1 ' ' •- - • ..,.,.- :i«»fciA3ijtrtift(i;isi(( (Kbto(Sti4iSt1rt1LtlA''ti.:7l4(i>iO:.:ii^;i'';Ji.,31i>(»i'H'.*s'1. :H:,;JJ:,\.' :':• ,• '.'• .. • '
^j^^r^i!*^^^^:'^^';?^™**$*'**-:Wv'•*!'•": '::'""! '•"• •""'.'":;:-'.' "" -'.'•. •:
"iiilitt f;-'•:^^1".•,';:,,•. < •• •,
lillliw^ • '•'
li'iSiiStt "•• " ,"'' ,',•.'•'
Bi^HTO^b«i«^^.i**(^S^* ^'•fflsiiiviKL'i-.''' !;;-'i:«r«<::;'''';''--: '!•'':,;': '••' •- • - ••
,f ' in*1 'i NIL iijpii, IB,' .,p IIHI|LII'I fcii \t u ,11, i' ,|»v, » ' • ]»|IU " i' i ""i,,i; ? •, ,i', :,' "i
R 'f|! .iBsiSSsiiiflinJi!'1*"1 ,aT:?r'it'WT.iS't lyiH, ^'VSlii .y1:1" i° il"l|ilii:,l,ii'!ii';j,li ;n," i: •• ,,'"':iir
...... Si: ..... j/.tJiK ;••.,»•;, 'it
-------
I,......--..—-..-.-, —.-~, ..,.^
II;,
!CfKsaiBilJfcsUi.:5:SiKl'i5.
iiililiiiitiiiii
*y
i*r
**%**
:sLi
ts
*W« !
•Tnr"'
•aii
"TS,^ itfi-*
-.^?5:''
1*43.^1,^
llHiil
IiagSlfe
Environmental
gcgyped mtoinne"serjes~The8e nine broad cate-_
-- -_—- -_ application Of en.
raoj.tiQO.al .grQUBing was consciously
®IH?y?V*£>Ji
maxirminl ...... interface in related fieos
gicalReseargh
Monttormig
bnergy-t
Research and Qevelopment
UgrfedTo'fhe SPECIAL" REPORTS series This series is
; needs .of specific
Sihe series incTudes proDlBm=orleniea reports, research application
include state-of-the-art
als, user manuals, and reports
. <•-
LlifiM' ':'SSt
sa\aabe to the
tnrpugh the NationaJecnpic^l mtormar
..... Wgma 61
-------
EPA 600/9-80-056
December 1980
URBAN STORMWATER AND COMBINED SEWER OVERFLOW
IMPACT ON RECEIVING WATER BODIES
Proceedings of the National Conference
Orlando, Florida
November 26-28, 1979
Edited by
Yousef A. Yousef
Martin P. Wanielista
Waldron M. McLellon
James S. Taylor
College of Engineering, University of Central Florida
Orlando, Florida 32816
Based on Grant No. R-806715
Project Officer
Robert Turkeltaub
Storm and Combined Sewer Section
Wastewater Research Division
Municipal Environmental Research Laboratory (Cincinnati)
Edison, New Jersey 08817
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for publica-
tion. Approval does not signify that the contents necessarily reflect the
views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
-------
FOREWORD
The Environmental Protection Agency was created because of increasing
public and government concern about the dangers of pollution to the health
and welfare of the American people. Noxious air, foul water, and spoiled
land are tragic testimony to the deterioration of our natural environment.
The complexity of that environment and the interplay between its components
require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem
solution and it involves defining the problem, measuring its impact, and
searching for solutions. The Municipal Environmental Research Laboratory
develops new and improved technology and systems for the prevention,
treatment, and management of wastewater and solid and hazardous waste
pollutant discharges from municipal and community sources, for the
preservation and treatment of public drinking water supplies, and to
minimize the adverse economic, social, health, and aesthetic efforts of
pollution. This publication is one of the products of that research; a
most vital communication link between the researcher and the user
community.
The proceedings contain presentations describing both the impact of
urban stormwater and combined sewer overflow on receiving waters and
methodologies for stormwater impact assessment for stormwater management.
Also included is an edited transcription of the taped workshop.
Francis T. Mayo
Director
Municipal Environmental Research
Laboratory
-------
r
ABSTRACT
The conference on "Urban Stormwater and Combined Sewer Overflow Impact
On Receiving Water Bodies" was held November 26-28, 1979 at Orlando, Florida,
The conference provided a forum for researchers, practitioners and
others to receive an update on the state-of-the-art and to learn about
research findings dealing with stormwater impact. It also served to stimu-
late dialogue among those who are interested in stormwater effects and
control, regarding the implication and applications of current research
results, particularly from those projects supported by the Environmental
Protection Agency, Municipal Environmental Research Laboratory's Storm and
Combined Sewer Program.
The topical areas considered included:
a. Combined sewer overflow control costs. vs. benefits.
b. Impacts on lakes, rivers and estuaries.
c.
d.
Ecological response to stormwater and methodologies for
stormwater impact assessment.
nr
Stormwater management through theousenof receiving water
quality models for planning and abatement methodology.
The proceedings contained herein include the contributions from the
scheduled speakers and an edited transcription of the taped workshop con-
ducted on practical applications of research findings and future research
needs.
iv
-------
CONTENTS
Disclaimer . • . ... • • • .... . • . . . ii
Foreword iii
Abstract • ; . -. • • . •'. iv
Acknowledgments . v viii
1. CONFERENCE SUMMARY .. . :. . . . ,1<.
2. FIRST SESSION - GENERAL OVERVIEW . . ,. . .'.., .7 .
Urban Runoff Receiving Water Impacts - Program Overview and
Research Needs
Richard Field and Robert Turkeltaub . . . . . . . . •; . . . .,;. 8
^sr^.-.t.ij v •.'-••' ..---•'.•...
3, SECOND SESSION - IMPACTS ON LAKES 31
An Assessment of the Impact of Urban Drainage on Eutrophication-
Related Water Quality in Urban Lakes
G. Fred Lee and R. Aiine^Oofies .....'. . . v. . -.. . ; '.' . ..', 32
The Effect of Urban Stormwater Runoff on the Water Quality of
Lake Jackson, Florida • r "
Christian Byrne, C.R. Donahue and W.C. Burnett ... . . .... = 58
A Comparison of Storm-Related Material Loadings to Two Glacial
Lakes from Urban, Wetland, and Agricultural Sources ,
Robert P. Glandon, Frederic C. Payne, Clarence D. McNabb, and
Ted R. Batterson 72
4. THIRD SESSION - IMPACTS ON LAKES AND RIVERS 89
Impacts of Stormwater on a Florida Lake Ecosystem: Effects on
Water Quality and Biota
Eldon C. Blancher, II 90
The Distribution of Sediments and Particulate Contaminants from
Combined Sewer and Storm Drain Overflows in Seattle's Nearshore
Waters
-------
R.D. Tomlinson, B.N. Bebee, S. Lazoff, R.G. Swartz and
D.E. Spyridakis 115
The Ecological Effects of Urban Runoff on Stream Communities
Donald B. Porcella 147
5. FOURTH SESSION - IMPACTS ON RIVERS 167
Analysis of Receiving Stream Impacts on the Milwaukee River
Thomas L. Meinholz 168
Urban Stormwater Impacts on Receiving Streams in North Carolina
E. Ryland Brown and Ross S. Green 197
Dissolved Oxygen Impact from Urban Stormwater Runoff
Thomas M. Keefer, Robert K. Simons and Raul S. McQuivey ... 223
The Impact of Combined Sewer Overflows on the Dissolved Oxygen
Concentration of a Small Stream
Thorkild Hvitved-Jacobsen 245.
6. FIFTH SESSION - IMPACTS ON ESTUARIES 260
Impacts of Intermittent Discharges on Receiving Waters
John L. Mancini 261
The Response of Great Lake Estuaries to Stormwater Runoff
John R. Adams and Stephen M. Yaksich 281
Water Quality and Urban Runoff in Selected Canal Communities
Along Texas Coast
Allen Messenger and Tom D. Reynolds 301
7. SIXTH SESSION - ECOLOGICAL RESPONSE TO STORMWATER 323
The Response of Infaunal Communities to Seasonal Variations in
Wastewater Discharge
Ronald M. Thorn 324
Productivity Responses of Lake Eola Water to Urban Runoff
Harvey H. Harper, III, Yousef A. Yousef, and Martin P.
Wanielista 341
Water Quality and Biological Effects of Urban Runoff on Coyote
Creek
Robert Pitt and Martin Bozeman 371
-------
8. SEVENTH SESSION - METHODOLOGIES FOR STORMWATER IMPACTS ASSESSMENT . 406
Nationwide Assessment of Urban Stormwater Impacts on Receiving
Water Bodies
James P. Heaney, Wayne C. Huber and Melvin E. Lehman 407
Statistics of Advective Dispersive System Response to Runoff
Dominic M. DiToro 437
Continuous Receiving Water Quality Modeling for Urban Stormwater
Management
Miguel A. Medina, Jr . 466
Potential of Stormwater Impacts Based on Comparative Analysis of
Wet and Dry Weather Pollutant Loading
Douglas C. Ammon and Richard Field . 502
9. EIGHTH SESSION - STORMWATER MANAGEMENT 523
The Use of Receiving Water Quality Models in Urban Runoff Pollution
Abatement: Application to Marginal Benefit: Marginal Cost Analysis
Cornelius B. Murphy, Jr., Gregory J. Weller and Dwight A.
MacArthur ....... 524
Methodology For Evaluating the Impact and Abatement of Combined
Sewer Overflows A Case Study of Onondaga Lake New York
Peter E. Moffa, John C. Byron, Steven D. Freedom and John M.
Karanik . . . . . _.C5 ££,f 557
H 1 r ? "J,
A Water Quality Planning Methodology for Urban Areas
Frank!i.n W. (Skip) Ellis and Ronald L. Wycoff 590
10. WORKSHOP ON PRACTICAL APPLICATIONS AND RESEARCH NEEDS FOR RECEIVING
WATER RESPONSES TO URBAN STORMWATER
Session I 618
Session II 631
Session I - Summary
James P. Heaney ,. 650
Session II - Summary
John L. Mancini 651
APPENDIX
List of Conference Participants 653
Vll
-------
ACKNOWLEDGMENTS
We wish to acknowledge with gratitude the assistance of the many indi-
viduals who helped make this conference possible. The financial and technical
support given by the United States Environmental Protection Agency, Municipal
Environmental Research Laboratory, Storm and Combined Sewer Program and the
American Society of Civil Engineers, Florida Section is very much appreciated.
Special acknowledgments are extended to Mr. Richard Field and Mr. Robert
Turkeltaub, EPA, Storm and Combined Sewer Program, Edison, New Jersey for
their invaluable contribution and assistance in the review process and
finalizing the program.
Appreciation is also extended to all speakers for their desire to share
their knowledge and to all participants who made this conference a success.
Special recognition must go to personnel from the College of Extended
Studies and the College of Engineering at the University of Central Florida
who helped arrange for the conference, edit the proceedings and type the
manuscript.
Yousef A. Yousef, Ph.D., P.E.
vm
-------
"URBAN STORMWATER AND COMBINED SEWER
OVERFLOW IMPACT ON RECEIVING WATER BODIES"
CONFERENCE SUMMARY
INTRODUCTION
Urban stormwater and combined sewer overflow (CSO) have been identified
as significant pollutant sources. Runoff from urban areas is usually contam-
inated with organic matter, sediments, nutrients, heavy metals, pesticides,
bacteria and other toxic substances. These pollutants will eventually enter
the adjacent environment in significant quantities depending on many factors
such as antecedent dry period, land use of the drainage basin, social and
economic status, degree of urbanization and volume and type of traffic, in-
dustry and air pollution fallout. The magnitude of these pollutant loadings
has led to a definite national need to study receiving water impacts.
Wet weather conditions and associated runoff may greatly affect the
quality of receiving waters due to shock effects and long-term buildup. The
extent to which the aquatic environment will be affected and courses of reme-
dial action can only be determined by experience and/or experimentation.
Wh'thout experience and knowledge, logical decision must depend on experimen-
tation. Decision makers throughout the country are looking for information
on receiving water impacts in order to preserve the integrity of their streams
and to abide by the goals set forth in PL 92-500 and its amendments.
Conferences provide a forum for effective and rapid dissemination of
available information. Also researchers and planners in all levels of Federal,
State and local government and the private sector are willing to present state-
of-the-art information and to engage in helpful discussions. This fact was
realized by the Storm and Combined Sewer Program, Municipal Environmental
Research Laboratory, U.S. Environmental Protection Agency who encouraged and
supported efforts by the Department of Civil Engineering and Environmental
Sciences, College of Engineering, University of Central Florida to host a
national conference on "Urban Stormwater and Combined Sewer Overflow Impact
on Receiving Water Bodies," November 26-28, 1979 at Orlando, Florida. The
conference was also sponsored by the American Society of Civil Engineers,
Florida Section. This conference drew national and international attention
judging from the list of participants and requests for the proceedings.
CONFERENCE FORMAT
The conference was divided into eight different sessions for formal
presentations on November 26-27, 1979 and was followed by a workshop on
-------
November 28. The opening session set the theme for the conference. Myron
Tiemens, EPA, Washington, D.C., started the technical presentations with a
presentation on "Combined Sewer Overflow Control: Costs vs. Benefits"
(no formal paper). His presentation was followed by general remarks (no
formal paper) from Dennis Athayde, EPA, Washington, D.C.; then Robert
Turkeltaub and Richard Field, EPA, Edison, New Jersey followed with a pre-
sentation on "Urban Runoff Receiving Water Impacts - Program Overview and
Research Needs." The other sessions covered studies and analysis demon-
strating methodologies for stormwater impacts assessment and stormwater
management. The workshop was divided into two sessions to discuss practical
applications and research needs for receiving water responses to urban
stormwater.
The conference was attended by approximately 180 persons in addition to
40 speakers and moderators. The magnitude and complexity of the problem was
evident from discussions that took place by chemists, biologists, engineers,
economists, management and planning personnel and others who participated in
the conference. The sessions were very well attended and good interaction
between speakers and attendees was developed. This voluminous proceedings
represents the contribution from participating speakers and edited copy of
the taped workshop.
SUMMARY OF PRESENTATIONS
During the first general session, Myron Tiemans, EPA, Washington, DjQu/i
indicated that secondary treatment requirements will be met before a CSQ* ortv?
remedial program is initiated and funding will be allocated for polluti013De-
duction only. Also, planning money for CSO will be accelerated by construe-"
tion grants after they are provided better cost estimates, better enfor^e-;^
ment policy and better plan review. He was followed by Dennis Athayde^EPAy
Washington, D.C., who stated that the U.S. EPA's National Urban Runoff Program
(NURP) will be determining effects of urban runoff on receiving waters. The
first session was ended by Robert Turkeltaub, EPA, Edison, New Jersey, who
presented a comprehensive paper on the status of the U.S. EPA's Storm and
Combined Sewer Program receiving water impacts efforts and future research
needs.
The second technical session started with G. Fred Lee, Colorado State
University, Fort Collins, Colorado, who presented an approach for assessing
the water quality significance of chemical contaminants in urban lakes.
Dr. Lee stated that the OECD (Organization for Economic Cooperation and
Development) results should be used to describe the nutrient - load response
relationship for urban lakes and most importantly, should be used to estimate
the magnitude of the water quality improvement resulting from reduction of
nutrient loads to phosphorus limited lakes. Also, an environmental hazard
assessment based on aquatic toxicology and environmental chemistry - fate
information should be conducted to define the magnitude of the impact of
toxicants on beneficial uses of receiving water bodies. This was followed
by Christian J. Byrne, Florida State University, Tallahassee, Florida,
who investigated the petroleum hydrocarbon content in urban stormwater and
initiated studies to model the hydrocarbon budget of Lake Jackson, Florida.
The second session was concluded by Robert P. Glandon, Michigan State
-------
University, East Lansing, Michigan, who showed a comparison of storm-
related phosphorus and nitrogen loading to two glacial lakes in Lake
Michigan drainage basin, from urban, wetland and agricultural sources.
The third session started with Eldon C. Blancher II, Marine Environmental
Sciences Consortium, Dauphin Island, Alabama, who presented data indicating a
direct impact of stormwater runoff on water quality and subsequent changes, in
biota of Lake Conway ecosystem located in Orange County, Florida based, on
Secchi disk transparency, chlorophyll a and total zooplankton numbers.
Richard D. Tomlinson, Municipality of Metropolitan Seattle, Seattle,
Washington, discussed the distribution of sediments and particulate con-
taminants from combined sewer and storm drain overflows in Lake Washington.
This session was concluded by Donald B. Porcella, Tetra Tech, Inc.,
Lafayette, California, who presented a literature review of the ecologi-
cal effects of urban runoff on streams. He discussed the concept" that:" ..
large scale variations and instabilities of stream pollutant input a'fnd
concentration would result in greater impact to stream communities than
steady inputs.
j,
Thomas L. Meinholz, Ecolsciences, Inc., Milwaukee, Wisconsin, started
session four and presented an analysis of receiving stream impacts on5the
Milwaukee River. He reported tremendous DO sags in the lower reaches of
the river following runoff events due to scouring of the bottom sediments '
by submerged combined sewer outfalls. He was followed by E. Ryland Brown,
North Carolina Department of Natural Resources, Raleigh, North Carolina,
who predicted that, under present conditions, almost all urban streams in
-North Carolina will be unable to meet the 1983 water quality goals.
Thomas N. Keefer, The Sutron Corporation, Arlington, Virginia, examined
records from 104 water quality monitoring sites throughout the country.
He -concluded that the stream flow increases and the dissolved oxygen diur-
nal cycle disappears following a storm event. The minimum DO drops from
1 to 1.5 mg/1 below the minimum values observed during steady flows and
remains constant there for periods ranging from one to five days.: 'Thorkild
Hvitved-Jacobsen from Denmark talked about immediate and delayed effec:£fse
on the DO concentration in a stream following a storm event.
The fifth session covered impacts of storm and combined sewer overflow
on estuaries. John Mancini, Manhattan College, Bronx, New York, presented
methodologies to illustrate impacts of intermittent CSO discharges o'n bathing
beach areas of New York City and discussed the fate and effect of inter-
mittent discharges of toxics. Then John R. Adams, Army Corps of Engineers,
Buffalo, New York, discussed the sources of pollution in the estuaries reach
of the Maunee River. He outlined the problems and complexities of chemical
measurements of water quality in estuarine systems of the Great Lakes during
varying hydrologic conditions. This session was concluded by Allen1 £."-
Messenger, Austin, Texas, who presented data to demonstrate deterioration
in water quality of canals in the Galveston Bay Area resulting from storm-
water runoff and possible resuspension of bottom sediments.
The sixth session started with Ronald M. Thorn, University of Washington,
Seattle, Washington, discussing the response of infaunal communities to
seasonal variations in CSO discharges. He indicated that the communities
-------
r
at the sites nearest the overflow appeared to be primarily affected by
intense scouring and deposition, whereas those communities further away
appeared to respond largely to input of organic debris. He was followed
by Harvey H. Harper, University of Central Florida, Orlando, Florida, who
presented results from algal bioassay on impact of runoff on Lake Eola
water. He demonstrated the impact of nutrients and toxic substances re-
leased to Lake Eola in urban runoff water. The session was concluded by
Robert Pitt, Woodward Clyde Consultants, San Francisco, California, who
presented his study to demonstrate biological effects of urban storm run-
off on Coyote Creek, San Jose. Fish, benthic macroinvertebrate and
attached algae were used as indicators and concluded that the urbanized
portion of the creek is significantly degraded by urban stormwater.
The seventh session on methodologies for stormwater impact assessment
was started'by James P. Heaney, University of Florida, Gainesville, Florida,
who indicated that well-documented cases of receiving water impacts are
scarce. His nationwide assessment will be used as a basis for devising
simple criteria for analyzing an urban area to determine whether a poten-
tial impact does or would occur. He was followed by Dominic M. DiToro,
Manhattan College, New York, who presented methods to evaluate the statis-
tical behavior of receiving water concentrations that result from runoff
events. These methods are specially useful for screening many alternative
treatment schemes that abate runoff mass discharge. He was followed by
Miguel A. Medina, Jr., Duke University, Durham, North Carolina, who pre-
sented Level III - Receiving Water Quality Model. This model has been .. .-^o
developed to permit preliminary planning and screening of area-wide urban
wastewater treatment alternatives, in terms of frequency of water quality
violations and more traditional approaches such as dissolved oxygen pro- •-
files. This session was concluded by Douglas C. Ammon, EPA, Edison, •••• s:
New Jersey, who summarized much of the stormwater characterization data
for trace contaminants and several conventional parameters. He also
identified the significance of each contaminant.
Session eight discussed stormwater management tools. Gregory J. Welter,
O'Brien and Gere Engineers, Inc., New York, New York, discussed the use of
receiving water quality models in the analysis of the Syracuse, New York;
Rochester, New York and the District of Columbia combined sewer facilities
planning activities using the marginal cost - marginal benefit approach.
He was followed by Peter E. Moffa, Stearns and Wheler, Cazenovia, New York,,
who presented a general methodology for the evaluation of the impact and
abatement of CSO on receiving waters. This methodology was developed from
experience with Onondaga Lake in Central New York. A 27-segment, three-
dimensional, dynamic water quality model of the lake was developed with
capability of predicting enteric bacteria, dissolved oxygen, nutrients and
toxic materials as a function of pollutant load. The model made it possi-
ble to predict DO deficit and fecal coliform content following treatment
alternatives for CSO. The technical sessions were concluded by Franklin W.
(Skip) Ellis, CH2M Hill, Reston, Virginia, who presented a two-phase
approach to water quality planning for urban areas. Phase I involves
the use of the continuous stormwater pollution simulation system (CSPSS)
-------
for economic optimization for various levels of pollution abatement.
Phase II will produce a description of the optimal mix of control alter-
natives, the total plan costs and the receiving water quality response.
SUMMARY OF WORKSHOP -:,;.,
'—- •—.!.— n— •• - II. IP I - '• ••• - •. •» " •• I '-• - k- „.' .,'
Upon completion of the technical presentations, arrangements were made
for follow-up discussions and interactions between speakers and attendees ,
during a two-session workshop held November 28, 1979. Each session lasted
for one and one-half hours. It started by opening statements from five
panelists for five minutes each followed by discussion from the floor. , -'
A summary of important comments presented during the workshop follows:
1. There are very limited cases where evidence of stormwater and CSO
impacts on receiving water bodies is adequately documented. Also,, 1;h£re, is
no quality assurance in data collected. A simple and reliable qual~Cty con-
trol methodology needs to be developed.
2. Chemical standards and pollutional loadings alone are not adequate
criteria to determine impacts. Biological indicators need to be used,,, and,
biological sampling must represent various levels of food chain. .. ~."•'''•
UV '' .1
3. Bioassay methodologies need to be improved and expanded in order to
be able to predict biological response as related to.changes in chemical
constituents. •
4. Fate and effects of toxic substances and heavy metals released to
receiving streams and their relationship to public health problems need to
be investigated.
5. Water quality problems should be defined as those which interfere
with stated beneficial water usage. The magnitude and frequency of ;the
problem should be investigated and supplemented with a realistic cost
benefit analysis. Therefore, simplified methods of assessing benefits-
need to be developed.
6. Frequency-duration-response curves for stormwater and CSO release
to receiving water need to be developed and measured in terms of impairment
of beneficial use. r.~~
7. Modeling of stream impacts is a complex process which needs better
identification of parameters involved. Fundamental research is needed to
define deoxygenation and reaction kinetics, transport and fate of pollutants,
sediment transport and its impact on benthic uptake and other geophysical and
hydrodynamic processes. :?'.:•':.
8. Simplification of existing models is desired and guidelines for
better matching the model to data base are needed. Models based on statisti-
cal analysis could be useful and a process for model calibration and verifi-
cation is needed.
-------
r
9. Socio-political constraints should be considered in assessing re-
ceiving water impacts.
In conclusion, the conference was well attended, good presentations were
made and the response was encouraging. Stormwater impacts will continue to
attract more attention and the need for continued investigation will increase.
Similar conferences are most beneficial for fast dissemination of available
information.
-------
First Session
GENERAL OVERVIEW
Moderator: Carl Brunner
EPA, Cincinnati, Ohio
-------
URBAN RUNOFF RECEIVING WATER IMPACTS:
PROGRAM OVERVIEW AND RESEARCH NEEDS
by
Richard Field* and Robert Turkeltaub**
Environmental Protection Agency, Edison, New Jersey
ABSTRACT
Recieving water impacts are a major national concern. We are spending
huge sums of money on secondary treatment plants, meanwhile major culprits,
stornwater and combined sewer overflows, are still uncontrolled. To attain
the goals set forth in PL 92-500 and PL 95-217 of abating pollution and
achieving water quality standards in an economical and efficient manner those
analyzing, planning and designing controls must have a thorough understanding
of the impact of pollutants on receiving waters. Receiving water impacts are
the bottom line justification for funding countermeasure campaigns and the
passage of abatement legislation. This conference will provide a forum for
the attendees to acquire first hand knowledge of the state-of-the-art and to
consider ongoing and recently completed research.
Data on the environmental impacts of urban stormwater and combined
sewer overflow are being gathered by projects of the Storm and Combined
Sewer Program, (SCSP) of the Municipal Environmental Research Laboratory
(MERL) as a first step in developing control needs and a methodology to quan-
tify pollutant stress and evaluate the impact in relation to receiving water
standards and desired uses. This paper will contain a brief history of our
ongoing projects. The projects will be briefly described including project
objectives and an outline of significant results to date. Also, future
Program needs will be discussed and areas in which we anticipate concentrating
our efforts will be outlined.
BACKGROUND
Recognition of the significance of stomwater-induced water pollution
has been slow. Current interest is largely attributable to the efforts of
*Chief, Storm & Combined Sewer Section, Municipal Environmental Research
Laboratory (Cincinnati), U.S. Environmental Protection Agency, Edison, New
Jersey 08817
**Staff Engineer, Storm & Combined Sewer Section, Municipal Environmental
Research Laboratory (Cincinnati), U.S. Environmental Protection Agency,
Edison, New Jersey 08817
8
-------
the USEPA Storm and Combined Sewer Research, Development, and Demonstration
Program.
Only within the past 20 years, has it been recognized that waters
discharged from separate storm sewers contain pollutants. Eain falling on
an urban area picks up pollutants from the air, dusty roofs, littered and
dirty streets and sidewalks, traffic by-products (gasoline and oil drippings,
tire residuals, brake ware, vehicular exhaust), metallic corrosion, hazar-
dous spill materials and chemicals applied for fertilization, control of ice,
rodents, insects, and weeds.
Urban Runoff Characterization
Figure 1 illustrates representative strengths of wastewaters. The
average five-day biochemical oxygen demand (6005) concentration in combined
(domestic and storm) sewer overflow is approximately one-half the raw sani-
tary sewage 10)5. However, storm discharges must be considered in terms of
their shockloading effect due to their great magnitude. A conmon rainfall
can produce flow rates up to 100 times dry-weather flow. Even separate
stormwater is a significant source of pollution, having solids concentrations
equal to or greater than untreated sanitary wastewater, and BC^'s approxi-
mately equal to secondary effluent. Bacterial contamination of separate
stormwater is two to four orders greater than concentrations considered safe
for water contact.
Microbiological studies of storm runoff have shown consistent recoveries
of pathogenic organisms (1). This indicates that all types of urban runoff,
in general, may be hazardous to health.
Representative Loads
From 40 percent to 80 percent of the total annual organic loading en-
tering receiving waters from a city is caused by sources other than the
treatment plant. During a single storm event, 95 percent of the organic
load is attributed to wet-weather flow sources. The runoff of toxic
pollutants, particularly heavy metals, and petroleum hydrocarbons is also high.
Potential Impacts
Approximately one-half of the stream miles in this country are water
quality limited and 30 percent of these stream lengths are polluted to a
certain degree with urban runoff. Hence, generally speaking, secondary
treatment of dry-weather flow (DWF) may not be sufficient to produce required
receiving water quality; and control of runoff pollution becomes an alterna-
tive for maintaining stream standards. For example, Figure 2 illustrates
that if Durham, North Carolina provided 100 percent treatment of municipal
wastewater, the total annual reduction of ultimate BOD and suspended solids
would only be 59 percent and 5 percent, respectively. Durham has a separate
system. Combined systems offer a greater pollutional impact since additional
loads come from domestic wastewaters, dry-weather sediment washout, and more
impervious and populated lands.
-------
200
PI
$s
'Pza
A> 4J<1
Wn
zoom
%
&
$
fa-
.•.•>•*•
3
ES RAW
^ COMBINED
CH STORM
- . .
BOD
SS
DO
5x10'
4
m
^M
40
10
_|2J RAW
^ COMBINED
II STORM
^ — ,
TOTAL COLIFORM TOTAL TOTAL
MPN/100ml NITROGEN PHOSPHORUS
Figure 1. Representative Strength of Wastewaters (Flow Weighted Means in
mg/1)
10
-------
LU
tc
UJ
X I
< CO < (O
UJ O UJ Q
D
o
O
o
in
co
CO
O
m
D
SSVW INVinilOd 1VONNV iVlOl %
cd
TD
ni
o
c
cd
-p
D
o
cu
-p
td
0)
Q
c
ni
S-,
(D
,c
•P
cfl
(D
CO
I O
-p J-,
(U nJ
:s u
cd -P
-P SM
o o
CM
0)
S-,
bO
•H
11
-------
Now that urban runoff has been identified as a significant pollutant
source the emphasis has been refocused from direct analysis of urban runoff
to substantiating its impacts on receiving waters and developing source
control methodologies based on these impacts. Identified and quantified
impacts and their associated impairment of designated beneficial uses of
receiving waters are fundamental factors providing justification for funding
stormwater and CSO countermeasure campaigns and passing abatement legislation.
For instance, in order for a CSO control project to qualify for Federal
funding a careful review must have shown that out of all the alternatives
considered, the selected alternative is cost effective and required to protect
the beneficial use of the receiving water, even after industrial effluent
limitations and secondary treatment for dry-weather flows are achieved (2).
Very little information of this kind can be found; therefore, several
years ago the U.S.EPA Storm and Combined Sewer (SCS) Program began a modest
effort to fill this data void. Ties between receiving water quality and
stormwater discharges must be clearly established and delineated. Quanti-
fication of the impairment of beneficial uses and water quality objectives
by such discharges is a major goal. Results from earlier SCS Program projects
and limited results from others are now available. The basic objective of
this paper is to present an overview of the SCS Program's receiving water
impacts subprogram. Many of the Program'siprojects will be described in
detail during the course of this conference.
RECEIVING WATER QUALITY IMPACTS
Data on the environmental impact of urban runoff are being gathered as
a first step in developing methodology to quantify pollutant stress and eval-
uate the impact in relation to receiving water standards and desired uses.
Dissolved Oxygen Depletion
The classical problem related to organic pollution of receiving waters
is the consumption of instream oxygen by the bacterial breakdown of organic
material. The resulting low levels of oxygen will destroy sensitive species
of fish and aquatic organisms and may cause anaerobic conditions which pro-
duce objectionable end products. Accordingly, the SCS Program's first attempt
at evaluating receiving water impacts was a project by the Sutron Corporation
(3) wherein data from existing rain gages, flow and water quality monitors
in and downstream of urban areas was assessed in an attempt to correlate
dissolved oxygen (DO) depletion and urban runoff.
Although sites with specific instances of standards violations were
located, no conclusions could be drawn on a national basis. Several other
correlation attempts in Ohio and Indiana streams also proved fruitless
(H, 5).
Except for one occurrence in the Chattahoochee River, after a July thun-
derstorm, when the DO reached a low of 1.5 mg/1 the Chattahoochee project (6)
also failed to ascertain adverse DO impacts despite BOD^ and other pollutant
concentrations in the CSO ranging from 1 to 10 times that in the receiving
streams.
12
-------
The 1.5 mg/1 DO level recorded in the Chattahoochee River is.an indication
of the potential negative impacts from oxygen demanding material in. urban
storm runoff. At this point although the SCS Program has not had complete
success in ascertaining DO impacts from urban runoff the Program should not
attempt to write off DO demand as being a wet-weather related impact. Based
on annual mass balance determinations urban wet-weather organic loads are the
same order of magnitude as dry-weather loads as indicated by Table 1, and 10
times greater during storm periods (7, 8). However, a large fraction of the
oxygen consuming material is associated with settleable and separable solids
which in storm flow are usually larger and in a more non-biodegradable state
than in sanitary effluent therefore the deoxygenation rate tends to be lower
and time delayed. The fate of this material is related more to the hydro-
dynamics of the specific receiving water and to the specific wet-weather
oxygen transfer rates rather than to the advective transport and lumped and
conventional rate assumptions used in traditional DO analysis.
Table 1. National Annual Urban Wet- and Dry-Weather Flow
and COD Load Comparisons*
TYPE
COMBINED
STORM
UNSEWERED
TOTALS
PERCENT
OF
DEVELOPED
AREA
14.3
38.3
47.4
100
ANNUAL DWF**
BOD5
MIL LB.
340
710
310"
1330
COD
MIL LB.
910
1890
830
3630
ANNUAL WWF"
BOD 5
MIL LB.
880
440
360
1640
COD
MIL LB.
2640
2500
2250
7390
PERCENT WWF
BOD5
72
36
54
55
COD
74
57
73
67
'ASSUMING
cso
URBAN
STORMWATER
DRY-
WEATHER
BOD5
(MG/L)
100
20
30
COD
(MG/L)
300
125
80
'LB = 0.454 KG
13
-------
Part of the problem in ascertaining DO impacts lies in locating urban
areas free from all interferences except urban runoff. And, as Figure 3
indicates, part lies in trying to derive data from sites set up for con-
ventional dry-weather monitoring applied for unsteady state flow regimes
caused by storms. Also, storm related changes in deoxygenation rate,
stream velocity and reoxygenation rate are variables which can cause the DO
sag to occur at remote locations as compared to relatively steady state
dry-weather conditions. And still another factor is money. Receiving water
studies are expensive and the SCS Program and others can ill afford to spend
substantial sums from their limited budgets on one project.
D.O.
DISTANCE (TIME)
RECEIVING WATER
CSO
WATER QUALITY MONITORS
URBAN AREA
Figure 3- Comparative Locations of DO Sag and Water Quality Monitors
14
-------
In Milwaukee, these problems were somewhat overcome (9). The City
was desirous of providing the necessary receiving water impact analyses
for evaluating CSO pollution and control and the SCS Program was able
to interface with their 201, step 1 construction grant.
River sampling results provided strong evidence of CSO impacts on
intensifying DO sag. The study has recently been completed and it was
found that resuspension of the sediments caused rapid DO sags in the
Milwaukee River after CSO events. Combined sewer overflows were found
to contribute 40 to 50 percent of the annual BOD loading to the sediments
and the sediment oxygen demand (SOD) of the disturbed sediments was
100 times greater than SOD rates in situ (10 g 02/m2/day vs. 1000 g
02/m2/day). Figure 4 illustrates the marked effect that the disturbed
sediment has on the-DO,in the river. The I-beams represent the range
of the observed real river data and offers verification of model predictions
using disturbed SOD rates.
14.0-
12.0-
o>
10.0-
UJ
CS 8.0-
X
o
Q
LU
6.0-
O 4.0-
W
O
2.0-
0.0,
PREDICTED D.O. WITH
TIME VARYING S.O.D.
PREDICTED D.O. W/O
TIME VARYING S.O.D.
OBSERVED D.O. RANGE
I
2400 1200 2400 1200 2400 1200 2400 1200 2400
i 4 AUGUST I 5 AUGUST I 6 AUGUST I 7 AUGUST I 8 AUG
TIME
Figure 4. DO Storm Profile - Milwaukee River, Milwaukee, Wisconsin
15
-------
Based on results identifying sediment as a major problem, the SCS
Program undertook a project to track the fate of settleable materials
discharged to flowing streams (10). The project surfaced limitations
with the current state-of-the-art on transport modeling.
Conventional receiving water quality models exclude sediment transport
while existing sediment transport models only consider discrete sediments
and not biochemical reactions. Furthermore, the present data base is
inadequate to validate these sediment transport models. For them to be
useful in wet-weather discharge analysis, it will be necessary to interface
the sediment routines with the runoff routines and to build-in the
appropriate biochemical reactions. As indicated by Figure 5, the questions
boils down to the level of effort and cost-effectiveness in developing
a more complex model and data base.
VERIFICATION DATA
DEVELOPMENTAL DATA
VERIFICATION DATA
SEDIMENT TRANSPORT,
DEPOSITION & SCOUR ROUTINES
+
LINKAGE WITH RUNOFF QUALITY MODEL
+
BIOCHEMICAL REACTIONS
Figure 5. Receiving Water Sediment/Pollutant Impact Analysis
Pathogen Concentrations
Excess concentrations of bacterial indicator organisms in urban
runoff will hinder water supply use, recreational use and fishing/shell
fishing use of the receiving water. A common bacterial standard for
recreational use of water is a total coliform concentration of less
than 1000 organisms/ 100 ml and a fecal coliform concentration of less
than 200 organisms/100 ml. However, coliform are merely indicating
organisms and not direct pathogens. Their use as .an indicator of human
enteric contamination in storm runoff waters is not valid because some
coliform in this flow originates from soils and animal fecal matter
inherent in storm runoff. Therefore, the SCS Program supported the
Johns Hopkins University study (1) whose conclusions showed that the
frequency of occurrence of human pathogenic organisms in storm flow
relates to cross contaminations from sanitary sewage. All stormwaters
were found to contain disease causing viruses and bacteria because
none of these waters are totally separated from sanitary or other contaminating
waste sources. As a result some pathogens were even found in the
background waters used for control observations.
16
-------
The aforementioned Milwaukee project (9) found that CSO is the major
source of fecal coliforms in the Milwaukee River and that removal of these
overflows will nearly eliminate fecal coliform standards violations in the
river.
In a study conducted at Syracuse, New York (11) the only water quality
parameter which was significantly and adversely affected by CSO into Lake
Onondaga was fecal coliform. As Figure 6 illustrates, of the 65 artnUal CSO
producing events, 38 resulted in discharges of significant magnitude to cause
violation of the bacterial standards in the contact recreational zone of the
lake. A hypothetical one-year, two-hour storm would result in bacterial
numbers in excess of the standard of 200 cells/100 ml in all "upper" segments
of the lake, and a maximum concentration of greater than 11,000 fecal coliform
cells/100 ml in the northern basin where contact recreation is anticipated.
A storm of this magnitude would cause violations of coliform standards for
a period in excess of three days. However, it is probably not necessary to
treat a one year frequency storm. It can be seen from Figure 6 that treat-
ment provided for a two month frequency storm will eliminate 82 percent
of the coliform violations. The cost of providing additional treatment for
the larger storms at the so-called knee of the curve is disproportionately
large compared to the rapidly decreasing benefits achieved much beyond this
point on the curve. It should be emphasized that the concept of dispro-
portionate returns is a decision-maker's tool and that the bottom line lies
with the desired frequency of water usage at the local level.
u
U
UJ
EC
rr
3
O
S
EC
10 20 30 40
NO. WATER QUALITY VIOLATIONS/YR.
FOR STORM INTENSITIES < RECURRENCE INTERVAL
Figure 6. Water Quality (Fecal Coliform) Violations Vs. Storm
Recurrence Interval - Lake Onondaga, Syracuse, New York
17
-------
At Lake Washington, human viruses were readily detected in CSO's (12).
Based on literature indicating that viruses present at detectable levels
can cause infection CSO may be a health hazard for people using Lake Washing-
ton beaches soon after overflows.
Nutrients
Surface water bodies with long detention times, such as lakes and
estuaries, tend to concentrate nutrients, organics, and metals in both
supernatant and bottom muds. Under anaerobic conditions, boat mixing,
concentration changes, and others, these pollutantsTcah be resuspended
and become available for plant growth or a depletion of DO.
In Lake Eola, Florida, urban runoff was found to be the sole source of
lake degradation (13). Figure 7 depicts the Lake Eola watershed. Urban
runoff is the only flow entering the lake. Phosphorous concentrations in
the runoff were found to significantly increase algal productivity. Table 2
indicates the lake's algal growth response to urban runoff. Based .on the
first phase recommendations of the SCS Program's project, the 314 Clean
Lakes Program is funding a phase two restoration.program. It is intended to
jointly fund a phase three post restoration evaluation.
Table 2. Algal Response to Urban Runoff - Lake Eola,
Orlando, Florida
Control
Lake Water
+ 5% Runoff
Lake Water
+ 10% Runoff
Lake Water
+ 25% Runoff
Lake Water
•*• 50% Runoff
Lake Water
+ 75% Runoff
Lake Water
+ 100% Runoff
Average Chlorophyll a Concentration (mg/m^)
Initial
11.7
11.5
11.7
11.7
11.7
11.7
11.7
4 Days
12.1
25.2
38.9
64.5
33.9
22.4
23.3
7 Days
6.5
26.7
38.4
74.2
41.8
16.6
16.2
8 Days
7.1
30.6
43.3
91.6
44.2
20.7
14.5
9 Days
5.0
28.5
38.6
84.7
44.2
16.8
15.3
14 Days
1.3
17.7
31.5
68.6
15.8
7.1
3.7
18
-------
2
CO
cd
"d
•H
^1
O
fi
cd
t-H
-------
Biological Investigations
To develop an assessment of the environmental impact of urban storm-
water runoff requires a comprehensive in-depth analysis of water quality and
the biological community in the receiving stream. Not only the physical and
chemical characteristics of the receiving water body and its sediment should
be examined, but also the aquatic and benthic organisms should be analyzed.
Such an investigation is going on in San Jose, California (14) where
preliminary sampling found that the non-urbanized section of Coyote Creek
supported a diverse assemblage of fish and benthic macroinvertebrates as
compared to the urbanized portion which was completely dominated by pollution
tolerant mosquito fish and tubificid worms. Figure 8 illustrates this point.
More pollution tolerant algae also dominated the urbanized reaches of the
creek. Similar results were found in. the previously mentioned Lake Washington
project (12) where bottom organisms (aquatic earthworms) near storm and
combined sewer outfalls were more pollution tolerant relative to organisms
away from these outfalls. Aquatic earthworm numbers and biomass was found
to be enhanced within .the zone of influence of the monitored combined sewer
and storni drain outfalls in Lake Washington.
STATIONS (relative locations)
Figure 8. Abundance of Benthic Taxa
San Jose, California
20
- Coyote Creek,
-------
For example, at one combined sewer outfall•the-biomass: enrichment-
amounted to 82 percent of the background control _value and aquatic earthworms
constituted 90 percent of the biomass. ' , "
r , . , ^ , . , . , , ,-..,,_, (. ..--•;.- f ,l( s
• • - -•','.--••,<•'
Toxicity . .'..!."'""
Toxicity problems can result from minute discharges of metals, pesti-
cides and persistent organics, which may exhibit a subtle long-term effect
on the environment by gradually accumulating in sensitive areas.
The SCS Program has been analyzing for toxic materials at the street
surface', in urban runoff and CSO's for more' than ten years. This has resulted
'in a large data base that identified urban runoff .as a principal source'of
toxic pollutants. For example, New York Harbor receivers jnetals from 'treatment
plant effluents;' combined sewer discharges; 'separate storm sewer"..discharges;
and untreated wastewater'(15). As .seen .in Table 3, urban runoff is.the major
contributor 'of hea'vy metals; to the Harjbb'r. ';;_ ''"'' ''',''.. ..',
Table 3- ' Metals Discharged to the''Harbor Fr6m''New; jfork^City'Sdurpes, _ .
= -..-••-$.
PLANT EFFLUENTS
RUNOFF" ,
UNTREATED WASTEWATER
TOTAL WEIGHT (LB/DAY)
WEIGHTED AVERAGE CONCENTRATION (MG/L)
*>,., -, V -
1.410
1,990
980
4,380
0.25
,y fi ,- r*,r-i ::. "
780
690
570
2.040
0.12
930
650
430
2,O10
0.11
f -.
2,520
6,920
1,500
10,940'
0.62
- ' r* .1
,95
, :,no
-60
;265
'* 0.015
"IN REALITY, SHOCKLOAD DISCHARGES ARE MUCH GREATER.
•; ; it. • -1; ;,". i'< ^>—..--»» ' '"•'-",, >,'''
Table 4 shows the. to-tal annual mass of selected ,;constituenfes from a
combined sewer and storm overflow point in Seattle (12). -: A high percentage
of the heavy metals ;and toxic materials is associated wit-h the suspended
solids or particulates which may tend to concentrate"in the sediments. - This
can help us in terms of treatment since it is easier to separate pollutants
attached to suspended solids.
21
-------
Table 4. Total Vs. Particulate Mass From a Combined Sewer and Storm
Sewer Overflow Point - Lake Washington, Seattle, Washington
SUSPENDED SOLIDS
Cu
Hg
Pb
Zn
Al
ORGANIC C
TOTAL P
OILS AND GREASES
CHLORINATED HC
SELECTED
TOTAL"
MASS(KG)
3920
1.33
.026
1.46
3.20
66.0
458
45.1
330
ND
CSO POINT
PARTICULATE*
MASS(KG)
3920
1.04
.023
1.01
2.20
63.4
193
12.8
NA
.460G
SELECTED STORM
DRAIN POINT
TOTAL'
MASS(KG)
2238
1.16
ND
6.04
2.74
97.2
299
8.73
113
ND
PARTICULATE"
MASS (KG)
2238
.745
ND
5.32
1.76
93.9
168
4.60
NA
.388G
NA: NOT APPLICABLE ND: NOT DETERMINED
"KG=2.2 LB.
Sediment samples were analyzed for metals, organic carbon, phosphorous,
chlorinated hydrocarbons and pplychlorinated biphenols (PCB's). As can be
seen from Figure 9, a composite index to assess wet-weather impacts was 16
times the minimum background control value. Also, pesticide levels in
sediments along the Seattle shoreline of Lake Washington were up to 37 times
background concentrations. The measured pesticide content of the sediments
was predominantly DDT and its degradation products. Although DDT is no
longer used this flags a potential problem with the use of newer toxic
chemicals and of their long-term fate and transport.
In the previously mentioned San Jose project (14) urban sediment compared
to non-urban sediment from Coyote Creek contained higher concentrations of
lead (10 times greater), arsenic (9 times greater) BOD5 (up to 4.4 times
greater) and ortho phosphates (up to about 4 times greater). Significantly
greater concentrations of high molecular weight hydrocarbons and oxygenated
compounds were also found in the urban samples.
Lead concentrations in urban samples of algae, crawfish and cattails
were two to three times greater than in non-urban samples, while zinc con-
centrations were about three tim.es the non-urban concentrations. Bio-
accumulation of lead and zinc in the organisms compared to water column
concentrations was at least 100 to 500 times greater.
22
-------
tt 70
g 65
8 60
"• 55
Z 50
LU
2 45
O 40
E 35
5 30
£ 25
LU 20
UJ 10
CO
5
CSO STORM CONTROL
DRAIN SITE
J
ORGANICS i METALSj CH's i
— - .p-^ —-''PCB'S ^
Figure 9. Illustration of Urban Sediment Enrichment - Lake Washington,
Seattle, Washington
23
-------
Petroleum enters urban runoff from many dispersed sources — motor
vehicles being the most significant. It may be attributed to spills, leaks
and disposal of motor vehicle lubricants, antifreeze and hydraulic fluids.
Petroleum hydrocarbons, particularly the polynuclear aromatics have
been shown to be carcinogenic and mutagenic. At New York City's Newtown
Creek treatment plant (15) 24,000 gallons (90,840 liters) of oil and grease,
[equivalent to a moderate harbor spill (16)] were bypassed during one four
hour storm. The SCS Program's Jamaica Bay study (17) found that 50 percent
of the hexane extractable material contributed to the Bay is due to CSO's.
It is clear that further investigation into receiving water impacts due to
urban runoff warrants consideration.
A recently funded project (18) with Onondaga County is screening urban
runoff and CSO for bacterial mutagens and is assessing the mutagenic/carcin-
ogenic potential of the toxicants.
Recently, the SCS Program arranged with EPA's Region II Surveillance
and Analysis Division to analyze for priority pollutants in urban samples
supplied from ongoing projects. Results from a few locations show that a
significant amount of priority pollutants are present in urban runoff.
In FY80, the Program is looking to supplement this with extramural
assistance. In any case, in all appropriate projects, priority pollutant
runs are taken to initially identify toxic substances for source identifi-
cation and source control studies.
CONTROL METHODOLOGY BASED ON IMPACTS
The optimal goal of an urban stormwater pollution control study is to
evaluate the relative impact of runoff and CSO's on the receiving waters
and decide what control alternatives would be most cost effective in reducing
wet-weather pollution. Impact's are often site specific and the extent of
the problems will depend heavily on local conditions, such as rainfall
quantities, point sources of pollution and their treatment, land use, and
the sensitivity of the receiving water.
In an effort to catalog existing data, case studies of documented
receiving water impacts are being reviewed in a project with the University
of Florida (19). The project is expected to identify the types of impacts
requiring further verification and the research needed to quantify receiving
water stresses.
The concept for a simplified continuous receiving water quality model
has been developed during the nationwide evaluation of CSO's and urban
stormwater discharges project (7) and refined into a user's manual by a
subsequent project (20). This model, termed Level III - Receiving, permits
preliminary planning and screening of areawide wastewater treatment alterna-
tives in terms of frequency of water quality violations and using more
traditional approaches such as DO profiles. Figure 10 represents the type
of analysis facilitated by the model.
24
-------
100.
WET-II (75% Rem) OR
CWET-I (PRIM)/LM
J V\ (75% Rem)
i »
O «>
UJ 0)
oc o
uj (0
CO
5
H
A|
\N
\-*-WET-1
V.(25% Rem)
\\
D.O. (mg/l)
CONTROL
ALTERNATIVES
EXISTING
TERTIARY
WET-I (PRIMARY)
WET-II (ADV)
WET- 1/ LAND MGMT.
% BOD REMOVAL
DRY WEATHER
85
95
85
85
85
WET WEATHER
0
0
25
75
75
COST
($x106)
—
6
1
6
3
Figure 10. Hypothetical Example Solution Methodology
25
-------
r
By this analysis a truer cost-effectiveness comparison can be made based
on total time of impact to receiving water and associated abatement costs.
For example, if a 5 mg/1 DO is desired in the receiving water 75 percent of
the time, an advanced form of wet-weather treatment or primary wet-weather
treatment integrated with land management is required. The latter is the
most cost-effective at $3M. This can help set cost-effective standards as
well as select alternatives.
Also a general methodology has been developed for evaluating the impact
of CSO's on receiving waters and for determining the abatement costs for
various water quality goals (11). It was developed from actual 201 facility
planning experience in Onondaga Lake in Syracuse, New York. An important
goal of studies to determine the impact of waste discharges on a receiving
water is to predict the waste loads that can be assimilated without violation
of water quality standards so that a loading curve such as shown in this
figure can be defined.
Figure 11 illustrates the potential effect storm loads may have in
violating a 5 mg/1 DO standard after dry-weather treatment is upgraded. It
further implies that CSO pollution loads should be abated next, since they
are the easiest of the storm loads to control and capture.
„ STORM _
o
E
I
X
o
Q
LU
O
CO
w
Q
WATER QUALITY GOAL
FUTURE
DRY
WEATHER
I O
03
zee
\
if
\
CSO
O O LU
LUC/5 DC
OCOh-
PRESENT DRY WEATHER
TOTAL OXYGEN DEMAND DISCHARGED TO LAKE (Ibs/day)
Figure 11. A Typical Loading Curve Relating Pollutant Load to Water
Quality Response
26
-------
The methodology outlines the interrelationships -of various phases of a
CSO study and develops.procedures necessary to arrive at the costs to meet
various water quality goals.
A project has been initated with Manhattan College (21) to develop a
methodology for defining criteria for wet-weather water quality standards.
The methodology will provide assistance in determining the degree of
treatment of intermittent urban discharges. Recognizing an important gap
in the aforementioned methodologies .the duration of water quality standards
violations vs. species survival will be taken into consideration. This may
be difficult to accomplish since there is a severe lack of data in this area.
The research will include the definition and testing of a method for
evaluating wet-weather water quality criteria considering both classical
pollutants and toxics for the different types of receiving waters. Load
matrices will be developed for determining the load of_ a constituent as, a
function of urban area size and the size and frequency of wet-weather events.
These arrays, will, act as input to effects matrices. • ..•
Other factors influencing, the effects matrices, are the receiving
water size and type and the organism/contaminant transformations.. Thus,
situations where water quality problems associated with discharges from
urban runoff will be defined. An effort will be made to generalize the
results on anticipated problem solutions to;-yield a national scale assessment
of the magnitude of the problem. '' •., "' ';'
This project will coordinate with other ongoing receiving water impact
studies both within and outside the SCS Program.. ; ; :
A user's manual/guidebook for wet-weather flow quality and quantity
monitoring is currently being developed (22). There are various reasons
for entering into a wet-weather sampling and monitoring program and based
on the desired objectives the approach philosophy will vary. This- manual
will delineate various approach philosophies to statistically satisfy
various wet-weather sampling/monitoring needs/objectives,-including separate
sections on best management practice (BMP) evaluation methodologies and
receiving water impact methodologies to assure statistical reliability and
transferability of data within these areas.
CONCLUSION AND RECOMMENDATIONS
Under certain conditions'storm runoff can govern the quality of receiving
waters regardless of the level of dry-weather flow treatment provided. Hence,
control of runoff pollution can be a viable alternative for maintaining
receiving water quality standards. • . .
The SCS program has had only partial success in finding urban storm flow
generated receiving water impacts-employing the conventional DO parameter.
The problem appears to be in the use of conventional dry-weather monitoring
applied for unsteady-state flow regimes caused by storms. Based on a
comparative analyses of wet vs. dry-weather oxygen demanding substances
loads there remains a high potential for these impacts-to occur in receiving
27
-------
waters. Further studies should be considered to resolve the anomaly between
actual runoff loadings and observed receiving water impacts. The Program has
been more successful in sediment analysis than in water column analysis for
finding DO depletions. Direct evidence has been obtained [from the Milwaukee
River project (9)] of how a disturbed benthos depletes DO from overlying
waters. Studies have also shown that storm and CSO adversely affect
sediment by toxics enrichment and resultant biological upsets. Since
particulate matter in untreated storm and CSO is larger, heavier, and in
significant quantities when compared to treated sanitary effluent more
needs to be known about the fate and transport of settleable and separable
materials. Hydrodynamic solids separation and sediment transport routines
must be added to receiving water models to take care of the neglected or
presently omitted but significant particulate and bed load flow fields.
Urban runoff has been identified as a major source of toxic substances
including heavy metals, which have been shown to concentrate in biological
species and sediment, and petroleum hydrocarbons some of which are known
carcinogens and matagens. Very little is known about the long term effects
of these substances. Further investigation into their receiving water impacts
is needed. Integration of various research program efforts and results are
needed in this area to optimize the benefits, avoid duplications and minimize
the costs involved.
J
fl • . •
In the area of control methodology, a simplified continuous receiving
water quality model has been developed for preliminary planning and screening
of areawide wastewater treatment alternatives in terms of frequency of water
quality violations. Also a general methodology has been developed for
evaluating the impact of CSO's on receiving waters and for determining the
abatement costs for achieving various wate^r; qualaty goals. Still more work
is needed on methodologies.
Mandates of the law are upon us, including wet-weather flow pollution
control; monies are being spent at large scale by EPA and others for water
pollution cleanup. In order for governments to execute their function in
this area properly, wet-weather flow pollution must be considered,
and research and development be fostered to back this need.
28
-------
REFERENCES ..
1. Olivieri, V.P., _et_ al_. Microorganisms in Urban Stormwater.
EPA-600/2-77-087, U.S. Environmental Protection Agency, 1977,
NTIS No. PB 272 245.
2. Lager, J.A., et al. Urban Stormwater Management and Technology:
Update and Users7" Guide. EPA-600/8-77-014, U.S. Environmental
Protection Agency, 1977, NTIS No. PB 275 654.
3. Keefer, T.N., et al. Dissolved Oxygen Impact from Urban Runoff.
EPA-600/2-79-156", U.S. Environmental Protection Agency, 1979, NTIS
No. Pending.
4. Stiefel, R.C. Dissolved Oxygen Measurements in Ohio Streams
Following Urban Runoff. EPA Draft Final Report, EPA Grant
No. R-805201.
5. Ketchum, L.H., Jr. Dissolved Oxygen Measurement in Indiana
Streams During Urban Runoff. EPA-600/2-78-135, U.S. Environmental
Protection Agency, 1978, NTIS No. PB 284 871.
6. McConnell, J.B. Impact of Urban Storm Runoff on Stream Quality
Near Atlanta, Georgia. EPA. Draft Final Report, EPA Inter-Agency
Agreement No. IAG-D6-0137. ,
7. Heaney, J.F., jet_ al_. Nationwide Evaluation of Combined Sewer
Overflows • and Urban Stormwater Discharges, Volume II - Cost
Assessment and Impacjtsj5,tEP3A-600/2-77-064b, 'U.S. Environmental
Protection Agency, 1977, 'NTIS No. PB 266 005.
8. Field, R. and D.C. Ammon. Potential of Stormwater Impacts
Based on Comparative Analysis of Wet and Dry Weather Pollutant
Loading. In: paper presented at National Conference on Urban
Stormwater and Combined Sewer Overflow Impact on Receiving
Water Bodies, Orlando, Florida, November 26-28, 1979.
9- Meinholz, T.L., _et_ al. Verification of the Water Quality Impacts
of Combined Sewer Overflows. EPA-600/2-79-155, U.S. Environmental
Protection Agency, 1979, NTIS No. Pending.
10. Klemetson, S.K., et al. Movement and Effects of Combined Sewer
Overflow Sediments in Receiving Waters. EPA Draft Final Report,
EPA Grant No. R-806111.
11. Moffa, P.E., et al. Methodology for Evaluating the Impact
and Abatement of Combined Sewer Overflows: A Case Study
of Onondaga Lake, New York. EPA Draft Final Report, EPA
Grant No. R-805096.
29
-------
r
12. Tomlinson, R.D., et_ al_. Fate and Effects of Sediments from
Combined Sewer and Storm Drain Overflows in Seattle's Nearshore
Waters. EPA Draft Final Report. EPA Grant No. R-805602.
13. EPA Grant No. R-805580. Stormwater Management to Improve
Urban Lake Water Quality. University of Central Florida, Orlando,
Florida.
14. Pitt, R. and M. Bozeman. Water Quality and Biological Effects
on Urban Runoff on Coyote Creek. EPA Draft Final Report,
EPA Grant No. R-805418.
15. Mytelka, A. I., et al. Combined Sewer Overflow Study for the
Hudson River Conference. EPA-R2-73-153, U.S. Environmental
Protection Agency, 1973, NT IS No. PB 227 341.
16. Congressional Federal Register. National Oil and Hazardous
Substances Pollution Contingency Plan. Title 40, Chapter 5,
Part 1510, Subpart A, Section 1510.5(1) (2) (Definition of Medium
Discharge: 10,000 to 100,000 gal. oil to costal waters), July 1,
1976.
.•. . i it JL
17. Feuerstein, D.L. and W.O. Maddaus. Wastewater Management
Program, Jamaica Bay - Volume IfOSurrfmary Report. EPA-600/2-
76-222a, U.S. Environmental ProtectfoiWrkgency, 1976, NTIS
No. PB 260 887. • " r' !v?! v,
18. EPA Grant No. R-806640. Evaluation of Urban Runoff and Combined
Sewer Overflow Mutagenicity. Onondaga County Department
of Drainage and Sanitation, Syracuse, New York.
19. EPA Grant No. R-805663- Nationwide Assessment of Receiving Water
Impacts from Urban Stormwater Pollution. University of Florida,
Gainesville, Florida.
20. Medina, M. Level III: Receiving Water Quality Modeling for
Urban Stormwater Management. EPA-600/2-79-100, U.S. Envirbh"-
mental Protection Agency, 1979, NTIS No. Pending.
21. EPA Grant No. R-806828. Development of Methods to Define Water
Quality Effects of Urban Runoff. Manhattan College, Bronx,
New York.
22. EPA Grant No. R-806345. Prepare Guidelines for Wet-Weather
Monitoring. Urban Drainage and Flood Control District, Denver,
Colorado.
30
-------
Moderator:
Second Session
IMPACTS ON LAKES
3i.i-
Waldron M. McLellon
University of Central Florida
Orlando, Florida
31
-------
AN APPROACH FOR ASSESSING THE WATER
QUALITY SIGNIFICANCE OF CHEMICAL
CONTAMINANTS IN URBAN LAKES
G. Fred Lee and R. Anne Jones
Colorado State University
Fort Collins, Colorado
ABSTRACT
The conversion of rural lands to urban area is often accom-
panied by a significant increase in the total amount of phospho-
rus and nitrogen derived per unit area of a waterbody's watershed.
For rural lands phosphorus without livestock, export coefficients
typically range from 0.005 to 0.5 g P/m2/yr. while in urban areas
the typical export coefficient is 0.1 g P/m2/yr. A significant
part of this difference is due to the much greater water yield
per unit area of watershed in urban areas compared to rural areas.
Urban runoff typically contains appreciable quantities of
both soluble ortho P and particulate forms of P. Studies have
been conducted to evaluate the amounts of available forms of
phasphorus present in typical urban stormwater drainage for
several municipalities located across the U.S. It has been found
that on the order of 10 to 30 percent of the particulate phospho-
rus present in urban stormwater drainage would likely become
available to affect algal growth in a lake or stream. As a re-
sult of these findings, the focal point of the control of nitro-
gen and phosphorus from urban stormwater sources should be direc-
ted toward the soluble orthophosphate component. Most stormwater
drainage control programs are directed toward control of particu-
late matter. Such programs are likely to have limited impact on
eutrophication-related water quality in urban lakes, since only
a small part of the particulate phosphorus will likely become
available to stimulate aquatic plant growth in the waterbody.
The OECD (Organization for Economic Cooperation and Develop-
ment) eutrophication modeling study which included about 40 water-
bodies across the U.S. and 200 waterbodies in Western Europe,
North America, Japan, Australia, etc., has shown that the phos-
phorus load normalized by waterbody mean depth and hydraulic
retention time is correlated to the planktonic algal chlorophyll
concentration, planktonic algal-related water clarity and hypo-
limnetic oxygen depletion rate. The results of the OECD study
provide the tools necessary to quantitatively assess what water
quality improvement can be achieved as the result of various
nutrient control efforts.
32
-------
INTRODUCTION
Increasing attention is being given to management of water
quality in urban lakes,, While these waterbodies have previously
been generally neglected by pollution control agencies, the US
EPA and state and local governmental agencies are beginning to
provide funding to investigate sources of contaminants for urban
lakes and to develop water quality control programs for them.
Two of the major forces behind- these efforts are the US EPA
Urban Lakes and the Combined Sewer Overflow Programs. These
programs, if carried out properly, will provide municipalities
with the information needed to develop cost-effective, techni-
cally valid water quality management programs for urban lakes.
This paper considers a number of factors that must be evaluated
in the development of such programs.
CHARACTERISTICS OF URBAN LAKES
The first issue that must be addressed in the development of
urban lakes' water quality management strategy is whether or not
urban lakes behave differently from rural lakes with respect to
contaminant load-water quality response. Generally urban lakes
tend to be shallow,.with maximum depths on the order of a couple
of meters. With few exceptions they are highly fertile and would
generally be classified as eutrophic. The littoral areas of many
urban lakes are dominated by macrophytes and attached algae.
They normally have "pea soup" green to brown color arising from
planktonic algae and suspended sediment. Occasionally they will
have dense algal blooms which can create obnoxious surface scum.
Thermal stratification, if it develops, is usually short-lived;
rarely would a thermocline persist throughout the summer growing
season. With few exceptions, urban lakes have appreciable popu-
lations of rough fish such as carp and bullheads. All of these
characteristics tend to make urban lakes fall into the group of
shallow, weedy, moderately-to-highly turbid waterbodies which
have appreciable mixing between the sediments and overlying
waters. The overall water quality in urban lakes would normally
be classified as poor; it is generally believed that if the pub-
.lie had a choice, they would recreate in rural waterbodies. How-
ever, many urban lakes have greater fishing and other recreation-
al pressures placed on them than their rural counterparts. This
pressure will likely increase significantly as the energy and
33
-------
economic situation continues to deteriorate. It is therefore
imperative that an understanding of the factors controlling
urban lake water quality be achieved in order to maximize water
quality improvements for funds spent in control programs.
In those municipalities where salt is used for road de-
icing, the first major snow melt in the spring can bring appre-
ciable amounts of salt into the lake. Some urban lakes have
become meromictic (temporarily or permanently stratified) be-
cause of this problem.
In addition to water quality problems there is also the
aesthetic problem of littering of the shores and bottom of urban
lakes by the public. Unless effectively policed, the littering
can be a significant deterrent to the use of urban lakes.
While most urban lakes are freshwater, in coastal areas
some are marine-estuarine which may be subject to the influence
of the tide. In some municipalities urban lakes are dammed
rivers and therefore may have appreciable contaminant input from
upstream sources. Many urban lakes are used as receptacles for
combined sewer overflow in municipalities where combined sewers.
exist. Routinely they are the receptacle for urban stormwater
drainage and frequently they receive domestic wastewater inputs
from separated sewerage sys.tems during periods of high f low,.,..,...
electrical outages, pump failure, etc.
WATER QUALITY MANAGEMENT IN URBAN LAKES
WATER QUALITY PROBLEM IDENTIFICATION
The first step in the development of a water quality man-
agement program for any waterbody is an assessment of the water
quality problems that exist within the waterbody. Generally,
the most significant and in many cases the only water quality
problem that exists in urban lakes is excessive fertilization.
While many urban lakes may receive relatively large loadings of
a variety of toxic chemicals such as heavy metals, pesticides,
etc., it appears that these contaminants rarely cause signifi-
cant water quality problems in urban lakes. This is a result of
the fact that more eutrophic waterbodies are able to "detoxify"
many contaminants as a result of contaminant interaction with
suspended and deposited particulate matter.
Urban lakes tend to receive relatively large loads of oxy-
gen demanding materials in urban stormwater drainage and com-
bined sewer overflows and/or domestic wastewater bypasses to the
lakes. While oxygen demand problems can occur in urban lakes,
they tend to be infrequent„ The overall trophic status of many
of these waterbodies is such that rarely would low dissolved
34
-------
oxygen be a significant cause of water quality deterioration
during ice-free periods. However, as discussed by Lee and Jones
(1979c), in the winter in areas where waterbodies develop an ice
cover, shallow lakes of this type, especially during periods of
snow cover preventing penetration of light through the ice5 tend
to be prone to winterkill where more desirable game fish would
be lost because of low dissolved oxygen.
Urban stormwater drainage and the entrance of domestic
wastewaters into urban lakes tends to cause elevated coliform
counts which can serve as a basis for closing beaches and re-
stricting other body contact with the water. Urban lakes in
regions where there is appreciable new construction can receive
large amounts of erosional materials which will result in more
rapid siltation especially near the point where the stormwaters
enter the waterbody. These situations can cause water dis-
coloration due to the presence of large amounts of suspended
sediment in the urban stormwater drainage.
While it is possible to generalize on water quality prob-
lems of urban lakes, the water quality problems of any particu-
lar-urban lake must be assessed on a case-by-case basis in which
deterioration of water quality is judged based on impairment of
a beneficial use of the water. Water quality should never be
judged based on the concentration of a chemical contaminant in
the water or the sediment. Many pollution control agencies make
a significant error in assuming that one can relate water qual-
ity to the chemical composition of the water as determined by
standard chemical analytical procedures normally used in the
water quality field. Chemical contaminants exist in aquatic
systems in a variety of chemical and physical forms, only some
of which are available to affect water quality. Basing an urban
lake water quality control program on the total contaminant load
or concentration for a particular contaminant can easily result
in the public spending large amounts of money in the name of
water pollution control with little or no improvement in urban
lake water quality<, As indicated above, those responsible for
developing water quality management programs for urban lakes
must first assess and document'how the public's use of the urban
lake is impaired by a particular water quality characteristic.
Once this is known, it is then possible to begin to formulate
the studies that are needed to provide the information which can
serve as a basis for developing a water quality management strat-
egy for a particular waterbody. "-!',.:
WATER QUALITY CRITERIA AND STANDARDS
The traditional approach that has been followed in the water
quality control field during the past 15 years and has been used
as a basis for developing water quality management programs has
involved the use of water quality criteria and/or standards.
Generally fixed numeric values have been established against
35
-------
which results of chemical analyses of water and/or sediments
have been compared. This approach has worked reasonably well
for- the control of water quality deterioration arising from many
municipal and industrial point source contaminants, where the
objective of the control program was the elimination of gross
v?ater quality deterioration such as aquatic organism acute tox-
icity. Public Law 92-500, the 1972 Amendments to the Water
Pollution Control Act, requires that the US EPA develop water
quality criteria which are to become the backbone of the water
pollution control program in the U.S. The US EPA released the
first set of these criteria, commonly called the "Red Book"
criteria, in July, 1976 (US EPA, 1976)„ Another set is current-
ly under review. Public Law 92-500 requires that every state
review its existing water quality standards every three years
and that standards acceptable to the US EPA be promulgated by
each state. Some states are adopting, without modification,
US EPA July, 1976 criteria as fixed numeric standards. These
criteria are "worst case" values developed by exposing organ-
isms for chronic - usually life-time periods of time to forms
of the contaminants which are 100% available. They are designed
to protect fish and other aquatic life against some of the more
subtle effects of contaminants such as impairment of reproduc-
tion and growth rate. For many chemicals there is a several
order of magnitude difference between the critical concentration
which will kill fish within 96 hours and the chronic safe con-
centration. Adoption of the US EPA July, 1976 Red Book criteria.
as water quality standards will, in general, provide the basis ;';
for ecologically protective water quality management programs/;1;^-
It is likely that there will be few instances where a waterbpdy'2
containing concentrations of contaminants equal to or less than'"
the Red Book criteria numeric concentrations would experience
deteriorated water quality due to chemicals considered in the
Red Book.
While ecologically protective, there are many who question
the cost-effectiveness of application of standards numerically
equal to the worst case criteria since that approach does not
consider the fact that substantial parts of many chemical con-
taminants present in aquatic systems are in chemical forms which
are unavailable to affect aquatic organisms and water quality.
This is especially true for contaminants associated with partic-
ulate matter. Lee et al. (1978a), Jones and Lee (1978b), and
Jones et al. (1979) have conducted a comprehensive study of the
water quality significance of chemical contaminants associated
with aquatic sediments from locations across the U.S. It has
been found that most of the contaminants associated with aquatic
sediments are in forms unavailable to affect water quality.
They have pointed out that there is no relationship between the
total concentration of a contaminant in a sediment and the po-
tential for this contaminant to affect water quality. They
further pointed out that it is not appropriate to use water qual-
ity standards numerically equal to the Red Book criteria for many
36
-------
aquatic systems, especially those which contain large amounts of
sediment-associated contaminants. Since the behavior of con-
taminants in urban lakes is frequently greatly influenced by
sediments both suspended and deposited, application of water
quality standards equal to Red Book criteria would not in mpst
instances be a cost-effective approach for developing a water
quality management strategy for these systems. Since the pri-
mary sources of contaminants for urban lakes are urban storm-
water drainage and in some instances combined sewer overflow and
domestic wastewater diversions, and since numerous studies have
found that appreciable parts of the1 contaminants in these
sources are in particulate forms, it is likely that water qual-
ity programs directed toward the total contaminant load to an
urban lake could readily result in expenditure of large amounts
of money for contaminant control with little or no improvement
in urban lake water quality. Because' of the relatively high
cost associated with the control of contaminants in urban'storm-
water drainage, it is imperative that all funds spent for this
purpose be directed to the maximum -extent possible to the con-
trol of available forms of contaminants.
HAZARD ASSESSMENT APPROACH - :
General Aspects*
„, /•-
„ ;..T^In situations where chemical water quality standards numer-
ically equivalent to the US EPA Red Book criteria are not suit-
able" indicators of water quality deterioration, -such as in
ass.essing the water quality -impact of contaminants in urban
stormwater drainage on urban lakes, it is necessary to use the
hazard assessment approach to develop a technically valid, cost-
effective, ecologically protective-program for 'water quality
management. This approach was originally developed for pre-
screening new chemicals for their environmental impact. Several
papers in the book edited by Cairns et al. (1978) described the
application of this approach for this purpose; :Lee and Jones
(1979a,b) and Lee et al. (1979b) have described the application
of this approach for existing chemicals. Pollution control
agencies will have to follow an approach similar to that dis-
cussed by Lee and Jones (1979a,b) and Lee et al. (1979bTin de-
veloping water quality control programs for urban lakes if these
programs are to maximize water quality improvement for the money
spent. ' ,
An aquatic environmental hazard assessment is built"on two
basic components: aquatic toxicology and environmental chemis-
try-fate information. In the aquatic toxicology phase, the con-
centrations of the contaminant, effluent, or other source (such
as urban stormwater drainage) of concern which cause no adverse '
impact on aquatic organisms is established. As part of this
*This discussion is adapted from Lee and Jones (1979b).
37
-------
evaluation, concentration-duration of exposure, couplings are de-
veloped for each contaminant or potentially .hazardous input which
define the "no adverse effect" level for a spectrum of exposure
durations which may be encountered in the aquatic environment.
Figure 1 shows a general-case concentration-duration of exposure
relationship. It shows that for short durations of exposure,
organisms can tolerate higher concentrations of contaminants
without being adversely affected. As the exposure duration in-
creases, the no effect concentration decreases eventually to. a
chronic safe level. For particular contaminants, these couplings
must be based on concentrations of available forms; for a
specific effluent containing a mixture of potentially hazardous
materials, this can be based on, dilutionofeffluent or other
source of contaminants„ In the aquatic toxicology phase of.haz-
ard assessment the bioaccumulation potential of the contaminant
is also assessed. Assessment of this potential impact will be
discussed further in a subsequent section.
The environmental chemistry-fate portion of an environmental
hazard assessment considers the chemical processes that occur
within the aqueous system as well as interactions with other en-
vironmental components that alter the form of the contaminant of
concern and its transformation products in the aquatic system.
Major types of chemical reactions that commonly occur in the
aquatic environment which could cause significant changes in the
form of a potentially hazardous chemical in terms of its water
quality impact are acid-base, precipitation, complexation, oxi-
dation-reduction, hydrolysis, photolysis, gas transfer, bio-
chemically mediated reactions - biotransfqrmation, and sorption,
both biotic and abiotic. The environmental chemistry-fate phase
of a hazard assessment for an aquatic, system also determines the
transport pathways of the contaminant and its transformation
products from the point at which it enters the aquatic system,,, to
its final disposition or the point at which it leaves the system.
In association with this, the physical processes of adveetion^
transport, and dilution-dispersion must be defined for the ter-
restrial, atmospheric, and aquatic environments with which the
contaminant of concern and its transformation products come in
contact. All of the pertinent environmental chemistry and fate
information is formulated into a series of differential equations
which describe each potentially significant transformation and
transport pathway. Such an environmental chemistry-fate model is
described schematically in Figure 2. Once an environmental chem-
istry-fate model has been verified it can be used to predict for
a particular contaminant input, the concentration of the con-
taminant of concern and each potentially significant transfor-
mation product in each aquatic environmental component, i.e., in
solution, associated with'particulates, and associated with organ-
isms, etc.
During the process of making an assessment of the environ-
mental hazard associated with the particular source of contami-
38
-------
CONCENTRATION
OF AVAILABLE
FORMS OF
-CONTAMINANT
ORGANISM IMPACT
. WATER QUALITY IMPAIRMENT
CHRONIC SAFE CONCENTRATION (RED BOOK VALUE)
96 MRS
DURATION OF EXPOSURE
FIGURE 1 - General Available Concentration - Duration of Exposure
Relationship.
39
-------
ATMOSPHERE
INPUT
V
GAS EXCHANGE
SORPTION
BIOCONCENTRATION
P> DEPUR/
DEPURATION
DESORPTION
(CHEMICAL TRANSFORMATION)
SEDIMENT EXCHANGE
SEDIMENT
OUTPUT
D (AVAIL. FORM)
, ^ , .
= K^(GAS EXCHANGE) + K2 (BIOCONCENTRATION) +
K5(SORPTION) + Kjj(CHEMICAL TRANSFORMATIONS) +
FIGURE 2 - Environmental Chemistry - Fate Model
40
-------
nants, .the predicted concentration-duration of exposure coupling
is compared with the "no effect concentration-duration of ex-
posure" relationship developed for the chemical or source in
question in the toxicological portion of the hazard assessment
(Figure 1). Figure 3 shows schematically a variety of the
possible couplings and their relationship to the area of "im-
pact." The hatched area in this figure represents an area in
which there is sufficient duration of exposure to sufficient
concentrations of available forms of contaminant to have an ad-
verse effect on aquatic organisms and/or water.quality. This
relationship, must be defined for each contaminant or source of
contaminants of concern in formulating a hazard assessment. The
numbered curves (1-5) in Figure 3 show results, that could be ob-
tained through environmental chemistry-fate modeling, where a
combination of dilution and chemical reactivity bring about a
certain'concentration-duration of exposure relationship. Curve
1 represents that coupling typical of spill situation, where
there is toxicity for a short time associated with the point of
entry before any reactions or dilution takes place. This might
also be the situation associated with the mixing zone for a
particular discharge such as urban stormwater drainage or com-
bined sewer overflow.
Curve 2 in Figure 3 is a case where the contaminant at
levels that are normally found in the environment does not show
any acute toxicity but is. chronically toxic either to an organ-
ism or to higher forms that may use the organism as food. PCB's,
DDT, and mercury would all fall into this category. However,
because of concentration of some of these types of chemicals
within the higher' trophic, level fish, there is' a potential for
harm to man and other animals that use these fish as a source of
food.
Curve 3 in Figure 3 represents the type of situation which
might be associated with municipal wastewater discharges which
contain ammonia, where for short durations of exposure there is
no impact because possible durations of exposure are too short.
However, there is an intermediate zone at some distance from the
point of discharge where expected durations of exposure of the
organisms to available forms are sufficient, where there could
be toxicity to fish or other organisms that reside in the area.
Eventually the ammonia would be oxidized or diluted to non-toxic
levels as one proceeds further down the stream from this zone.
Curve 4 is representative of the situation where there is a
transformation of the contaminant added•to the system which
causes it to be more toxic as It goes downstream. Eventually it
is either diluted or detoxified through other reactions. An
example of this type of situation is one involving the addition
of a complexed heavy metal to the environment where the complex
is biodegradable, releasing the heavy metal at some distance
downstream in sufficient concentrations to be toxic to aquatic
41
-------
CONCENTRATION
OF AVAILABLE
FORMS
DURATION OF EXPOSURE OR
TIME IN THE ENVIRONMENT
FIGURE 3 - Examples of Concentration of Available Forms - Duration of
Exposure Couplings Compared to Area of Water Quality Impact
42
-------
life in that region. Curve 5 is the case that exists for most
chemicals for which there is sufficient treatment or controlled
use so that there is no toxicity associated with it, either
acute or chronic.
An environmental hazard assessment for any contaminant
source should be conducted in a series of levels or tiers in
which the aquatic toxicology and environmental chemistry-fate
are determined with increasing sophistication and reliability
with succeeding tiers, with a decision point at the end of each
tier. The decision choices are: 1) do not allow discharge be-
cause of excessive expected hazard, 2) restrict discharge
through use, or treatment to reduce -environmental hazard to an
acceptable level, 3) proceed with discharge as- is - expected
environmental hazard is acceptable, or 4) continue testing to
more precisely defined expected impact. The ovjerall approach is
to screen for gross effects of a contaminant such as acute tox-
icity and 'estimated chronic toxicity in the earlier tiers. At
the higher tiers, the -fo.cus is on conducting tests for the more
subtle effects 'of the contaminant such as impairment of repro-
duction. These higher tier tests only need5 be conducted if the
lower tiers suggest that there is a potential for this type of
water quality problem. As shown in Eigure M-, with each .succeed-
ing tier of testing a more precise estimate - of 'both the'actual
expected environmental concentration and -also t.he '-'ho impact"
concentration for organisms is obtained. '• l ''••
The tiered environmental hazard assessment approach will
likely be of particular significance in,developing water quality
management plans for urban lakes. It is likely, that the results
of a few screening bioassays normally associated with lower tiers
of testing will show that urban stormwate^; drainage does not con-
tain sufficient amounts of contaminants to'cause toxicity. Find-
ings of this type can greatly simplify the study program and
thereby reduce its -. cost, compared to the normal procedure for
studying water quality involving routinely collecting data for a
year or two and then attempting to'-interpret the data at the end
of the study period in terms of the waterbody's water quality.
Most water quality.studies as conducted today provide limited
amounts of information that can be used to formulate cost-effec-
tive, technically valid and at the same ecologically protective
water quality management programs. •
Control of Toxics
A key component of the hazard assessment approach for urban
lakes as well as other waterbodies is the determination of the
availability of the contaminants in the inputs to the lake. As
discussed by Lee et al. (1979a), Jones and Lee (1978b), and Jones
et al. (1979), chemical leaching tests for particulate matter
such as are frequently used, do not in general properly assess
the amounts of available forms of contaminants that can affect
43
-------
RELIABILITY
OF ESTIMATE
CONFIDENCE LIMITS
NO IMPACT
ESTIMATED
CONCENTRATION
2345
TIER OF TESTING
FIGURE 4 - Aquatic Toxicology and Environmental Chemistry-Fate
Modeling Results for a Tiered Environmental Hazard
Assessment
44
-------
water quality for conditions typically found in urban lakes. In-
stead, bioassay procedures must be used on the sources of con-
taminants because of the variety of contaminants in the sources
and for urban lakes, the typically high suspended particulate
levels. If chemical leaching tests are used, the results must
be verified against an extensive series of bioassays conducted
at various times of the year in order to determine if there are
any seasonal effects on the availability of the contaminants.
The first tier of an environmental hazard assessment for
toxicants associated with contaminant inputs to urban lakes
should be a 96 hour static bioassay screening test of the type
described by Lee et al. (1978a) and M.ariani (1979), on the un-
diluted water from each input. If no acute toxicity is found in
these waters, and if' there is ,a population of warm water pan fish
(i.e., bluegill, sunfish, etc.) which are reproducing in the
waterbody, and if -there, is,-at .least a. 10 0-f old, .dilution of the
input waters with the lake water within a few -days to a week or
so of input, then the likelihood of there being a;ny type of tox-
icity, including impairment,of reproduction due to this source,
is remote. The 100-fold dilution is based on the factor that is
typically found to relate acute lethal toxicity to the chronic
safe limit-. - -•-.-•
If acute toxicity is found in the screening bioassys of the
undiluted urban storniwater drainage or other source of contami-
nant for an urban lake, then additional bioassays.must be con-
ducted on diluted contaminant source water, using- appropriate
dilutions with lake water based on the characteristics of the
waterbody, to check for both acute and chronic.toxicity. If
potentially significant toxicity is found, it would be important
to attempt to identify the specific toxic component. This can
possibly be accomplished using the standard additions approach
described by Lee and-Jones .(19 7 9d) ........ For .further., information on
the use of bioassay procedures in a hazard assessment, consult
Lee et al. (1979b).
Since bioaccumulation of contaminants in fish is of concern
in urban lakes, it is imperative that some work also be done in
the early tiers of any hazard assessment program to determine
whether excessive buildup of contaminants is occurring in organ-
isms used for human food. In general, the only.criteria by
which one can judge the significance of the body burden of some
contaminants within an aquatic organism is the Food and Drug Ad-
ministration (FDA) Action Limit. Therefore the most appropriate
and only reliable method for assessing potential water quality
problems associated with certain contaminants is to collect from,
the lake of concern seasonally over at least one year, represen-
tative samples of fish being used as human food and analyze their
flesh for each of those contaminants for which the FDA has es-
tablished Action Limits such as PCB's, mercury, DDT and its ana-
logs, as well as several other chlorinated hydrocarbon pesticides.
45
-------
If the concentrations of any of those contaminants would cause
the fish to be unsuitable for use as food, i.e., the concentra-
tions are above the FDA Action Limits, then each source of that
contaminant must be investigated and, control programs initiated
where possible.
It is important to note that for developing control programs,
while first priority must be given to analyzing fish flesh for
those chemicals for which Action Limits exist since those are
the only interpretable data at this time, as funds permit, the
fish should be analyzed for any other contaminant which may be
present in large amounts or may be of concern for a particular
system. While these data may not at this time.be interpretable
in terms of water quality and fish usability, they may be_a clue
to potential water quality problems that should be investigated
further since organisms in aquatic systems tend to be integrators
for some types of persistent contaminants.
Control of Eutrophication
While it is expected that toxicants present in urban storm-
water drainage and other urban lake water sources would,not in
general have a significant adverse effect on urban lake water
quality, nutrients (nitrogen and phosphorus) frequently cause
significant water quality deterioration in these waters. The
hazard assessment approach for evaluating the significance of
nutrient inputs to waterbodies, including urban lakes, is based
on the OECD eutrophication modeling approach. The Organization
for Economic Cooperation and Development (OECD) has sponsored a
22 country, 200 waterbody (lake and impoundment) study of the
relationships between the nutrient load to a waterbody and the
eutrophication response of that waterbody. Based on concepts
originally developed by Vollenweider (1968) and the data on the
approximately 4-0 OECD waterbodies in the U.S., Rast and Lee
(1978) found that good correlations existed between the P load
to a waterbody normalized by the waterbodyrs mean depth and
hydraulic residence time, and the average summer planktonic algal
chlorophyll concentrations, planktonic algal-related water clar-
ity as measured by average summer Secchi depth, and for those
waterbodies which thermally stratify, hypolimnetic oxygen de-
pletion rate. Figure 5 shows these relationships. Included in
this data set were several urban lakes including Lake Wingra in
Madison, Wisconsin. While some have claimed that the nutrient
behavior in small shallow lakes of this type is not the same as
that in larger, deep waterbodies, the OECD eutrophication model
has demonstrated that when normalized by hydrologic and morpho-
logic characteristics, small and large lakes and impoundments
behave in a similar manner in terms of eutrophication response
to nutrient inputs. Lee et al. (1977, 1978b) have presented
summaries of these results for the US OECD waterbodies.
Subsequent to the completion of the Rast and Lee (1978)
46
-------
-------
work, Lee and his associates have evaluated the nutrient load-
response relationships for an additional 60 or so waterbodies
primarily in Jrhe,.U, S,. and,have found thatw these waterbodies
follow the same relationships shown in Figure 5"« : Further, the
other approximately 150 OECD waterbodies' nutrient load-eutro-
phication response relationships also are in agreement with what
Rast and Lee (1978) found for the US OECD waterbodies.
Lee and Jones (1979c), who have conducted an extensive re-
view of the effect of eutrophication on fisheries., have used the
data provided by Oglesby 01977) and Rast and Lee (1978) to de-
velop the relationship between normalized P load and fish yield
shown in'Figure 6. As would be expected, it shows that the
greater the nutrient load, thefgreater the overall fish yield.
Lee and Jones (1979c) should be consulted for further information
on this topic. = *
There are a number of constraints on t)ie use of the OECD
eutrophication model that must be considered*-for,-each waterbody
to which the approach is to be applied. While' studies- are being
conducted on application of this approach to N-limi.ted. water-
bodies, currently this approach is only applicable'to those
waterbodies whose maximum summer phytoplankton biomass is limited
by phosphorus. The limiting nutrient in a waterbody or an area
of a waterbody can be determined by measuring the concentrations
of available N (NO1! plus NH~) and available P (soluble ortho P)
during the period of_maximum_ phvjtqglankto:ri biomass:.; In general,
if the available P, cbncentration"*"is "reduce'd "to' a fSw ug P/l, the
phytoplankton growth at the time the samples were "collected was
most likely limited by P. If the available N concentrations are
reduced to about 30 to 50 ug/1 or so, N is likely limiting phyto-
plankton production. If neither nutrient is reduced to these
levels, some other factor is likely limiting maximum planktonic
algal biomass.
The ratio of available N to available P is also used to
indicate which nutrient would be depleted first (i.e., poten-
tially limiting) in a water, based on the theoretical uptake
ratio of these nutrients by algae of 7.5 N to 1 P on a mg/1
basis. Algal assays are also used to estimate the limiting
nutrient by determining which nutrient if added would promote
increased algal growth. Caution must be exercised in using the
latter two approaches to determine the limiting nutrient; they
must be performed near the time of maximum algal production since
the results of such tests run at other times ,of the year will not
necessarily give an accurate representation of the limiting nu-
trient during the period of water quality concern. Further,
analyses for available N and P during peak"biomass production
should be conducted in conjunction with these procedures to ver-
ify that one of these nutrients is actually limiting the growth;
while an N-to-P ratio may indicate a lesser relative abundance of
one nutrient, some other factor such as light may in reality
48
-------
I01
IOU
ID
toO
N^X
0>
•H
>>
io3
IO1
10s
IO3
(L(P)/q )/(! + v^T
=> (JO
mg P/m"
Line, of best fit:
Fish Yield =0.7 log [(L(P)/qo)/ (1 + /T~)J - 1.86
S (jj
(r2 = 0.86)
FIGURE 6 - Relationship between P load and Fish Yield
49
-------
limit algal growth during the period of concern.
The OECD eutrophication modeling approach is also only ,, _ ..
applicable to waterbodies in which aquatic plant growth is. domi-
nated by phytoplankton. In its current degree .of development it
cannot be used to assess eutrophication response as measured -by
a macrophyte or attached algal growth parameter. As discussed ...
by Newbry et al. (1979), it is likely possible to develop nutri-
ent load-response relationships for macrophytes where, response
is assessed in terms of the percent of the area of the waterbody
with depth less than 2m that is covered by aquatic macrophytes
and attached algae.
Another constraint on the application of the OECD eutrophi-
cation model is that the average hydraulic residence time (i.e.,
filling time - volume divided by annual input) of the waterbody
must be two weeks or more. For waterbodies having an annual
hydraulic residence time shorter than two weeks, it may be possi-
ble to modify the model as was done by Jones and Lee (1978a) for
Lake Lillinonah, CT. For this waterbody the summer averagehy-
draulic residence time was used since each spring, the waterbody
is essentially completely flushed.
WATER QUALITY MANAGEMENT IN URBAN LAKES
The overall approach that should be used for management of -
water quality on urban lakes is the same as for any waterbody,
namely, limiting the input of available forms of contaminants to
the extent necessary to achieve the desired water quality. In
assessing available forms, it is important to assess, not..only
those forms which are immediately available but also those which
can become available in the waterbody. Because of the potential
significance of urban stormwater drainage as a source of con-
taminants for urban lakes, in those situations where the hazard
assessment shows that this source is .a significant source of
contaminants causing water quality deterioration, studies should
be done to define the source of contaminants in the urban storm-
water drainage. At this time there is a relatively poor under-
standing of the specific sources of total as well as available
forms of contaminants in this source. The studies of Cowen and,
Lee (1973) and Kluesener and Lee (1974) have shown that in the
fall, an appreciable part of- the P present in urban stormwater
drainage is from fallen leaves. Lee (1972) proposed that the
rapid removal of leaves and frequent street sweeping using
vacuuming could significantly reduce the phosphorus content of
urban stormwater drainage. Ahern and Armstrong (1979) found that
street sweeping reduced the P content of urban stormwater drain-
age by 47 to 59%.
The importance of focusing nutrient control programs on
50
-------
available forms of nitrogen and phosphorus was demonstrated by
the studies of Lee and his associates in Madison, WI0 Kluesener
and Lee (1974) found that urban stormwater drainage contributed
about 0.1 g/m /yr of total phosphorus and 0.5 g/m2/yr of total
nitrogen. About 25% of.the total N was in the form of ammonia
and nitrate, while about 50% of the total P was in the form of
soluble ortho Po Cowen et al. (1978), using bioassay techniques
found that about 50% of the particulate organic and inorganic N
(i.e., the difference between the total N and the sum of nitrate
and ammonia) would likely become available to stimulate algal
growth in an urban lake. They also found that only about 20% of
the difference between the total P and soluble ortho P would
likely become available to support algal growth. These results
point to the fact that appreciable parts of the N and P entering
urban lakes from urban stormwater drainage are in forms not like-
ly to support algal growth in the waterbody. While the results
of the studies by Cowen et al. (1978) are in general agreement
with what other investigators who have conducted similar studies
have found, additional work needs to be conducted for each ' system
of interest until sufficient information is gathered to make
appropriate estimates of nutrient export from urban areas and the
availability of these nutrients to support algal growth.
A potentially significant source of nutrients for some urban
lakes is waterfowl. While some of these birds only alter recycle
of nutrients within the waterbody, such as ducks which feed on
aquatic organisms, other waterfowl such as geese feed on land but
spend most of their time in the water. The latter type of bird
can add nutrients from on-land sources which may never otherwise
reach the lake. An example of the potential importance of this
problem is seen in urban lakes in the City of Fort Collins," .Colo-
rado, where the population of resident geese number well into the
tens of thousands.
Typical urban stormwater and/or combined sewer contaminant
control programs are usually directed toward the control of par-
ticulate forms of contaminants. Most of these programs are based
on the erroneous premise that there is a relationship between the
total load of a contaminant and' its impact on water quality0 As
discussed above, it is rare that the total load or concentration
of a contaminant can be used to predict the impact of the con-
taminant on water quality. While control of particulate forms of
a contaminant may be useful for siltation control, it may not be
cost-effective for reducing the'magnitude of chemical related
water quality problems. In many instances the most cost-effective
way to deal with siltation, is periodic dredging,,
The first priority in any urban lake's water quality main-
tenance program should be the removal of debris-trash usually left
by the users of the area. The next priority should be given to
the control of excessive fertility. Obviously, all significant
point sources of nitrogen and/or phosphorus depending on^what is
51
-------
limiting planktonic algal growth in the waterbody, should be
identified. If the hazard assessment study shows P is limiting
or can be made limiting through point source control, and the _
point-source(s) already has some treatment for removal of solids
and BOD, then P removal by alum or iron co-precipitation tech-
niques should be added to the treatment process. Generally, P
can be controlled 'in domestic wastewater to less than 1 mg P/l
total P at a cost of less than a half a cent (0.5O per person
per day for the population served. As noted above, the OECD
eutrophication study modeling results can be used to estimate the
magnitude of planktonic algal-related water quality improvement
associated with point source and diffuse source nutrient control.
Before embarking on large-scale street vacuuming or other diffuse
source control programs, the magnitude of water quality improve-
ment that could be achieved by these measures should be deter-
mined using the OECD eutrophication model. Such control programs
are often relatively expensive and frequently provide limited
reductions in available nutrient loads. As shown in Figure 5,
relatively large reductions in P load are needed to significantly
change the planktonic algal-related chlorophyll .concentration.
This is especially true for eutrophic lakes in which even rela-
tively large reductions in chlorophyll may not result in a per-
ceivable improvement in water quality.
As a result of their shallow character, many urban lakes
have significant water quality .problems due to attached algae
and macrophytes. As discussed by Lee (1973) and Lee and Jones
(1979c), a variety of techniques such as aquatic herbicides,
harvesting, etc., are available for control of these growths.
In general, the overall control philosophy for aquatic macro-
phytes and attached algae should be to remove the fewest number
of these plants as necessary to make the waterbody usable for
the desired recreational purposes. As discussed by Lee and Jones
(1979c), aquatic macrophytes and especially attached algae com-
pete with planktonic algae for available nutrients. Extensive
removal of attached algae and macrophytes can result in produc-
tion of phytoplankton blooms and could destroy fish habitat and
nursery grounds.
The dredging of urban lakes for the removal of deposited
sediment can also have a significant impact on the relative
utilization of nutrients by various types of aquatic plants. Ex-
tensive dredging to depths of 3m or more can greatly decrease the
macrophyte and attached algae in eutrophic lakes with a con-
comitant increase in phytoplankton. Dredging should be restrict-
ed to those parts of the waterbody for which the shallowne.ss
permits excessive growths of attached algae and macrophytes which
significantly interfere with recreational ,uses of the water such
as swimming, boating, or fishing.
The direct alum addition to urban lakes is a technique that
deserves much greater attention for managing urban lake water
52
-------
quality than it has received in the past. In many situations it
may be the only technique available to manage excessive fertili-
zation problems. Since urban lakes'tend to normally be turbid,
alum flocculation of inflowing waters can help reduce turbidity
and also remove P from the watercolumn. The alum 'floe thats forms
can be settled either directly in the lake or in a pre-impplund-
ment constructed for this purpose. Systems can be developed to
automatically feed alum to the stormwater, combined sewer over-
flow, or domestic wastewater bypass inputs to the lake. This
approach does not appear to be harmful to fish as long as" the're
is adequate alkalinity in the water to maintain a desirable: pH.
For low alkalinity waters, it may be-necessary to add lime to
increase the buffer capacity of the1water. The,duration of
effectiveness of this approach is directly dependent on the :hy-
draulic residence time of the waterbody. Waterbodies with -•"
several years1 residence times will require relatively infr'e"-
quent treatment. However, the typical urban lake would require
the more or less continuous treatment of 'all inflowing waters to
be effective. It should be noted that such treatment may result
in increased amounts of attached algae and macrophyte growth -be-
cause of the greater water clarity arising from alum treatment.
This may necessitate additional harvesting or .other control pro-
grams especially for macrophytes which can derive at least part
of their nutrients from the sediments.
Those responsible for' developing nutrient control programs
for waterbodies are -frequently concerned' about the rate of re-
covery of the waterbody upon altering the nutrient load. The
shallowness of many urban lakes has led some to believe that
these waterbodies will not recover upon reduction of nutrient
load because of the release of nutrients from -the sediment.-• As
discussed by Sonzogni e_t' al. "(1976), the rate of recovery of ;a
waterbody upon reduction of P input depends on the P residence
time of the waterbody. About 95% bf the expected recovery from
the degree of altered load occurs within a period equal to three
times the phosphorus residence time.' The P residence time of
many waterbodies is less than one year; which means that within a
few years after altering the load, a new equilibrium P-chloro-
phyll concentration.will be achieved for the waterbody.
CONCLUSIONS
Increasing pressures are being placed on water pollution
control agencies to manage water quality in urban lakes. The
first step in developing a water quality management plan for such
a waterbody is a careful definition of the water quality problems
that may exist therein. In this evaluation, for many urban lakes,
particular attention will have to be given to eutrophication-
related water quality which will require reduction in the amounts
of available forms of nitrogen and phosphorus entering them. The
53
-------
OECD eutrophication study results, should beused to describe the
nutrient load-response relationships for the urban lakes and
most importantly, should be used to estimate, .the magnitude of
water quality improvement that will arise, as, a result of alter-
ing the nutrient loads to P limited urban lakes. In many in-
stances it may not be possible to control nitrogen and/or phos-
phorus to a sufficient degree to attain desired water quality in
urban lakes because of the difficulty of limiting nutrient input
from urban stormwater drainage and/or combined sewer overflow.
Direct alum addition to the inflowing waters is a technique that
should be more thoroughly investigated as a tool to manage ex-
cessive fertility in urban lakes. Also selective dredging and
attached algae and macrophyte control through harvesting can be
effective tools for minimizing impact of aquatic plant nutrients
on urban lake water quality.
For those waterbodies,, impacted by toxicants as well as nu-
trients, an environmental hazard assessment, should be conducted
to define the magnitude of the impact of contaminant sources on
water quality, i.e., beneficial uses, of the lake. Such an
assessment can also be used to develop the information necessary
to control the excessive input of the contaminants. Water qual-
ity standards adapted directly from the US, EPA Red Book criteria
have limited applicability in serving as a basis for developing
water quality management programs for urban lakes. The environ-
mental hazard assessment involving the tiered, development of en-
vironmental chemistry-fate and aquatic toxicology information
with emphasis given to use of bioassays to define available forms
of the contaminant should be used.
54
-------
REFERENCES
Ahern, J. and Armstrong, D. E. "Phosphorus Control in Urban
Runoff by Streetsweeping," Personal communication to,
G. Fred Lee, Civil Engineering, Colorado State University,
Fort Collins, CO (1979).
Cairns, J. , Dickson, K., L. , and Maki, A. W. (eds.) Estimating
the Hazard' of Chemical Substances to Aquatic Life, American
Society for Testing and Materials Special Publication 657,
ASTM, Philadelphia, PA (1978).
Cowen, W. and Lee, G. F. "Leaves as a Source of Phosphorus,"
Environ. Sci. and Tech. 7_:853-854 (1973).
Cowen, W. F., Sirisinha, K., and Lee, G. F. "Nitrogen and Phos-
phorus in Lake Ontario Tributary Waters," Water, Air, and
Soil Pollution 10_: 343-350 (1978).
Jones, R. A. and Lee, G. F. "Evaluation of the Impact of Phos-
phorus Removal at the Danbury, Connecticut Sewage Treatment
Plant on Water Quality in Lake Lillinonah," Environmental
Engineering, Colorado State University, Fort Collins,
Occasional Paper No. 31, April (1978a).
Jones, R. A. and Lee, G. F. "Evaluation of the Elutriate Test
as a Method of Predicting Contaminant Release during Open
Water Disposal of Dredged Sediment and Environmental Impact
of Open Water Dredged Material Disposal, Vol. I: Discus-
sion," Technical Report D-78-45, U.S. Army Corps of Engi-
neers WES, Vicksburg, MS (1978b).
Jones, R. A., Mariani, G. M., and Lee, G. F. "Evaluation of the
Significance of Sediment-Associated Contaminants to Water
Quality," Presented at American Water Resources Assoc.
Conference, Las Vegas, NV (1979). To be published in con-
ference proceedings.
Kluesener, J. W. and Lee, G. F. "Nutrient Loading from a
Separate Storm Sewer in Madison, Wisconsin," Water Poll.
Cont. Fed. 46:920-936 (1974).
55
-------
Lee} G. F« "Ways in Which a Resident of the Madison Lakes'
Watershed May Help to Improve Water Quality in the Lakes,"
A report of the Water Chemistry Program, University of
Wisconsin (1972).
Lee, G» F. "Eutrophication," Transactions of the Northeast
Fish and Wildlife Conference, pp 39-60 (1973).
Lee, G. F. and Jones, R. A. "The Role of Biotransformation in
Environmental Hazard Assessment," To be published in Proc.
of 'Biotransformation and Fate of Chemicals in the Aquatic
Environment' workshop held in Pellston, MI, August (1979a).
Lee, G. F. and Jones, R. A. "The Role of Environmental Chemis-
try-Fate Modeling in Environmental Hazard Assessment: An
Overview," Presented at ASTM Symposium on Aquatic Toxicol-
ogy, Chicago, IL, October (1979b). To be published in con-
ference proceedings.
Lee, G. F. and Jones, R. A. "Effect of Eutrophication on
Fisheries," Prepared for and to be published by the Ameri-
can Fisheries Society (1979c).
Lee, G. F. and Jones, R. A. "Water Quality Characteristics of
the US Waters of Lake Ontario during the IFYGL and Model-
ing Contaminant Load - Water Quality Response Relationships
in the Nearshore Waters of the -Great Lakes," Report to
NOAA, Ann Arbor, MI, June (1979d).
Lee, G. F., Rast, W., and Jones, R. A. "Recent Advances in
Assessing Aquatic Plant Nutrient Load-Eutrophication Re-
sponse for Lakes and Impoundments," Environmental Engi-
neering, Colorado State University, Fort Collins, Occasion-
al Paper No. 14, May (1977). , ,
Lee, G. F., Jones, R. A., Saleh, F. Y., Mariani, G. M., Homer,
D. H., Butler, J. S., and Bandyopadhyay, P. "Evaluation
of the Elutriate Test as a Method of Predicting Contami-
nant Release during Open Water Disposal of Dredged Sediment
and Environmental Impact of Open Water Dredged Materials
Disposal, Vol. II: Data Report," Technical Report
D-78-45, U.S. Army Corps of Engineers WES, Vicksburg, MS,
1186 pp, August (1978a).
Lee, G. F., Rast, W., and Jones, R. A. "Eutrophication of
Waterbodies: Insights for an Age-old Problem," Environ.,
Sci. S Technol. 12:900-908 (1978b).
56
-------
f
Lee, G. F., Jones, R. A., and Rast, W. "Availability of Phos-
phorus to Phytoplankton and Its Implications for Phosphorus
Management Strategies," Presented at IJC-Cornell Univer-
sity Conference on Phosphorus Management Strategies for the
Great Lakes, April (1979a). To be published in conference
proceedings.
Lee, G. F., Jones, R. A., and Newbry, B. W. "Use of Bioassays
in a Hazard Assessment," Presented at American Water Re-
sources Assoc. Conference, Las Vegas, NV (1979b). To be
published in conference proceedings.
Mariani, G. M. Bioassay Assessment of Toxicity of Contaminants
in Dredged Sediment^Ph.D.dissertation,Univ..of Texas at
.Dallas, Richardson, TX (1979).
.Newbry, B. W., Jones, R. A., and Lee, G. F. ."Application of the
OECD Eutrophication Modeling Approach to Cherokee Reservoir
and Other TVA Impoundments," Report to the Tennessee
Valley Authority, Chattanooga, TN (1979).
Oglesby, R. T. "Relationships of Fish Yield to Lake Phytoplank-
ton Standing Crop, Production, and Morphoedaphic Factors,"
Journal of the Fisheries Research Board of Canada, 34; 2271-
2279 (1977).
Rast, W. and Lee, G. F. "Summary Analysis of the North American
(U.S. Portion) OECD Eutrophication Project: Nutrient Load-
ing-Lake Response Relationships and Trophic State Indices."
US EPA, EPA-600/3-78-008, Corvallis, OR, 454 pp (1978).
Sonzogni, W. C-. , Uttormark, P. C. , and Lee, G. F. "A Phosphorus
Residence Time Model: Theory and Application," Water
Research 1(3:429-435 (1976).
US EPA, Quality Criteria for Water, EPA-440/9-76-023, U.S.
Gov't. Printing Office, Washington, D.C. (1976).
Vollenweider, R. A. "Scientific Fundamentals of the Eutrophi-
cation of Lakes and Flowing Waters with Particular
Reference to Nitrogen and Phosphorus as Factors in Eutro-
phication." Technical Report DAS/CSI/68, OECD, Paris
(1968).
57
-------
THE EFFECT OF URBAN STORMWATER RUNOFF ON
THE WATER QUALITY OF LAKE JACKSON, FLORIDA
Christian Byrne, Charles R. Donahue, William C. Burnett5'
Department of Oceanography, The Florida State University
Tallahassee, Florida 32306
ABSTRACT
Analyses of the runoff from the southern end (Meginniss
Arm) of Lake Jackson revealed high concentrations of suspended
solids, dissolved nutrients , heavy metals p articulates,,, and
petro-chemicals. It is clear that the overwhelming majority
of pollutants in this end of the lake is due to urban storm-
water runoff. In response to the serious water quality in the
southern areas of Lake Jackson, the Florida Department ,of Envi-
ronmental Regulations in association with the United States
Environmental Protection Agency have proposed to construct,and ;
maintain a bio-filtration system in the watershed of Meginniss
Arm to restrict the urban pollutional loading.
The primary source of petroleum hydrocarbon in Meginniss
Arm is the waste automobile oil from the roads and parking areas
of the commercial establishments o,f the area. Lake Jackson of-
fords an area for a very interesting study of pollutant loading.
There are two watersheds in Lake Jacks:bn, which are similar in
size, topography, and geology but very dissimilar in land usu'age.,,
Ox-Bottom Creek Arm in the Northern area is primarily forested- , :
agricultural with little mechanical activity while Meginniss
Arm, is highly urban containing two shopping malls. These major
watersheds will make possible a comparison of the concentration
of hydrocarbons in the stormwater runoff from the two areas and
make possible the separation of the contribution of biogenic
and anthropogenic sources in the two areas. ! By sampling the
stormwater runoff (for dissolved and particulate forms), atmos-
pheric dust, rainfall, sediments, and water samples from several
locations of Lake Jackson, it might be possible to determine a
mass balance budget for hydrocarbons in Lake Jackson,, Florida.
INTRODUCTION
Lake Jackson, in Leon County Florida (Figure 1), is a
closed depression (1960 ha.) in an area of rapidly changing land,
usage. Urban development has increased markedly in the southern
watershed of Lake Jackson, while the northern areas have remained
essentially undisturbed with little or no development occurring.
Intensive studies of the lake and its associated watershed (Smith,
1972; Smith, 1973; Smith, 1974; Schamel et_ all. , 1974; Harris
and Turner, 1974, Turner et al., 1974) have revealed significantly
"^Department of Oceanography, The Florida State University,
Tallahassee, Florida 32306
58
-------
C1J
O)
rg
O.
E
•r-
-a
O
s-
-a
E
(O
o
CO
^:
o
fC
'-3
CD
O)
S-
LJ_
59
-------
poorer water quality in the southern areas. This decrease in
water quality can be directly related to,,,, the stormwater runoff
from the urbanized watersheds of these areas.
The watershed possessing the worst water quality is also the
one with the highest degree of urban development - the watershed
of Meginniss Arm. Intensive sampling and analysis of stormwater
runoff in the area (Harriss and Turner, 1974) demonstrated the
runoff to be a principal source of organic solids, dissolved
nutrients, and particulate forms of several heavy metals, includ-
ing lead. Recent investigations in other areas (Wakeham, 1977;
Hunter et al., 1979) have shown that stormwater runoff can be
a significant source of organic compounds and, in particular,
petroleum hydrocarbons. The object of this study was to determine
the presence of petroleum hydrocarbons in the stormwater runoff
to Meginniss Arm during a storm event and assess the concentra-
tions and forms of these hydrocarbons as they move through the
discharge channel and empty into Lake Jackson.
STUDY SITE
The study site was the urban watershed of Meginniss Arm
(Figure 2). A case study was made of .the stermwaterrunoff gen-
erated in the watershed during a storm ..event on November 27, 1978.
Table 1 indicates the general land usage and hydrologic charac-
terization of the storm event. Water grab samples were taken at
four different locations along the discharge tributary. . The first
sampling station was located below the discharge conduit at 24.8
ha. The second sampling station was located below Tallahassee
Mall (TM) discharge conduit system. The,drainage area was 202.9
ha. The third station was at the entrance to a series of small
holding ponds (P) above Interstate 10. This station received
the joint discharge from the Tallahassee Mal,,l,,,,and Northwood Mall
sheds. This total drainage was 356.9 ha. The final station was
located at the mouth of the discharge channel as it emptied into
Meginniss Arm at the southern portion of Lake Jackson,
Sampling times were previously determined from a study of
twelve hydrographs of storm events in this watershed (Harriss
and Turner, 1974). Samples were taken at the initial discharge
of runoff from the storm (time = 0 hours), 0.5 hours, 1 hour,
2 hours, M- hours and 8 hours after the initial discharge. Only
at the fourth station (MA) was this series changed, the first
sample was taken at 0.5 hour and the last 26 hours after the
initial discharge.
METHODS AND MATERIALS
Grab samples were taken at the selected time intervals using
steel cans (Firestone Challenger VI). These cans were pre-
washed with warm water and Alconox, rinsed, then washed with
single distilled methanol, and finally single distilled dichloro-
methane. All rubber seals were wrapped with teflon film to
prevent contamination.
60
-------
E
0!
OJ
M-
O
T3
Ol
-C
CO
i-
-------
r
•d H
6 Q) rd
5-i -H O
"d GO
CD S
d bO
•H Pi 0)
•d & -p
Q) O
•P CO (p.
rtf-rl O
Pi Td
Q) Q)
q-d o
CD d -H
Mrrt tLj
iy *t" i
it-) H O
CH H
O rd p
3 pi -H
rd 4->
!LI rt co
Q) -H
•P Q
rd •
|5 CD rd
g f- T3
Pi en -H
O rH i^
P O
CO ^H
c-~ U-t
CH CM
O ft*,
& -P
>. CD k
TJ ^ O
-P CD
CO > CD >,
O rd
CO f_,
rd d !>i JlJ
0 0 XI GO
.
H
H
m
<£j
EH
P
d
CD cococncocn CD CD
£-1 CMLOd-CMOO O LO
CD CM LO rH O CM
Pn rH
s~\
J-
r>-
cn
rH
0
CD
^
EH
•d
rd
CO
•rl
C-t rdrdfdrdrd rd rd
rd rd rdrdcc!rdrd .^ .*"!
ffi CD
v-^ C-i cococnr-~cM en LO
d HCOLOLOJ- CD rH
O OOrHCDd-rH LO CD
•H rH 00
P
rd
N
•H
C_j
CD
^j
O
rd
Cj
rd s~*
rd CO
O <4H
o
CD O
CO f-t
Z>
•d fao
d d
rd rd -H
J CD y.
rH d J d H d O 3 -H Td
E-M ' rd 3 o> Pi fc >rd
rH O T3 CD P Pi O
"d -H -H g CO CD Pi
O Pi CO R d ftx^'
O bO CD O O i=i
O
•H
P
rd
•H
Pi
CD
I)
T~^
O
td
P-!
rd
O
O
•H
bO
O
H
O
P)
Tj
^K,
ffi
d
o
•H
P
rd
P
•H
ft
•H
O
CD
ft CO
P rd
rd
O o
•H CM
v—/ v^/ C[—I
•H CO
O O o\° bO O
•H -H
rH H rH CO >
LO oo . CD
• • CD O f-t
CM O CM |Z| ft
rH rH
CD CD
O O
X X
d- en
CM CD
CD CM
d O
O P-<
•H
p p
rd rd
co
O
•H
P
•H
P
•H
13 < PS O O
H
•H
O
CD
PL,
rd
P
O
O
O
H P
CD d
PS -H CD
>-i 'd
d CD
CD c_, o
bO 0) CD
3 p p
rd rd PI
CD S£ <
62
-------
The water grab samples were separated into two fractions
(dissolved and particulate) and prepared for analysis as follows:
Dissolved Fraction
The dissolved fractions were filtered through pre-combusted
(500°C; 12 hours) GF/F glass filter filters (142 mm) under pre-
purified nitrogen gas pressure. The dissolved fraction was
transferred into separatory funnels and acidified below 2 pH and
extracted three times with dichloromethane at a ratio of.10:1.
This total organic extract (TOE) was dried over anhydrous MgSO^,
reduced to dryness and weighed on a MicroMettler M5 balance to
the nearest microgram.
Particulate Fraction
The particulate fractions were trapped on pre-combusted GF/F
filters (effective efficiency - 1.0 urn). The filtegs were then
folded, placed in aluminum foil, and frozen at - 80 C until ex-
traction. The filters were thawed at the time of extraction,
cut into 0.5 cm stips, and placed into glass thimbles for
Soxhlet extraction. The filters (particulates) were extracted
initially with methanol for 12 hours and then 12 hours with di-
chloromethane. The methanol extracts were back extracted twice
with hexane. The hexane and:dichloromethane extracts were com-
bined, dried over anhydrous MgSO^ and reduced to dryness under
a. vacuum. These extracts constituted the total organic extract
(TOE). The TOE were dried and weighed as in the case of the
dissolved fraction. From this point both fractions were treated
identically.
All samples were saponified in round bottom flasks with a
mixture of KOH/methanol: benzene: deionized water (Van Vleet
and Quinn, 1977). After refluxing the solutions for at least
four hours and then coecling, the samples were transferred to
separatory funnels where the phases were separated by addition
of deionized water. The non-saponificable extract (NSE) was
extracted twice with benzene, dried and weighed. The NSE was
then taken up in hexane for column chromatography. The NSE
was applied to activated silica gel columns with a column to
sample ratio of 200:1. The columns were previously charged with
four columns volumes hexane. The columns were successively elut-
ed with two column volumes of hexane to remove the aliphatic
species and with two column volumes to remove the aromatic
species. The two eluajits were transferred to pre-weighed micro-
vials, reduced, and weighed. The two eluants were then taken
back up in their respective solvents.
The aliphatic eluants of the dissolved and particulate
fractions were analyzed by gas-liquid chromatography (GLC) with
a Hewlett-Packett Model 5711 gas chromatograph equipped with dual
63
-------
flame ionization detectors and dual 6 ft X 1/8 in. stainless
steel columns containing 10% SP-1000 mesh Chromosorb W (HP)
The columns were temperature programmed from 80° to 250° at
a rate of 4°C/minute. The injection and detector temperatures
were 250°. The nitrogen carrier gas flow rate was 51 ml/min.
The hydrocarbons measured eluted from the columns between
nC12-nC34. The chromatographs were examined for CPI (Carbon
Preference Index) values and the presence of unresolved en-
velopes. Both are considered standard methods of assessing
the presence of petroleum hydrocarbons (Blumer and Sass, 1972).
All solvents used were doubly distilled in all glass distil-
lation systems unless otherwise stated. All glassware was washed
with Alconox, rinsed, submerged in Non-Chromix-sulfuric acid
solution for at least four hours, rinsed with water, air dried,
rinsed with single distilled dichloromethane, and dried at 150°C
overnight.
RESULTS
Concentrations of total dissolved hydrocarbons (TDH) and
total particulate hydrocarbons (TPH) for the November, 1978
storm event are given in Table 2. These total concentrations
were determined by the addition of the dry weights of the ali-
phatic and aromatic eluants of each fraction. Although dry
weights of organic extracts often lose low boiling-low mole-
cular weight compounds, gravimetric determinations of the con-
centrations of these extracts are considered,..mgre accurate
than analysis of the gas-liquid chromatographs. Very high
boiling-high molecular weight compounds do not elute from the
chromatograph columns and as such would not be included in the
final concentration values. In addition, the unresolved envelopes
that occur in chromatographs can not be easily quantified by
planimetry or computer assessment.
Dissolved Hydrocarbon Concentrations
An examination of Table 2 reveals that the TDH concentra-
tions ranged from 19 to 72 ug/1 (ppb). Although the range does
not appear to be very great, there is a significant maximum
wh±9h occurs two hours into the storm event. This could be
the result of bleeding or leeching of dissolvable hydrocarbons
from the impervious (i.e. asphalt) surfaces. This relationship
is shown clearly in Figure 3. This figure presents data for
the station at Tallahassee Mall (TM) during the storm event.
There is a definite relationship between rainfall, discharge,
and hydrocarbon concentrations.
Particulate Hydrocarbon Concentrations
The total particulate hydrocarbon (TDH) concentrations
64
-------
/— X
°Jj
s
p
'R
•S
(L)
P
1 1
^j
O
•H
"Fj
TO
H
frt
-H
1
ffi
V^»
?
o *
5s-
^1 £>
r^H
t|_J
"§ °
> CO
H P
0-iH
co C
co p
TJ Q
'rci CO
O ro
P
CO
0 O
•rH
CO -P
C rci
'•P S
S 8
§ o
CJ
d l — 1
O -^^
i — 1 ("I"!
o p
CO EH
CO N^
•H
O
O
I)
jrj
&
g
^M
O
•H
-p
&
•rT
^
CO CO HT
CO !>. J-
OO rH
O
C~~ J" CD
CM O H
CD rH
CD CD LO
LO CD CM
jj-
fNJ
OO OO rH
H CM CM
O CD CM
rH rH rH
00 CM CD
H
O O O
CO O CO
CD O O
rH rH
«
rH CM CO
IS S 2
Fi P EH
Jd- rH CM
0 HT j-
CD rH
CM OD CD
t-- LO rH
CM
CM OO OO
OO CO CM
OO
CO rH CD
d- CO CM
CD CD CD
rH rH
l>- CM O
CM CM rH
0 0 O
0 CM 00
rH CO C^
rH rH rH
J- LO CD
£H £-, 5-,
P Fi F)
LO l-O CD CM
CD CD CD CD
OO Csl C-- CD
CD rH
rH
CO rH 00 J-
J- CO CD CO
O LO CM LO
rH
CM J- CM OO
CM rH CD CM
co r-~ HT rt-
LO-
rH
H CD d- oo
C-- LO CM CD
CD CM E — OO
CM 00 rH CO
CM zt- C~- O
HT CM co
LO LO LO LO
OO O CO CO
CD O CD H
rH rH rH
rH CM 00 ^t"
!§ S ^ IS
^ ^ ^^
CM LO 0
c^ m co
LO CM CO
CM rH
H
H d- LO
OO LO CM
CO rH CD
CD
rH
rH rH LO
Zf 00 LO
c~- J-
H
*
•S
1 CO rH
I LO CM
CM CM
| CO CD
rH rH
O CD CD
.H- rH -Hr
CD O 0
rH rH
rH CM CO
P-) PH CL,
CM CM t~
r-~ t~- co
OO CD
rH O CO
— J" (~~1
CM OO
rH l>- CD
OO H CM
rH CO
CO CD O.
LO CO OO
HT CD HT
CM rH
CD O CD
CM 00 rH
O LO O
d- co HT
rH OO r-.
rH rH rH
d- LO CD
P-i OH DH
rH CO CM
CO O LO
rH rH
CD CO rH
rH 00 rH
CD J-
CO CO CD
CD J- rH
CO CD rH
rH
CD oo r-^
00 00 CM
rH LO J-
CM rH rH
CO CO OO
rH rH rH
0 O O
CM LO LO
0 O rH
rH rH rH
rH CM OO
s s s
CM co en
00 d- CM
rH rH rH
CD CD C--
r- r- oo
OO CM OO
IS) t^~ LO
O CD OO
.HT zj- LO
LO CM CO
rH CM CM
LO J- LO
CM CM CO
OO
CM
\
rH
rH
O CD O
CO LO OO
00 C~~ rH
rH rH rH
J- LO CD
s s s
T_i
f^
fd
•o"
g
Pn
II
p
€\
H
g
0
^
i-Cl
• S
II
g
r\
rH
| 1
s
cu
CO
•i
rH
, — i
rn
H
II
5^
§H
, .
0)
3
• P
CO
•H
^P
pj
•rH
CD
CO g
gco
O CO
•H -H
s|
> M
Jj S
^q n
<• ^
* S
"S
u*
g
s
CL)
T^
-p
o
e
-K
*
65
-------
c
'E
£ O
jjai
ft02
•i
-------
were at least a factor of 2 greater than the total dissolved
hydrocarbon (TDH) concentrations throughout the storm. In the
initial runoff discharge at station NW#1, the difference was
two orders of magnitude. Unlike the TDH concentrations the
TPH concentrations reach their highest values at the beginning
of the storm surge. After this surge had passed-, the TPH
mimicked the TDH concentrations with a secondary maximum at
two hours into the storm at which point the concentrations
began to plateau out very similarly to the TDH. The TPH as well
as the TDH concentrations are both shown in Figure 3.
There appears to be no subsequent increase in either
concentrations later into the storm with the additional dis-
charge that occurs after 1300 hours. This may indicate that
with continuous rinsing of the pavement no further hydrocarbons
can be easily transported in the particulate or dissolved form.
Analysis of the gas-liquid chromatographs of the aliphatic
eluants of both the particulate and dissolved fractions reveal
that petroleum hydrocarbons are definitely present in: these
fractions and that they are primary sources of the hydrocarbons
present. Figure 4 shows chromatographsat various stages dur-
ing the storm at the Tallahassee Mall station. The CPI (Carbon
Prefernce Index) for the first chromatograph (Figure M-a) is
approximately 0.99 CPI is a weighted measure of the ratio of
©dd carbon to even carbon hydrocarbons which can be useful in
distinguishing natural from anthropogenic sources. Samples of
natural biogenic hydrocarbons possess CPI's in a range from
1..5.-8 . Anthropogenic hydrocarbons (i.e. petroleum hydrocar-
bons ) have CPI values approaching unity (1) or less. Figure M-b
is a chromatograph of the aliphatic eluant of the TDH at Station
TM two hours into the storm event. The CPI is 1.5 and, in addi-
tion, a unresolved hydrocarbon envelope, often indicative of pe-
troleum hydrocarbons (Blumer and Sass, 1972),is present. This en-
velope, together with the CPI value, make a good case for the
presence of petroleum hydro.carbons in this sample. The fact that
this envelope appears after only two hours of flushing of the
surface of the watershed is another factor that would support the
theory of bleeding or leeching of the asphalt occurring, into the
stormwater runoff. A chromatograph of the aliphatic eluant of
the TPHs from station TM taken at time 0 is shown in Figure 4c.
The CPI value of 0.96 and the unresolved envelope give vali-
dity to our hypothesis that petroleum hydrocarbons are present in
the particulate fraction of the stormwater runoff very early
during the course of the storm.
DISCUSSION
Wakeham (1977) studied urban stormwater runoff which flows
into Lake Washington from surrounding areas of Seattle, Washington.
The values of total aliphatic hydrocarbons in the urban stormwater
67
-------
C22 24 26 28 3O
25.27
32
unresolved
envelope
\ j" B 3 •
TIME
Figure 4. Gas-liquid chromatographs of
(a) Tallahassee Mall (9:30 - initial discharge
surge); total dissolved aliphatic hydro-
carbon fraction
(b) Tallahassee Mall (9:30 - initial discharge
surge); total particulate aliphatic hydro-
carbon fraction.
(c) Tallahassee Mall (11:30 - two hours after
initial surge); total dissolved aliphatic
hydrocarbon fraction.
68
-------
runoff ranged from 200 to 7500 ug/1 with a mean value of 1200
ug/1. These values were total aliphatic hydrocarbons as the
particulate fraction was not separated from the dissolved.
Wakeham's total aliphatic concentrations are comparable with
the results of this study. The total aliphatic hydrocarbons
concentrations reported here range from 36 to 15000 ug/1.
Hunter ejt al. (1979) studied five storm events and the asso-
ciated urban stormwater runoff of a watershed of Philadelphia,
Pennsylvania. The average total hydrocarbon concentrations for
the five storms was 3.69 mg/1, with a total particulate concen-
tration (3.29 mg/1) accounting for 82.0% of the total. His
value for the average total dissolved hydrocarbon concentra-
tion was 400 ug/1, an order of magnitude greater than the values
for this study. Different methodology was followed, however,
and could be the reason for the variance. The trends for the
particulates and dissolved hydrocarbon concentrations, were
very similar to,ours. The total particulate hydrocarbons
concentrations peaked dramatically at the initial surge of
the runoff, diminished, and peaked a second time, later in the
storm event. The total dissolved hydrocarbon concentrations
were initially low and peaked just before or just at the sec-
ondary peaking of the total particulate hydrocarbons. The
particulate fraction of the water samples possessed the great-
est hydrocarbon concentrations. Various studies have shown
(Meyers and Quinn, 1973, Khan and Schnitzer, 1972, Meyers and
Oas,1978) that particulate material such as' clay particles
or organic detritus can hold organic species by adsorption and
absorption and act as vectors in the movement of these organic
species through the hydrologic cycle. This could be one factor
in the higher concentrations of hydrocarbons existing in
particulate form. In addition, continuous weathering of
biogenic and anthropogenic hydrocarbons in the environment re-
sult in these hydrocarbons becoming considerably more insol-
uble in water. As the particulate hydrocarbons are flushed
from the discharge channel into the lake, these particulates
rain through the water column and deposit themselves onto the
sediment layer. Since these forms have been heavily weathered,
they could be persistent and incorporate into the sediment
column. (Wakeham, 1977; Sheldon and Kites, 1978). Further
studies are progressing to model the petroleum hydrocarbon
budget of Lake Jackson.
69
-------
REFERENCES CITED
Blumer, M. and J. Sass. 1972. Indigenous and petroleum-derived hydrocarbons
in a polluted sediment. Mar. Pollut. Bull. 3: 92-94.
Giger, W., M. Reinhardt, C. Schaffner, and W. Stumm. 1974. Petroleum-derived
and" indigenous hydrocarbons in recent sediments of Lake Zug, Switzerland.
Environ. Sci. £ Tech. 8: 454-455.
Harriss, R.C. and R.R. Turner. 1974. State of Florida. Game and Freshwater
Fish Commission. 1973-1974. Job Completion Report. Lake Jackson Invest-
gation. 231 p.
Hunter, J.V., T. Sabatino, R. Gomperts, and M.J. MacKenzie. 1979. Contri-
bution of urban runoff to hydrocarbon pollution. Journal Water Pollut.
Control Fed. 51(8): 2129-2138.
Khan, S.U. and M. Sennitzer. 1972. The retention of hydrophobia organic
compounds by humic acids. Geochim. et Cosmochim. Acta. 36: 745-754.
Meyers, P.A. and T.G. Oas. 1978. Comparison of association of different
hydrocarbons with clay particles in simulated seawater. Environ. Sci.
S Tech. 12: 934-937.
Meyers, P.A. and J.G. Quinn. 1973. Factors affecting the association of fatty
acids with mineral particles in seawater. Geochim. et Cosmochim. Acta.
37: 1745-1759.
Schamel, S, G.A. Davidson, and T.J. Casey. 1974. "Urban" sediment in.Lake
Jackson, Florida - a case study in sediment pollution (abstract).
Proc. Geol. Soc. Amer. Southeast Section Meeting. April 4, 1974.
Atlanta, Ga. p. 396.
Sheldon, L.S. and R.A. Hites 1978. Organic compounds in Delaware River.
Environ. Sci. & Tech. 12: 1188-1194.
Smith, S.L. 1972. Lake Management and Research - Lake Jackson Studies.
Annual Progress Report, Dingell-Johnson Project F-12-13. Florida Game
and Freshwater Fish Commission. Tallahassee. 68 p.
, 1973. Lake Management and Research - Lake Jackson Studies. Annual
Progress Report, Dingell-Johnson Project F-12-13. Florida Game and
Freshwater Fish Commission. Tallahassee. 68 p.
70
-------
- .. 1974. Lake Management and Research - Lake Jackson Studies.. Annual
Progress Report, Dingell-Johnson Project F-12-15. Florida Game and
Freshwater Fish Commission. 94- pp.
Turner, R.R., R.C. Harriss, T.M. Burton, and E.A. Laws. 1974. The effect
of urben land use of nutrient and suspended solids export from north
Florida watersheds. In: Mineral Cycling in Southeastern Ecosystems.
Proc. Mineral Cycling Symposium. May 1-3, 1974. Augusta, Ga.
Van Vleet, E.S. and J.G. Quinn. 1977. Input and fate of petroleum hydrocarbons
entering the Providence River and Upper Naragansett Bay from • wastewater
effluents. Environ. Sci. £ Tech. 11: 1086-1091.
Wakeham, S.G. 1977. A characterization of the sources of petroleum hydrocarbons
in Lake Washington. Journal Water Pollut. Control Fed. 19: 1680-716^7.
71
-------
A COMPARISON OF RAIN-RELATED PHOSPHORUS AND NITROGEN LOADING
FROM URBAN, WETLAND, AND AGRICULTURAL SOURCES
R.P. Glandon, F.C. Payne, C.D. McNabb, and T.R. Batterson
Department of Fisheries and Wildlife
Michigan State University
East Lansing, Michigan 48824
ABSTRACT
Comprehensive watershed studies have been conducted for two
lakes located in the Lake Michigan drainage system. Studies
were conducted from March through October of 1979. During that
interval, large differences in storm-related nutrient loading
were measured from urban, wetland, and agricultural sources.
Eliminating runoff due to melt of the snow pack, it was found
that rain-related discharge from the urban area studied was
0.578 kg total-P and 3.688 kg total-N ha-1. Rain
induced runoff from marshes in the same drainage basin trans-
ported 0.023 kg total-P and 0.585 kg total-N ha~l.
Rainfall of approximately the same amount caused runoff from
agricultural land of 0.180 kg total-P and 5.965 kg total-N ha"1.
Algae of both lakes were phosphorus limited; nitrogen was
present in excess. Using constants from Nichols-Dillon relation-
ships in the literature regarding phosphorus, phytoplankton
biomass, and secchi disc transparencies, the urban input of
phosphorus ha~l of drainage was sufficient to bring 0.96 ha-m of
lake water to undesirable algal bloom status. Similarly, marsh
input ha~l would bring an estimated 0.04 ha-m into bloom. By
the same calculation, storm-related agricultural runoff would
result in 0.30 ha-m of lake water becoming undesirably rich in
algae. Knowing the number of hectares in these types of catch-
ment and the volume available in a particular lake for phyto-
plankton production, decisions regarding cost-effective treat-
ment of rain-related discharge can be made.
72
-------
"I
.'*•
INTRODUCTION
For a majority of lakes, the most important nutrient factors
causing an increase in phytoplankton populations are phosphorus
and nitrogen (Vollenweider, 1968). Consequently, a significant
portion of the limnological literature has been devoted to the
comparison of the annual export of phosphorus and nitrogen from
different watershed types (Danigian and Crawford, 1977; Dillon
and Kirchner, 1975; Colston, 1974). In general, the studies
show that almost every constructional or agricultural activity
on a watershed increases nutrient loading in drainage water.
The urbanized watershed has been characterized as an outstanding
exporter of nutrients (1.10 to 16.60 kg TP ha'lyr"1) to surface
waters. While untreated,sewage water contains the highest
nutrient levels, urban storm water runoff alone typically exports
nutrients at a rate well above that of other watershed types
(Weibel, 1969). Urban fertilization of surface water diminishes
valued supplies of suitable quality water for domestic and
industrial purposes and frustrates attempts to meet the increased
demands for recreational waters.
The potential of watershed runoff to maintain phytoplankton
productivity in a receiving lake is largely determined by the
level of nutrients exported throughout the summer when temp-
erature and light regimes are optimal. A sustained nutrient in-
put can maintain high productivity by replacing particulate
phosphorus and nitrogen sedimented throughout the season (Wetzel,
1975). Annual loading estimates from watersheds in the seasonal
north temperate region are often dominated by quantities carried
in a spring pulse of discharge. For this reason, annual esti-
mates tend to mask the degree of loading to a lake during the
period of stratification.
In this paper, we examine phosphorus and nitrogen losses
from marshland, urban, and agricultural watersheds with an
emphasis on rain induced loading. This component of surface
drainage is used as a means of comparing the watershed types
with respect to the ability to maintain phytoplankton produc-
tivity in a lake throughout the summer.
73
-------
STUDY AREAS AND METHODS
Lake Lansing, Ingham County, Michigan has a low relief
watershed of ca.842 hectares. Runoff from approximately 776 ha
of the watershed moves to the lake through adjacent marshes and
enters the lake through six discrete surface streams. Street
drains carry runoff from ca.2.5 ha of residential and commerical
areas in the watershed.
Skinner Lake, Noble County, Indiana receives runoff from a
rolling agricultural watershed of 3,813 ha. Runoff characteris-
tics were determined from data collected from the Rimmel stream
which carries water to the lake from approximately 3,090 ha of
the watershed.
Discharge of each marsh stream to Lake Lansing was measured
three times per week from the onset of flow (March 5, 1979) to
cessation (June 8, 1979). Discharge rates of four of the streams
were determined with standard use of a Pigmy-type Gurley current
meter. Rectangular weirs were installed across the remaining two
streams and discharge was calculated by applying stage height
readings to U.S.G.S. approximation formulae (U.S.D.I., 1967).
Water samples were collected from each stream each time
discharge was measured and aliquots were composited according to
a discharge-proportional scheme devised to reduce the number of
samples to be analyzed from 36 (6 stream samples x 3 collections
per week x 2 weeks) to 6 (1 two-week discharge-proportional com-
posite per stream). Central to the compositing scheme was the
idea that each sample (eg. collected at time ti) represented the
volume of water discharged during the time, interval extending
from halfway to the previous sample collection (ti-to/2), to
halfway to the following sample collection (t2~ti/2). This
method accounts for the possibility of detecting changes in water
quality resulting from an observed change in discharge.
If the frequency of measurement is high relative to the rate
of fluctuation in flow, we can assume a linear change from one
measured discharge value to the next. The volume of water dis-
charged represented by the sample collected at t]_ (shaded area
in the illustration),
Q
74
-------
can be calculated:
v = QQ+3Q1
8
3Q1+Q2
8
(1)
where, V = the volume of water discharged, and QQ = the discharge
rate measured at time 0 (to). An aliquot from a water sample
representing twice the volume of discharge as that of a second
sample would .be twice as large. Discharge proportional aliquots
were added to a composite for intervals of two weeks, after which
composite was analyzed for total phosphorus and total nitrogen.
"Running" composites were acidified (2 mis cone. H2SO4 per liter)
and refrigerated. Six composite samples (one per marsh stream)
were submitted to the laboratory biweekly.
During the interval of interest, beginning on March 5, the
urban drains of the Lake Lansing watershed discharged only
during rain events. Relationships between total urban drain
discharge and amount of rainfall were determined by plotting the
total volume of discharge against the corresponding amount of
rainfall for each of several.monitored storms. The total volume
of water discharged was calculated from periodic measurements of
the discharge rate of each drain throughout the monitored storm.
Weirs were installed to facilitate the measurement of discharge
rates. Water_ samples collected during the beginning, middle,
and end of each monitored storm were composited according to a
discharge-proportional scheme analogous to that described for
the marsh stream samples. The total amount of rainfall was read
from a rain gage set at Lake Lansing. .
Discharge of the RimmeL stream in the Skinner Lake water-
shed was measured daily from 2/23/79 to 4/15/79 with use of a
Gurley current meter. Water samples, collected each.time dis-
charge was measured, were composited and stored in the same
manner as those collected from the Lake Lansing watershed. After
4/15, discharge was measured at two week intervals and grab
samples were returned to the laboratory for .immediate analysis.
Rainfall data were obtained from an automatic precipitation
recorder located near the lake.
Total phosphorus concentrations were determined colorimet-
rically using the single reagent-ascorbic acid method on unfil-
tered, digested samples.
Total nitrogen is defined as the sum of the nitrate-nitrite
and total Kjeldahl nitrogen. Nitrate-nitrite nitrogen was de-
termined colorimetrically by the cadmium reduction method. Total
Kjeldahl nitrogen was determined colorimetrically after digestion
and distillation of an unfiltered sample. All values are reported
as P and N (A.P.H.A., 1975).
75
-------
RESULTS , :
The summed discharge rates of the six marsh streams that
empty into Lake Lansing reached a maximum of 257 1 sec"1 on
3/9/79 and fell to zero by 6/13/79 (cf. Figure 1). Early spring
fluctuations in the marsh hydrograph correspond to freezing
temperatures on 3/11 and from 3/14 to 3/16, and thawing temper-
atures on 3/19. Thereafter during the spring, the fluctuations
in the discharge rate of the marsh correspond to wet and dry
periods. During the interval from 3/5 to 3/16 the runoff co-
efficient, (i.e. m^ of observed discharge/m of rainfall x m2 of
catchment area) was 1.73, indicating that snow melt was contri-
buting to the observed discharge. Prom 4/28 to 5/11 the runoff
coefficient was 0.36, and from 5/11 to 6/8 the coefficient fell
to 0.09.
The peak discharge of 257 1 sec"1 resulted from the melting
snow pack. Separation of runoff due to snow melt from that due
to rain events was done by describing the expected recession from
a discharge of 257 1 sec"1 had there been no spring rain. It was
assumed that the expected stream flow recession can be character-
ized by those segments of the hydrograph that were undisturbed
by rainfall (Barnes, 1940; Davis and DeWeist, 1960). By isolat-
ing dry weather hydrograph segments (eg. from 5/4 to 5/11 of
Figure 1), the undisturbed recession could be described by:
Q = (a) (1.14)
-t
(2)
If the recession curve described by this equation is applied to
the peak discharge of 3/9, then:
Q = (257) (1.14)-ti
(3)
,-1
where, Q is the discharge in liters sec •L, and t is the number of
days beyond 3/9. This analysis allows for the separation of
total discharge into the component associated with water collect-
ed over winter and that induced by spring rain and enables a more
meaningful comparison of rain induced runoff between watersheds.
The occurrence of freezing temperatures in the middle of
March caused the discharge rate to fall below that expected from
the recession curve. This was followed by a warming trend and a
spike in discharge, in spite of a lack of rainfall. The area
under the curve, between the observed and expected rates of dis-
charge, from 3/10 to 3/16 was used to estimate the volume of
water held back by freezing temperatures. This volume of re-
tained water is considered to be a part of the discharge asso-
ciated with meltwater (cf. stippled areas of Figure 1).
Total phosphorus and total nitrogen content in composited
water samples was used with separated components of marsh dis-
charge to yield estimates of nutrient export from the marsh
(Table 1). The intervals of measurement correspond to composit-
76
-------
CNJ
CO
CO
I I I I
CXI
oo
-p
W
N tn
•H fl d (0 o
77
-------
Table 1. Total and rain induced loading* of total phosphorus
and total nitrogen from marshland in the Lake Lansing
watershed (1979).
Interval of
Measurement
3/5-3/16
Total
Rain
3/17-3/30
Total
Rain
3/31-4/13
Total
Rain
4/14-4/27
4/28-5/11
5/12-6/8
3/5-6/8
Total
Rain
Total Discharge
m3xlO~4 m3ha~l
13.
4
170
Total Phosphorus
kg kg ha~l
10.5
0.013
Total Nitrogen
kg kg ha"-*-
166.
1
0.
214
nil - - -
10.
3.
11.
9.
7.
9.
2.
54.
31.
8
1
6
4
0
7
4
9
6
140
40
150
120
90
130
30
710
410
5.7
1.6
7.9
6.4
4.3
4.8
0.4
33.6
, 17.5
0.007
0.002
0.010
0.008
0.006
0.006
0.001
0.043
0.023
134.
38.
141.
114.
95.
164.
43.
745.
455.
8
2
1
2
9
5
1
5
9
0.
0.
0.
0.
0.
0.
0.
0.
0.
174
049
182
145
123
212
056
961
585
* Rain induced loading = total loading after 4/13.
78
-------
ing intervals. Estimates of loading of TP and TN were divided by
the number of hectares of watershed to normalize the data for
comparison. During the entire period of flow, the marshes re-
leased a total of 0.043 kg TP ha~l and 0.961 kg TN ha~l. Of
that, 0.023 kg TP ha""1 and 0.589 kg TN ha~l was released as a
result of rainfall during the period.
A comparison of the total discharge from the urban drains
during monitored events and the corresponding amount of rainfall
reveals two distinct relationships, one associated with the
wet spring season and another with summer and fall (Figure 2).
The correlation coefficient (r ) between urban drain dis-
charge and spring rainfall was 0.87, and the relationship can
be described by:
V = 6.89 + 7.76x, (4)
where, V is the total volume of discharge from a spring storm
and x is the corresponding amount of rainfall in mm.
The discharge-rainfall relationship in the summer and fall
reflects a lower runoff coefficient. This relationship exhibited
a correlation coefficient of 0.96, and is described by:
V = -6.45 + 4.73x.
,(5)
The linear description of the spring relationship was
applied to the local rainfall record from 3/5 to 6/8 to estimate
the volume of urban drain discharge. The summer-fall relation-
ship was applied to the amount and occurrence of rainfall in the
interval from 6/9 to 10/22 (Table 2).
Total phosphorus and total nitrogen content in composite
samples collected during monitored events averaged 0.547 mg TP
1-1 and 3.493 mg TN I"1 with one standard deviation of 0.149 and
1.120 respectively. There was no statistically significant
difference («<.0.05) between the elemental concentrations in spring
and summer-fall samples. The average TP and TN levels of all
samples were used with urban drain discharge to calculate nutrient
export.
For the entire season under consideration (3/5 to 10/22),
the per ha loading of TP was 25 times the rain induced marsh
loading. Urban drains carried nearly 6 times the rain induced
TN loading from the marsh.
The discharge hydrograph of the Rimmel stream in the
Skinner Lake watershed shows a dramatic elevation in discharge
rate during the first week in March and falls to a stable flow
by the middle of May (Figure 3). Separation of the Rimmel
hydrograph was done in a manner similar to that of the marsh
hydrograph. Dry weather segments (eg. 4/12 to 4/15) were con-
79
-------
f
\
rH 4J
I I I I I
I I I I
o
m
o
o
o
m
01
0)
•H
En
39UVHOSIQ 1V101
80
-------
Table 2. Rain induced loading of total phosphorus and total
nitrogen to Lake Lansing from urban street drains
(1979).
Interval of
Measurement
3/5-3/16
3/17-3/30
3/31-4/13
4/14-4/27
4/28-5/11
5/12-6/8
6/9-6/23
6/24-7/8
7/9-7/23
7/24-8/8
8/9-8/23
8/24-9/7
9/8-9/29
9/30-10/14
10/15-10/22
Total Discharge
m3 m3ha~ 1
56
177
343
225
311
399
139
437
132
142
143
42
0
76
18
22
71
137
90
124
160
56
175
53
57
57
17
-;0
30
7
Total Phosphorus
kg kg ha"1
0.031
0.097
0.188
0.123
$
0.170
0.218
0*076
0.239
0.072
0.078
0.078
, 0.023
-
0.042
0.010
0.012
0.039
0.075
0.049
0.068
0.087
0.030
0.096
0.029
V0.031
0.031
0.009
-
0.017
0.004
Total Nitrogen
kg kg ha~l
0.196
0.618
. 1.198
0.786
1.086
1.394
0.486
1.526
0.461
0.496
0.499
0.147
-
0.265
0.063
0.078
0.247
0.479
0 . 314
0.434
0.558
0.194
0.610
0.184
0.198
0.200
0.059
-
0.106
0.025
3/5-6/8
1511
604
0.827
0.331
5.278 2.111
3/5-10/22
2640
1056
1.445
0.578
9.221 3.688
81
-------
(ma)
00
I n i i j ri
rH
m
-P O
H O
3 0) •
U S-) (Ti
•H r~
0) Cft
-------
sidered in order to describe the expected rate of recession of
discharge in the absence of rainfall. The expected decay of the
discharge rate from the level observed on 3/6 can be described
by:
Q = (1.25) (0.74)t, (6)
where, t is the number of days after 3/6.
TP and TN concentrations in composited samples were used
with separated components of discharge to calculate nutrient
export from the agricultural land (Table 3). The rain induced
loading over the 2/23 to 10/22 interval was 0.180 kg TP ha~l and
5.965 kg TN ha-1.
In summary, the normalized rain induced loading from ;2/2 3/79
to 10/22/79 in kg ha"1 was:
TP: 0.023, r^ < 0.180, . , . .< 0 . 587 ,•,-.-•
(marsh) (agriculture) «•-'« (urban)
TN: 0.585, .. < 3.688, , \ < 5.965, . , . •* .
(marsh) (urban) (agriculture)
DISCUSSION
In comparison, the marshland discharge had the lowest capac-
ity to support phytoplankton productivity throughout the summer
based on the low level of total nutrient contribution and the
early-summer termination of discharge. By contrast, the urban
and agricultural watersheds exported roughly 10 times the
nitrogen and phosphorus in sustained loading throughout the
summer and fall. The significance of this nutrient loading can
be given perspective by examining the potential impact on popu-
lations of phytoplankton in a receiving lake.
Other studies of nutrient levels in surface waters have
shown that phosphorus concentrations are generally the phyto-
plankton yield-limiting factor (Edmondson, 1972; Schindler,
1974). As a result, a number of empirically defined relation-
ships between measured phosphorus and the abundance of phyto-
plankton in lakes have been reported (Dillon and Rigler, 1974).
They are intended for use as predictive tools. Through an ex-
amination of the available literature, Nichols and Dillon (1978)
have found that the most well defined relationship is that be-
tween summer total phosphorus concentration in the euphotic
zone of a lake and the average summer phytoplankton cell volume.
The relationship between these variables is linear and can be
described by:
phytoplankton cell volume = 0.21x - 2.6, (7)
where, x is the total phosphorus concentration in mg m~3, and
83
-------
Table 3. Total and rain induced loading* of total phosphorus and
total nitrogen from agricultural land in the Skinner
Lake watershed (1979).
Interval of
Measurement
2/23-3/6
Total
Rain
3/7-3/31
Total
Rain
4/1-4/15
4/16-4/30
5/1-5/14
5/15-5/28
5/29-6/12
6/13-6/25
6/26-7/9
7/10-7/23
7/24-8/5
8/6-8/20
8/31-9/6
9/7-9/25
9/26-10/13
2/23-10/13
Total
Rain
Total Discharge
m3xlO~5 m3ha~l
11.2
4.6
11.1
3.5
9.6
3.2
2.6
0.9
0.5
0.4
0.4
0.4
0.5
0.8
0.8
0.9
0.9
44.2
30.0
362
150
358
114
311
103
84
30
15
14
14
13
15
25
27
29
28
1428
972
Total
kg
478
198
285
90
170
21
20
8
2
1
3
4
1
14
8
6
10
1031
556
Phosphorus
kg ha~l
0.155
0.064
0.092
0.029
0.055
0.007
0.006
0.002
0.001
0.000
0.001
0.001
0.000
0.005
0.003
0.002
0.003
0.333
0.180
Total Nitrogen
kg kg ha"1
8362
3468
9313
2957
7613
1295
1291
310
136
179
245
110
81
348
96
177
128
29556
18306
2.705
1.122
3.013
0.957
2.463
0.419
0.418
0.100
0.044
0.058
0.079
0.035
0.026
0.113
0.031
0.057
0.041
9.602
5.965
* Rain induced loading = total loading after 3/31.
84
-------
the cell volume is expressed in mm3 1 -'-. Nichols and Dillon
further present a well defined relationship between cell volume
and resultant secchi disc visibility. This relationship was
shown to be non-linear, and can be generally described by:
secchi disc visibility = (3.20) (1.08)~x,
where, x is the phytoplankton cell volume and secchi disc vis-
ibility is expressed in meters. ...
Using these relationships and selecting a measure of
nutrient-related water quality that is undesirable, for example
a secchi disc transparency of 1.5m or less, the impact of
nutrients discharged to lakes from urban, marsh, or agricultural
land can be predicted. The Nichols-Dillon relationships are
such that at a secchi disc measurement of 1.5m, the predicted
volume of phytoplankton cells would be near IQsnm3 1~1. This
corresponds to a total phosphorus concentration of 60 mg m~3.
If this level of phytoplankton density is chosen as an
index of minimum acceptable water quality, then by our data the
rain induced total phosphorus loading from one hectare of marsh-
land would be sufficient to degrade 0.04 ha-m of phosphorus
limited lake volume to that level. By the same calculation, the
discharge from one hectare of the agricultural watershed would
adversely affect 0,30 ha-m of lake volume. By contrast, summer
runoff from the urban watershed would debase 0.96 ha-m of lake
volume, roughly 3 times the impact of the agricultural watershed
and 24 times that of the marshland.
Assuming that watershed export of phosphorus increases ,
linearly with the size of the watershed, comparisons like 'these
can be used to predict the impact on a lake that would occur if
a watershed were converted from one type to another. For example,
our export estimates from marsh, agricultural, and urban areas
were used with the Nichols-Dillon relationships and the criterion
that a secchi disc transparency of 1.5m or less constitutes
undesirable nutrient-related water quality, to predict the volume
of phosphorus limited lake water adversely impacted by urbanizing
either marsh or agricultural land. The result is Figure 4.
The figure shows that for a given watershed area, the volume
of receiving water adversely impacted (minimum receiving volume)
increases to an upper limit as urbanization is increased,towards
100%. In approaching the upper limit, the figure suggests that
a given increase in urbanization of marshland would cause a
greater change over previous conditions than would the same in-
crease in urbanization of agricultural land.
While our dataware site specific to a degree, the technique
has general applicability. The approach can be used as a basis
for decisions regarding cost effective treatment of existing
rain-related discharges, or decisions regarding the need for
85
-------
1500-1
1000-
2! 500H
CD
20
PERCENT
I
40
I
60
I
80
URBANIZATION
Figure 4. The minimum volume of phosphorus limited lake
water that would be brought to an undesirable
phytoplankton density by urbanizing marshland
( ) or agricultural land ( ) according
to the data of this study. Numbers on lines
indicate selected marsh and agricultural
drainage areas in hectares.
06
-------
treating discharges from urban developments during the planning
stages of those developments.
ACKNOWLEDGEMENTS
This work has been supported by the U.S. Environmental
Protection Agency, Office of Water Planning and Standards, as
Grant No. R80504601 administered by the Corvallis Environmental
Research Laboratory as a part of the evaluation phase of the
Clean Lakes Program, and by the Michigan Agricultural Experiment
Station. John R. Craig, George W. Knoecklein, John M. McCabe,
Maureen, M. Wilson, G. Douglas Pullman and Mehdi Siami of the
Limnological Research Laboratory at Michigan State University
provided valuable assistance.
LITERATURE CITED
American Public Health Association. 1975. Standard Methods for
Examination of Water and Wastewater. 14th ed. American
Public Health Assoc., Washington, D.C. 1193 pp.
Barnes, B.S. 1940. Discussion of analysis of runoff charagter-
istics by O.H. Meyer. Amer. Sod, Civil Eng. Trans.™'"
105:1-106.
Colston, N.V. 1974. Characterization and treatment of urban
land runoff. EPA-i670/2-74-096. U.S. Environmental
Protection Agency, Cincinnati, Ohio. 170 pp.
Danigian, A.S., and Crawford, N.H. 1977. Simulation of nutrient
loading in surface runoff with the NPS model. EPA-600/
3-77-065. U.S. Environmental Protection Agency, Office
of Res. and Dev., Athens, Georgia. 110 pp.
Davis, S.N., and DeWiest, R.J. 1966. Hydrogeology.
and Sons, Inc., New York. 463 pp.
John Wiley
Dillon, P.J., and Rigler, F.H. 1974. The phosphorus-chlorophyll
relationship in lakes. Limnol. Oceanogr. 17:250-254.
Dillon, P.J., and Kirchner, W.B. 1975. The effects of geology
and land use on the export of phosphorus from water-
sheds. Water Res. 9:135-148.
87
-------
Edmondson, W.T. 1972. Nutrients and phytoplankton in Lake
Washington. In G.E. Likens, ed. Nutrients and Eutro-
phication: The Limiting-Nutrient Controversy. Special
Symposium, Amer. Soc. Limnol. Oceanogr. 1:172-193.
Nichols, K.H., and Dillon, P.J. 1978. An evaluation of phos-
phorus-chlorophyll-phytoplankton relationships for lakes,
int. Revue ges. Hydrobiol. 63:141-154.
Schindler, D.W. 1974. Eutrophication and recovery in experimen-
tal lakes: Implications for lake management. Science,
184:897-899.
U.S. Department of the Interior. 1967. Water Measurement
Manual. 2nd ed. U.S.D.I., Washington, D.C. 329 pp.
Vollenweider, R.A. 1968. Scientific Fundamentals of the Eutro-
phication of Lakes and Flowing Waters, with Particular
Reference to Nitrogen and Phosphorus as Factors in
Eutrophication. Rep. Organisation for Economic Coop-
eration and Development, DAS/CS1/68.27, Paris. 192 pp.;
Annex, 21 pp.; Bibliography, 61 pp.
Weibel, S.R. 1969. Urban drainage as a factor in eutrophication.
In^Eutrophication: Causes, Consequences, Correctives.
National Academy of Sciences, Washington, D.C., pp.
383-403.
Wetzel, R.G. 1975.
743 pp.
Limnology. W.B. Saunders Co., Philadelphia.
88
-------
Third Session
IMPACTS ON LAKES AND RIVERS
Moderator: Francis J. Condon
EPA, Washington, D.C.
89
-------
IMPACTS OF STORMWATER RUNOFF ON A
FLORIDA LAKE ECOSYSTEM:
EFFECTS ON WATER QUALITY AND BIOTA
by
Eldon C. Blancher II
Marine Environmental Sciences Consortium
P. 0. Box 386
Dauphin Island, Alabama 36528
ABSTRACT
A study of external nutrient loadings to the Lake Conway ecosystem, an
interconnected series of three lakes located in Orange County, Florida,
showed that both nitrogen (2.6 g-N/m2-yr) and phosphorus (0.22 g-P/m2-yr)
inputs were within the range of loadings that leads to mesotrophic conditions.
The major external sources of both elements were atmospheric inputs, storm-
water runoff and subsurface seepage,
Multivariate analysis of water quality data by discriminant analysis
showed differences among the three lakes of the Conway system. Seasonal
trends in several water quality indicators varied concomitantly with changes
in external nutrient loadings, especially with those from residential storm-
water runoff. Those lakes that had a proportionately larger share of phos-
phorus loadings from stormwater runoff showed a degradation in water quality.
Strong linear relationships were found between watershed area to lake area
ration and stormwater phosphorus loadings (r2 = 0.98); stormwater phosphorus
loadings and Secchi disk transparency (r2 - 0.88) and chlorophyll a^ and
total zooplankton numbers (r2 = 0.88). These findings indicate a direct
impact of stormwater runoff on water quality and subsequently on changes in
the lakes biota.
90
-------
INTRODUCTION
Numerous receiving water bodies have been adversely affected by the in-
troduction of wastewaters from various sources. While many investigations
have shown the direct impact of point source pollution, (for instance,
Edmondsons1 (1972) study as the recovery of Lake Washington after sewage
diversion) relatively few have shown the direct effect on non-point source
pollution, particularly stormwater, on water quality and biota. A large
number of excellent studies have recently been made on the quantity and
quality of stormwater runoff but in the majority of these studies the actual
impact of this pollution source can only be inferred. Both the paucity of
comparative data to systems receiving varying pollutional loadings and the
confounding influence of additional pollution sources have contributed large-
ly to this lack of conclusive studies.
Ideally, to test the hypothesis of the impact of stormwater runoff, one
should identify a group of receiving water bodies with similar physical,
chemical and biological features which vary in stormwater loadings and are
relatively free of other pollution inputs. Rarely do reasearchers have the
opportunity to work on such "ideal" systems. Fortunately, the Lake Conway
system discussed in this paper did meet the above criteria and served as an
excellent test case for demonstrating the impacts of stormwater runoff-to a
Florida Lake ecosystem.
DESCRIPTION OF THE-STUDY AREA ,
The Lake Conway system is comprised of a series.of fused dolines (Fig-
ure 1). Average depth for the system is 5.2 meters and surface areas range
from 0.28 km2 for Lake Gat!in to almost 3.0 km^ for the middle pool (Table 1).
These lakes are fairly representative of typical solution basin lake found
in Florida. The surrounding areas in the watershed (Table 1) can be classi-
fied as primarily residential, among which are interspersed areas classified
as agricultural, mainly citrus. However, few of these citrus areas have sig-
nificant agricultural activities and they could probably be better classified
as undeveloped land. A number of studies recently have been published on the
Lake Conway system in conjunction with the U. S. Army's Corps of Engineers
Project. These studies include a general description of the Lake Conway area
(Theriot 1977), and studies of its fishes (Guillory et al_. 1977), aquatic
macrophytes (Nail j?t. jj_. 1977, 1978), plankton and benthos (Fox e^t jil_. 1977;
Conley et^ aJL), water quality (Sawicki 1977), as well as system modeling
(Ewell and Fontaine 1977, Fontaine and Ewell 1978), nutrient budgets (Blan-
cher et a!. 1978; Blancher and Fellows 1979), and a study of the nitrogen
cycle~TSompongse 1978). The reader is referred to these articles for an
additional discussion of these topics.
91
-------
Figure 1. A map of the Lake Conway system indicating the location of
sampling stations. (0) indicates stations sampled from April
1976 through March 1978, (0) those sampled from April 1976
through March 1977, and (0) those sampled from April 1977
through March 1978.
92
-------
•
ea
:s
GO
ec£
02
LU
C3
^^*
I-H
^£
py;
Q
Q
UJ
i ,
•a:
1 — 1
o
o
GO
-
•=£
^
O
o
1 t 1
"=C
_J
LU
a:
i
i
Lu
0
GO
LU
ry1
ZD
j— :
LU
Lu
•
r—
CU
r—
f"»
ea
t—
s- -a
o cu
Q.
tO OJO
3 i— to
S- CU (O
4-> > cu
•i- cu s-
o -o «< to
re to
3 CU
1 —
-a
•— cu
ea .E (O
4-> to CU
Oe t
?« >—
H- CU ^-^
Q. CO
CU i-
-o cu
j \
c cu
(O £
CU
(O
CU cr
^" ^
•^C "^
o
vo oo "d- co i —
« • • • •
CM !— OO •* i—
CM 00 >=i- i— p—
co CM en co co
• • • • •
r-» oo o CTi "^t"
r^. vo uo vc i^
LO LO CM IO OO
r^ i — o^ «^j~ i —
• • • • •
r— CO CTl CTl CO
CM "5l- Ol >^- i—
r—< i — i — CD «^f-
CO to i — O OO
i — ^~
r- CO «* CM CO
o en LO r~^ i —
l£3 LO -=d" LO «^f
r^- en r-~ co oo
oo en CM <=i- CM
• • • • •
i— CM i— i—
t—
i— O
O O r— i—
O Q. O O
Q. O O
CU Q. Q. E
jT" p— •! —
4-> ~O 4^ 4-^ r~
~s ~o in to 4->
O •!- ea CU "3
GO SI LU S CD
CM
•
O"^
CNJ
r—
•
CO
vo
r_
•=d-
•
p^.
CM
0
oo
,_jl
oo
en
CM
tn
en
CO
•
r^
i~~
ea
] ^
o
f—
to
ea
cu
S-
fO
cu
ea
s-
o
to
E
0
to
cu
c
QL
CU
"°
s-
cu
o
-o
E
ea
to
-a
o
Q.
to
CU
-a
3
r—
O
E
fO
cu
fO
T3
CU
CM to
E S-
^ cu
] ^
E tO
•r- 3
"O CU
CU -E
s-
O <1-
Q- O
CU
S- CU
o
to E
fO (O
cu r—
&— fO
eC CQ
rO -Q
93
-------
METHODS AND MATERIALS
HYDROLOGY
Lake and watershed areas were calculated from United States Geological
Survey topographical maps with an electronic planimeter (Hewlett Packard
9810A-9864A). Extent of catchment area was determined from topographic fea-
tures and by consulting area tax assessment maps at the Orange County Public
Works Office for street drainage patterns. Land use and development patterns
for the watershed were estimated from a recent (1975) aerial photograph
supplied by county engineers (Orange County Public Works, Orlando, FL) and
updated by visual reconnaisance.
Monthly rainfall and evaporation data were obtained from U. S. Weather
Bureau climatological data reports (NOAA 1972-1978). Rainfall data reported
by the Orlando weather station at McCoy Air Base were used because of the
close proximity of the airport to the lake.
NUTRIENT INPUTS
Aolean sources were monitored utilizing a wet/dry precipitation collec-
tor placed on a boat house extending over the lake. Basically, the unit
consisted of two large buckets with a servo-operated lid and a precipitation
sensor.
Since there were no point sources entering the lake, our efforts were
directed to determining the non-point sources such as stormwater runoff from
the residential areas. The rational method was used, calibrated by monitor-
ing the quantity and quality of runoff using a v-notch weir in conjunction
with an automated flowmeter and sampler unit. The particular unit used had
a bubbler type flow meter which activated the sampler based on flow rate.
Thus flow weighted averages could be computed. Subsurfaces seepage measure-
ments were made by Fellows (1978) utilizing the methods of Lee (1977). Top
portions of 55 gallon drums were pushed into the sediments and allowed to
equilibrate for several weeks. Water was collected in collapsed latex bags
to determine seepage flow and then analyzed for nutrients.
WATER QUALITY AND BIOTA
Secchi disk transparency, dissolved oxygen, temperature, pH, chlorophyll
_§_, conductivity, turbidity, total hardness, and total alkalinity were rou-
tinely measured at all station in the Lake Conway system. Monthly primary
productivity measurements by the light-dark bottle method (APHA, 1975) were
made at 1 meter depth in the center of each pool of the Lake Conway system
from June 1976 through March 1978. Nutrient samples were obtained from each
of the pools from June 1976 through March 1978 in conjunction with biologic
sampling. Additional nutrient information was supplied by the Orange County
Pollution Control department. All analyses were perforned using standard
methods (APHA, 1975; U. S. EPA 1976) and average nutrient concentrations were
computed by weighting the samples on an areal basis.
Analysis of the trophic indicators of the Conway system were accom-
plished using multivariate techniques. The indicators used in this study
94
-------
were conductivity, total phosphorus, total nitrogen, Secchi disk transparen-
cy, functional chlorophyll a_, and primary productivity. Ordination of the
various pools according to trophic indicators was accomplished using stepwise
discriminate analysis (BMDP7M).
Monthly zooplankton samples were collected at littoral (< 3 m) and
limnetic (> 3 m) stations in each of the five pools of the Lake Conway system
following two sampling regimes. From April 1976 to March 1977, 16 littoral
and 5 limnetic stations were sampled, and from April 1977 to March 1978, 13
littoral and limnetic stations were utilized (Figure 1).
Zooplankton samples were collected at all stations by a vertical haul
with a U. S. Standard # 10 (153 y mesh) Wisconsin plankton net. Data ob-
tained by this method were expressed on both an area! and a volumetric basis.
RESULTS AND DISCUSSION
HYDROLOGIC AND NUTRIENT LOADINGS
Nutrient inputs to a lacustrine system are closely related to the hydro-
logic budget of the system. Thus, critical to the development of any real-
istic materials budget is the construction of an accurate water budget.
Every attempt was made to insure that all possible sources of nutrient inputs
were assessed. For the sake of brevity, only the most important sources are
considered here. Detailed discussion of all possible inputs to the system
have been previously presented (Blancher 1979).
Stormwater hydrographs for the Conway basin were calculated from data
collected on August 1976 and March, May, and June 1978. Measurements of
stormwater flows during 1977 were not possible due to the paucity of rain-
fall. Of the precipitation that fell during those storms, 3.5 percent on
an average entered the lake as stormwater runoff. Initiation of runoff and
time interval to peak flow were generally rapid as was return to base flow.
In situ seepage measurements provide direct measurements of seepage
flow. Prior to the development of this technique, seepage flow was estimated
by subtracting the measurable outputs from the inputs. The seepage meter
method used by Fellows (1978) provided values for seepage with an estimated
error margin of ± 20 percent.
Hydrologic inputs for the individual pools of the Lake Conway system
are presented in Table 2. Precipitation is the dominant hydrologic input
of all the pools. Seepage for the various pools varies from 15 to 23 percent
whereas inputs from stormwater varies from 3 percent in the south pool to 28
percent in Lake Gatlin.
In order to verify the accuracy of the parameters used in the hydrologic
budget, the calculated flows were used in a dynamic model to predict change
in storage within the total lake system. An information flow diagram for
the model is presented in Figure 2. The model was simulated for the period
from December 1972 through March 1978 on a monthly basis.
The results of the model are compared with the actual changes in lake
height in Figure 3.
Differences between the simulated and observed changes in lake height
are within five percent on a volume basis. For a detailed discussion of the
development of the model see Blancher (1979).
Aolean nutrient loadings based on the rainfall and dry fallout samples
95
_
-------
r
H-
eC
- —J—
• — L—
fe*-
LU z:
t_.._n . .
CO
>- 3
CO O
>- u.
3 HH
C^ i
C—5 "^
t—
LU O
rr" |—
1 —
•z. o
i— i
CO 2=
_J LU
0 0
o c£
Q- LU
O_
_J
Q 1—
l — i
> LU
l— l |—
f"~i c£
"Z. C_5
Q
•XlS
a: LU
o co
U_ LU
3=
££
feg
r\ rV
^y^ *^
1— 1 Q.
o 2:
1— 1 1 — 1
f"~i (/J
l r>^
O LU •
C£. CQ _J
Q SI O
>- =3 O
:n 2: DU
CM*
CO
f^
o
•r—
res
•t—
Q.
•r-
O
O)
s-
D-
"o
0
D-
, Q^
i — CO
CM «*
^^ ^-x
C7^ UO
i — ^-*
^^-^ *^~*
LO cri
CTi VO
co vo
* *
»*— -» *^^«.
«^j- CO
^ — " **—*
*JD ^O
r^« co
0 r—
* •
•••"•*-•. *~^.
CO CM
r-^ co
•>!- CT>
IO LO
• •
i— CO
O)
^ I
+J "O
3 -O
o •«-
CO S
co •* o
vo co r*-.
CM CM O
. ^ . — s , — s
01 LO co
i— i— CM
*. * ^..^ **^r
oo co si-
co co vo
«^J- ^J- r —
• • •
i^~* •*—** " — N
•* CM CO
CM CM CM
» — • — ' ^- •
vo cyi ^^*
c\j CM cr»
to ^o i —
• • •
*••— *» ^"^«* ** — •*
CO CO CTi
Lf> (^ «s3*
CM CO <*
• ur> r^* co
• • •
r~~ ^^
£
•!-> -t-> r—
t/> t/J +J
(C
-------
Q(t)
Q (t)
dt
Q0-Q,±Q(
dV
*Q
°o(t) **
Qlt\ ^
0ltJ ^
Q,(t) »-
dVc,
— r»/-' -Or> -lff<\/
df - Qoco QICI kciv
^ n (t\
Figure 2. An information flow diagram for the dynamic
hydrologic-materials systems model.
97
-------
87-i
h-
LJ
UJ
LL
co
UJ
O
00
h-
X
UJ
X
UJ
86-
85-
84-
co
UJ
83
0
10
20
30
MONTH
40
50
60
Figure 3. Comparison of the actual variations in lake height of the
Lake Conway System (A) and the simulated height from the
hydrologic model (o) for the period December 1972 (month
= 0) through March 1978 (month = 63).
90
-------
collected, provided annual goading values of 0.048 gP/m^-yr and 0.36
for rainfall and 0.080 gP/m^-yr and 0.57 gN/m^-yr for dry fallout. The
resulting bulk loadings (0.125 gP/m^-yr and 0.93 gN/m2-yr) compare well with
similar loadings of 0.105 gP/m^-yr and 1.0 gN/m^-yr for Gainesville Florida
(Hendry and Edgerton 1978, unpublished data, University of Florida, Dept. of
Environmental Engineering).
Analysis of nitrogen and phosphorus concentrations in stormwater runoff
from a small catchment in the Lake Conway watershed yielded average concen-
trations of 3.25 mg/L and 0.39 mg/L for nitrogen and phosphorus, respectively
(Table 3). These compare well to published values which range from 1.93 to
4.45 gm/L for N and from 0.19 to 0.98 mg/L for P (Wei be! 1969; Kluesener and
Lee 1974; Mattraw and Sherwood 1977).
Nutrient inputs from subsurface seepage to Lake Conway were determined
by Fellows (1978). Annual loadings per meter of Lake shoreline from this
source, calculated as an average of three sites were found to be 0.66 gN/m-yr
and 0.024 gP/m-yr. t
Additional external sources of nutrients were identified but accounted
on the average for less than 5 percent of the inputs to the lakes and hence
were considered insignificant.
Areal nitrogen and phosphorus loading rates from the three major ex-
ternal sources are presented in Table 4. Of these input sources, the storm-
water loading rate was most variable between the pools accounting for ap-
proximately 10 percent of the N and P inputs in the South pool to about 50
percent in Lake Gatlin.
TROPHIC STATE/WATER QUALITY VARIATIONS
Six parameters were chosen as trophic/water quality indicators of the
Conway system representing the physical (Secchi disk transparency and con-
ductivity), chemical (total nitrogen and phosphorus), and biolgoical (func-
tional chlorophyll _a and gross primary productivity) attributes of the lakes.
Comparisons between those parameters were used as a basis of determining
difference in water quality.
Limnologists have determined general guidelines for determining trophic
state based on total phosphorus, chlorophyll ja, and Secchi disk transparency.
Gakstatter _et ajL (1975) reviewed the literature and established criteria
used for classifying the lakes sampled for the National Eutrophication Survey
(NES). Comparing their values with annual average values of the five pools
of the Conway system (Table 5) we see that the south and middle pools border
on oligotrophy, whereas the east and west pools are decidedly mesotrophic.
Using the NES criteria, Lake Gatlin would be considered eutrophic. While
this comparison illustrates the overall (average) trophic states of these
lakes, better resolution between the study lakes would be more useful in ex-
amining the mechanisms and causative factors involved in the eutrophication
process. This resolution is necessary for any comparisons of the importance
of various loading sources to overall lake response.
The multivariate method of discriminant analysis provides linear func-
tions of the variables from a multidimensional data set. These functions
then allow the classification of the multidimensional data vector into one
of several multivariate normal populations. A modified procedure known as
stepwise discriminant analysis (BMDP7M, Dixon and Brown 1977, using default
options) was used to identify the variables that would produce a "best-fit"
99
-------
Table 3. NITROGEN AND PHOSPHORUS CONCENTRATIONS IN STORMWATER RUNOFF
ENTERING LAKE CONWAY. CONCENTRATIONS ARE FLOW WEIGHTED
AVERAGES EXPRESSED AS g/m3.
21 August 1976
23 March 1978
4 May 1978
16 June 1978
21 June 1978
Number of
Samples
9
27
28
12
8
Total
Nitrogen
2.3
4.7
4.6
4.0
5.1
Total
Phosphorus
0.33
0.45
0.47
0.42
0.63
Total Flow
On3)
986.3
45.5
665.8
45.8
4.6
Mean Values
(flow weighted)
3.25
0.39
100
-------
oo
UJ
h-
•=£.
*
OO
UJ
0
C£
"~>
o
oo
_i
~^*
rv*
UJ
1—
X
1 1 1
1 ' *
rv*
O
•~3
S
CD
i ,
OO
UJ
1—
^£
Qr^
CD
Q
•a:
o
oo
ro
o
o
31
O-
oo
o
Q. •
Q >>
•z. \
^^ cr
^z
UJ O5
CD
o 2:
rv i i
HH OO
2: uj
h-
(0
^_
o
•»-»
oo 2:
f^
E
(O
cu
r— •
o
ei
^
CTl
00
^
^o
r-«
•
O
CO
<£>
•
^~
,_).
CM
O
•
O
CM
<~O
O
LO
O
O
CO
' —
CD
LO
CM
CD
CO
cn
*
o
^C7
^_>
3
0
oo
CO
I""""*
LO
LO
r—
•
O
i—
LO
*
1
CTt
O
•
0
CO
CD
CM
0
CD
0
CD
LO
CM
*
O
co
en
CD
cu
^
•o
^
vo
^j~
cr>
co
,
(T>
CM
•
O
O
•
CM
CM
CO
O
•
o
CO
CO
o
^J-
co
r—
•
O
r—
' —
r_!
LO
CM
CD
CO
CTl
0
4.)
t/>
CO
UJ
*sd~
^J*
CO
co
UD
CM
•
O
CM
LO
, •
CM
LO
CM
0
O
CO
LO
o
vo
f-^
•
o
LO
en
CD
LO
CM
*
O
co
CT>
•
O
.4.3
C/J
CU
s
r-.
LO
r^.
«^-
LO
o
•
o
*^"
1—
•
•=!-
cn
o
•
O
f"**
CM
,
CM
CO
CM
•
0
^J-
en
,-J
LO
CM
CD
CO
cn
CD
c
•r-
r—
4_>
10
CD
101
-------
UJ
UJ I—
5 0
l —
CO CO
CD CD
D-
Q£
UJ CO
tg
1—
SI LU
CO h-
o s:
1-1 O
:n ce:
Q- LU
o
en. co
o
CD h_
1—1 CJ>
I— i Q
§ i— i
LU
1 — LU
UJh-
\—
DC CO
0
LU 0
i—i
co :n
LU D-
2: o
i— i a;
—I r—
LU
i — i cC
• — i fV
CD LU
>- UJ
UJ CO
a: u.
ZD O
co
CO
o -
__1 CO
-
0 _J •<
h- 2:
•=C I— 0
^ LU O
*
in
QJ
r—
ro
1—
c
CD -r-
CO 4->
—I CO
CD
4-5 ^~
CO O
CD O
ns D-
4->i—
tO O
CO O
LU D_
dJ
r— i —
•a o
-a o
•r- a.
s:
c"
40 >~
CO Q-
O
•r—
c~
Q-
O
S-
4-5
3
LU
O
•r—
JC
Q.
0
s_
4-5
O
O)
s:
o
[£
CL
O
S-
o
CO
o
, —
CM
O
•
o
o
CM
CD
O
^~
CM
O
•
O
CO
1—
CD
•
CD
to
0
o
CM
O
•
0
A
CM
0
•
O
I
(—
o
0
0
•
o
V
^
CD
E
^— ^
to
3
o
c*"
Q-
to
0
JC
a.
i —
CO
o
1—
o^
^"
•
O"i
p*^»
^*
^
^
•
^t"
r*^^
*
CO
CO
CM
O
t — '
•
o
1 —
A
0
r—
|
•=*-
o
•
•=3-
V
,— ,
1
^^
CD
3-
C0|
r—
^~
~*">
,_E^
Q.
O
t-
0
r—
-C
O
to
co
•
1 —
to
to
CM
0
r--
*
CM
O
CO
•
co
to
CO
o
•
CM
V
r-.
•
CO
1
o
CM
f^.
•
CO
A
* — *
E
-C
4-5
a.
CD
*^3
•i —
-f
o
o
O)
CO
102
-------
model to discriminate between the populations. Each of the pools in the
Conway system was treated as a "population", and the observed variables in-
cluded the monthly mean values of the six trophic state indicators. The
procedure was performed using all six indicators, as well as several subsets
of the variables.
The results of the analyses (Table 6) show that conductivity and Secchi
disk transparency were the only variables that explained most of the varia-
tion in the data. The first analysis (Test 1, Table 7) proved good for
separating out the seasonal data for Lake Gatlin, but showed poorer discri-
minating ability when trying to discriminate samples from the other pools,
especially between the west and east pools and the south and middle pools
(Table 7). This problem was due to the high degree of similarity of the
data of the contiguous pools. A considerable improvement of the discriminant
functions was obtained by combining observations from the pools of the same
lakes (i.e., the east and west pools and the middle and south pools). The
first two canonical variables from this final analysis (Test 4) can be used
to cluster the data into three groups with some reliability as illustrated
in Figure 4.
While condictivity was the best variable in terms of discriminating the
lake pools, this variable reflected changes in the hydro!ogic regimes more
than actual changes in lake trophic state. The discriminant analysis showed
that Secchi disk transparency was almost as good as a discriminator by it-
self and also reflected changes in the trophic conditions in the system with
some regularity. For this reason it was felt that the transparency alone
could be used to indicate changes in the trophic state of the Lake Conway
system.
ZOOPLANKTON
Seven species of copepods, 14 species of cladocerans, and 24 species of
rotifer were identified in samples collected from the Lake Conway system.
At any particular time, two to four species of copepods and cladocerans,
and one to three species of rotifers were encountered in the lakes. Shannon-
Weaver diversity indices computed for the individual samples were generally
low, ranging from 0.80 to 2.46. However, low zooplankton diversity is
characteristic of Florida lakes (Nordlie 1976) and North American lakes in
general (Pennak 1962). All of the zooplankton species encountered during
this study have been collected form other Florida lakes (Nordlie 1976;
Confer 1971; Mas!in 1969; Cowell et a]_. 1975).
The 24 month averages of area! abundances of the major taxa reveal some
interesting trends (Table 8). Copepods comprised over 43 and 42 percent of
total zooplankton in the south and middle pools, respectively, with clado-
cerans and rotifers of lower importance. Diaptomus floridanus was the
dominant organism in these two pools. Rotifers dominated the west pool
while in the east pool all three of the major taxa were nearly codominant.
Forty-three percent of the total zooplankton in Lake Gatlin were cladocerans,
primarily Bosmina longirostris. Average annual abundance of total zooplank-
ton ranged between 201,351 and 378,906 individuals per square meter of Take
surface area (Table 8).
The results of the zooplankton study indicate several important trends
that may have significance in the development of zooplankton communities in
Florida lake systems. Total zooplankton abundance increased moving north
103
-------
co
O
o
I—I
Q
o_
o
G!.
O
o
UJ
DC
r-
O
CO
t—i
CO
o
CO
co
I—I
Q.
UJ
co
u_
o
co
CO
UJ
cc.
CU
r—
rO
CJ
•^
•— t
LO
CT>
CD
1 |
1 I i
.
u.
* •
LO CD
I— 1—
•i— CO
^ *r~
•r- -a
•p
O "r-
3 .E
T3 O
E O
O CU
O CO
LO
to
s-
o
•i _TC"*
re| ex
00
\S | Q
co r— .E
•r- >> CX
•a j=
CXr—
•r- O
o o o
O r— -P
cu ^:
co o »
E
>> >, en >,
-P O O -P
•r- E J- -i-
> CU -P >
•r- S- T- -r-
•P re E -P
O CX O
3 to r- 3
"T3 ^" fO "XU
E re -P 0
0 S- 0 S-
CJ -P -P Q.
1 —
O 0
O CD
«* W-
-I-
•r- T3
•P
O T-
3 .C
-a o
E O
0 CU
O CO
LO
t^
re|
sx ^~
to p—
•r- >)
CX
•i- O
JC S-
0 0
U r-
CU -E tO
CO O 3
S-
w rt CD
•POO.
•r- E tO
> CU 0
•r- S- .E
•4-5 (O f^>
u ex
3 tO i —
•0 E re
E rO -P
0 i- 0
0 -P -P
(M
o
o
cjjl
W3
CM
•
o
I—
"CO
•I—
-a
•r-
c^
o
0
CU
co
LO
«t
E
CU
en
o
« S-
^> -p
O -i-
E E
CU
S- i—
rO re
f"i -p
tO O
E -P
rO
i— 1
•P rel
^r- ^
tO i r-
cx-p
•r- O O
J= 5- 3
O O T3
0 r— O
CU J= S-
co o ex
co
O 0
o o
•=d- -=d-
1'^ »^J'
r— CO
• •
CM i—
CO CM
•r- tO
^ •!—
•P- ~o
O T-
3 J=
-o o
E 0
O CU
o co
CO
CO
3
s-
o
rOlQ.
to
-i<£ i— O
tO I— -E
•i- >, Q.
•o ^
Q-i—
•r- O rO
.E S- -P
o o o
O r- -P
CU .E
CO O «
E
« •* CU
>, >, CD >>
+? C? 0 4?
•r- E S-T-
> cu -P >
•P ro E -P
o a. o
3 tO i— 3
TD E rO "O
E ro -P O
O i- O S-
O -P -P CX
^
.
to
CU
r—
ro
'si
ro
o
o
II
^
CU
g-
ro
to
M-
o
•
o
E
re
o
I—
II
z:
to
CX
3
0
S-
en
q_
O
•
O
z:
u
CD
^ *
^»
1
' CD
•
z:
u
CM
• n
^~
1
C3
II
i-H
*
104
-------
LiJ
1C
ii i
* ill—
^H-
"^ i i
^^ *-)-•
UJ O i— i
Q re oo
oo oo
oo D;
LU 2: Q;
O 0
1— I-H 0
s: i—
> _
•p -^
•r- tO
>•!-
•i- -a
•p
CO o *i- -P
i — 3 -E E
J3 -o o re
re c o -p
•r- O CD V)
S- CJ 00 E
re o
=• 0
Q.
3
o
s_
CD
CO
0
to
CO
I/)
re
CJ
• J
0
o
•z.
-p -p
E O
CD CO
0 S-
i- i-
CO 0
a. cj
a.
3
o
CD
a.
3
O
CD
S_
CD
Q.
to
CO
to
re
o
t.
*t^
o
s-
co
JD.
E
2:
-P 4->
E O
CO CD
O S-
f~ S_
CO O
Q- O
Q_
3
e
CD
d
r— o co CD r^
•P r— CM
re
CD
-p -p
to to co LO co to
re CD r— CM •*
1 1 1 3
1 CD
_E i—
-p -a to i— o t-~
3 -a CM i— co
O 'I-
oo s:
i — CO «d- to
* • • *
CD to Id CO
LO LO CO tO
Itl
IdJ
-o
-a -P
•r- tO
s: co
1 3 E
_E 1 •!- 1—
•r™* •!— * r1" " flj
3 10 -P -P
o re re o
tO LU CD 1—
E
i — CD O CO CO CD LO
-P i— OJ
re
CD
.J_>
tO LO tO CO CD CM O
CD CO
3
-p
to LO OJ •=*• to i— co
rrt r-
iw r '
LU
CD
TD i— !- i— CD IO
O i— •— CM
oo
O CM CM CT> «3- I~^
O CO CO O tO CM
LO i — i — ^ CO "vf
CO E
.E i— • »r- r— -
•P -a -P -P i— re
3 -o to to -p -p
o •!- re CD re o
oo s: LU 3 CD l—
105
-------
r
•1
O
q
to*
1
o
in
*
i
o
o
d
_Q
i
i i
0 0
10 q
i i
o
^
CO
O)
o
1/5
- 1
^ p.
^0
-a
S-
\
\
\
/
/
4->
CO
OJ
-^.2
to
CO
UJ
c
•1 —
+->
S^ fO
^«. CD
O)
.^
(0
O
— •
CM
LJ
•
^•J
CD
_
a:
<
-9 >
0
_i
<
0
z
o
z
<
o
0
—, •
CM
1
w/
to
,_?
fO
fO
-p
fO
c
• I—
E
•i—
5-
O
co
-a
O)
5
E
O •
i~ CO
M- 5-
0
CD4->
CZ fQ
•r- O
+J •!-
i— T3
3 C
CO -r-
O)
±- cu
CO fO
ai -M
i — co
J3
03 O
•1 — •! —
<- JC
fO d.
> 0
i.
(J X
C CO
O
C O)
O -I-5
O)
S-
3iavmvA HVOINONVO
106
-------
o
I—
2:
1
Q-
o
CD
_1
^
O
I—
0 •
-
D-
^f_
^^
^C ' '
•
co
O)
i —
JD
fO
1—
c
•r~
r—
^->
i03
CD
H^
* OO
LO •* . ,
«=f
LO
CTt
I~~
LO r-
CO CM
o
<£
CO CO
LO CM
o
co
1 . , • .
* *
E' . E
0 0
4-> 4->
E ^^^^ E
ctf CM CO • — ^
r- E r— _J
0 O O O
O2: O2:
r— • i —
fO ' rO
O O
H- h-
CO
E
CO
•r- O
•i— IM
i— O
CX4->
3 O
CO i-
CO CO
CD
"O "O
3 3
O O
E E
i — i i — t
* *
*
107
-------
through the pools, associated with increasing trophic state. This trend
also has been observed for other lake systems (Patalas and Salki 1973). The
increase in the Lake Conway system is attributable to an increase in the
abundance of small cladocerans and rotifers, taxa which are indicative of
eutrophic conditions.
INTERRELATIONSHIPS
Nitrogen and phosphorus stormwater loadings to the various pools varied
concommitantly with watershed lake area ratios (Table 9). The South and
middle pools of Lake Conway which had a watershed Lake area ratios of about
1.0 received a minor portion (< 10%) of its nutrient loadings from stormwater.
Conversely Lake Gat!in with a Aw/Am ratio of 11.18 received about half of its
nutrient load from this source.
Table 10 compares nitrogen and phosphorus stormwater loadings to the
water quality and biological indicators from the Lake Conway system. Signi-
ficant correlations between the various parameters are presented in Table
11. Strong relationships were found between watershed - lake area ratios and
stormwater nutrient loadings (r2 = 0.98); between stormwater nutrient load-
ings and secchi disk visibility (r2 = 0.98); between secchi depth and
chlorophyll a_ (r2 = 0.93.) and total zooplankton (r2 = 0.98); and between
chlorophyll a. and zooplankton (r2 = 0.88). Significant relationships were
indicated between zooplankton and watershed - lake area ratios and zooplank-
ton and nitrogen stormwater loadings but these were considered to be spurious.
These results indicate that those pools of the Lake Conway system which
received a larger share of their loadings from stormwater sources exhibited
a degradation in water quality as indicated by changes in secchi disk visi-
bilities and chlorophyll
-------
s:
UJ
1—
oo
00
oo
rv 1 1 1
O OO
:n uj
CL. 3=
oo 1—
o 2:
zn uj
r> fV
^^
Q Q.
«c 2:
i — i
2:
uj oo
CD Ct:
CD UJ
C£ CO
i — s:
1 — 1 — i
2: 2:
•
CTi
^
i
CM CM
E E
cn
i
^^
i— i **•»
2
**~^*
C\J
r— ^ — *» r~" * » CO **"** P" " * — s CO ^^^»
OO O^O i— CO r— i— OJO
*i> •• •• •• • •
o 0*1 o f*^ o ^o o *^" o r-*.
"-' ^ • S S ^
OO *^""- t~~* ' — ^ r— *"^ *^5 ^ * «^~ - — ^
i— LD r— co i— oo cnr— cncn
• • •• •• * r ••
CD f — ' O ^ i — CTi O CO i — *£)
— * — cf) m "*i-
-. — — - <~-
CO UV i— OO CO
CM o oo r~* i—
• • • • •
i — i — r-- 10 i —
f—
LO LO CM \O OO
r^^ r— CT> «v!" r—
• • • • •
i— oo en LO oo
I*--* crv r^^ co co
OO LO CM "=d- CM
r— CM i-^ i— O
r— E
I— 0 •!-
O O i— r— r—
O D- O O +J
Cu O O to
O) D- Q- CD
C"" f—
-P "O *P 4-^ CU
3 T3 IO CO -i^
O -r- n3 CU (0
oo S uj 3 _J
•
0
'J3
fO
s-
03
CU
£^
ft3
a;
nu
1
O
to
CU
re
-a
CU
c/>
S-
QJ
•t^
to
3
II
_J
^^
^s»^
S.
<— i
en
c
•r-
-a
ro
o
I—
c:
CU
§>
S-
-p
•r—
C
s-
cu
•f™*
03
2
^
^.
o
•p
oo
II
C
1
CM
cn
•r—
-a
to
o
i—
t/>
=3
i-
0
•§_
(/)
O
c~
Q.
S-
cu
*4~^
03
3£
£
^.
o
.{_3
oo
II
Q.
l
OO
109
-------
co
ID
CD
i — i
r*s
^^
^>
LU
I—
O
U-
co
a:
CD
I-
O
Q
i — i
Si-
t—i
CD
O
CD
I — i
CQ
Q
z.
CO
ex:
0
o
i— i
Q
Z.
t—t
1—
i— <
i
et
=>
CD"
I— I—
=a: co
3: >-
CO
CO >-
CD «=C
Z 3
0 CD
«=C 0
O
_1 LU
^C
UJ
I 1 .
p~" UL.
•< O
s: co
a: _i
O 0
1— 0
CO D-
•
CD
r—
0)
p—
J71
fO
H-
0
r-~ ^o
fo ^: ^-~-
•4-> E r—
o to -^.
f- •— =a=
o
o
*l
i —
*>) r— '
c~ ^^^
Q. OJ
O E
5^ s ^
O
S
C_3
_£-
4->
Q.
O)
"O ' — ^
!S '
o
o
o>
co
t
PI "*v^
H- CD
E
•z. ^-~.
V- CD
f ,
£^
>^
Q- 1
_J CM
E
*^,
CD
S-
>i
—1 CM
^
^*^
CD
^---
(U
(tj
^1
co
CM
CO
•
CM
LO
co
10
o
CD
CO
f^
•
o
LO
r—
O
•
o
CO
,—
•
o
,_
0
o
Q-
c~
^-}
3
0
CO
p^x
CM
^^.
•
OO
CM
CO
co
co
0
CD
^
P-^.
•
O
CM
^—
o
•
o
0
r—
*
CD
O
0
D-
O)
"U
-a
s
CO
«^"
«;J.
f—~
•
«^j-
r—
P~-
CM
(
CM
O
*
O
LO
•
o
^t"
CO
1—
•
o
r—
r—
•
p— .
I—
0
0
D_
•P
(/>
fO
LU
CM
^"
1^
«^J-
«
(/)
CD
rs
^*
UD
CTt
*^J"
•
O^
Lf)
CO
,_!
i —
CM
O
O
,_,.
co
CD
CM
CO
CM
*
O
•vf
C31
•
r—
-r_
r^
-P
fO
CD
CD
^s
to
—1
110
-------
CD
!^
1 — 1
Q
0
00
O
fc
LU
DC •
1— 00
CD O
i— < 5
1— 0
eC i — i
_J -Z.
LU H- 1
C£
o <:
CJ CJ
s: CD
0 O
Lu 0
i — i
O CD
LU
> Q
S §
111
I.I..J
Q 00
o;
CM (—
0
•Z. Q
0 S
i — i i — r
1—
l—) i — i
H 5:
LU ^3
I— o-
LU
Q a:
LU
1 1 1 —
I.L. | — •
og
oo
t f.
\ — **
•z. oo
LU Q£
h-H LU
CJ \—
1— l UJ
Lu "Si
1 1 1 Q£
0
ii •
O
^
•|c
oo
1 1 CTl
1 1
CD
^
*^J"
i i cn
CD
If
D- Z. *j -j-^
j£-* , J _\^
Q. (O E
O 03 +-> i —
0 O
CM • •
$- O O
II II , II
1 CM CM
1 * -X
111
-------
ACKNOWLEDGEMENTS
I am expecially indebted to the U. S. Army Corps of Engineers Waterways
Experiment station, Aquatic Plant Control Research Program, Vicksburg,
Mississippi for its support and cooperation throughout this study. Special
thanks are extended to Mr. Lewis Decell, Mr. William Rushing and Mr. Russel
Theriot for their sustained interest in the project.
Acknowledgement is given to the Department of Environmental Engineering
Sciences, Univeristy of Florida, and especially to Dr. Jackson L. Fox and
Dr. Patrick L, Brezonik whose advise and guidance made this study possible.
The assistance of Charles R. Fellows both in the field and laboratory was
greatly appreciated.
The Assistance of the Marine Environmental Sciences Consortium, Dauphin
Island, Alabama, and in particular the cooperation of Dr. George Crozier,
director, Linda Lutz, illustrator and Pam Barbour and Lynn Bryant, typist,
is gratefully acknowledged.
LITERATURE CITED
APHA. 1975. Standard methods for the analysis of water and wastewater, 14th
ed. New York.
Blancher, E. C. 1979. Lake Conway, Florida, Ecosystem: Nutrient Dynamics,
Trophic state. Zooplankton relationships. Ph.D. dissertation Univ.
of Florida, Gainesville. 146 p.
eco-
, C. R. Fellows, D. Sompongse, and J. L. Fox. 1978. Nitrogen
and phosphorus loading characteristics of the Lake Conway, Florida,
system. Preliminary Rep. Dep. Env. Eng. Sci. Univ. of Florida, ecosys-
tem. Preliminary Rep. Dep. Env. Eng. Sci. Univ. of Florida, Gainesville
Fla. 47 p.
_^__ and C. R. Fellows. 1979. Nitrogen and phosphorus loadings of
the Lake Conway ecosystem: Final report. Final report to U. S. Army
Engineer Waterways Experimental Station. Department of Environmental
Engineering. University of Florida, Gainesville. 62 p.
112
-------
Confer, J. L.
ural prey
1971. Intrazooplankton predation by Mesocyclops edax at nat-
densities. Limnol. Oceanogr. 16: 663.
Conley, R. A., E. C. Blancher, F. S. Kooijman, and C. Feerich. 1978. The
plankton, periphyton and benthic invertebrates of Lake Conway, Florida.
Dept. of Environmental Engineering. University of Florida, Gainesville.
Cowell, B. C., C. W. Dye, and R. C. Adams. 1975. A synoptic study of the
limnology of Lake Thonotosassa, Florida. Part I. Effects of primary
treated sewage and citrus wastes. Hydrobiologia 46: 301.
Dixon, W. J. and J. Brown. 1977. BMDP-77, Biomedical computer programs, P-
series. University of California Press. Berkeley. 880 p.
Edmondson, W. T. 1972. Nutrients and phytoplankton in Lake Washington. _In_
G. E. Likens, ed. Nutrient and Eutrophication: the limiting nutrient
controversy. Special Symposium, Amer. Soc. Limnol. Oceanogr. 1: 172-
193.
Ewel, K. C. and T. D. Fontaine, III. 1977. Proposed relationships between
white amur and aquatic ecosystem at Lake Conway, FJorida. U. S. Army
Engineer Waterways Experiment Station. M. P. A-77-3: 159-176.
Fellows, C. R. 1978. The significance of seepage in the water and nutrient
budgets of selected Florida lakes. M. S. Thesis. College of Engineering
University of Florida, Gainesville.
Fontaine, T. D. and K. C. Ewell. 1978. Baseline studies for evaluating the
response of an ecosystem to the introduction of white amur. Second
Annual Report to the U. S. Army Waterways Experiment Station. Center
for Wetlands. University of Florida, Gainesville (mimeo).
Fox, J. L., E. C. Blancher, F. M. Kooijman, R. A. Conley, and C. P. Feerick.
1977. Biological baseline studies of the Lake Conway, Florida, system.
U. S. Army Engineer Waterways Experiment Station. M. P. A-77-3: 123-145.
Gakstatter, J. M., M. 0. Allum, and J. M. Omernik. 1975. Lake eutrophica-
tion: Results from the National Eutrophication Survey. 'Corvallis, Ore.
32 p.
Guillory, V., R. Land, and R. Gasaway. 1977. Baseline data report — Lake
Conway grass carp project. Fishes. U. S. Army Engineer Waterways Ex-
periment Station. M. P. A-77-3:36-112.
Kluesener, J. W. and G. F. Lee. 1974.
storm sewer in Madison, Wisconsin.
936.
Nutrient loading from a separate
J. Water Poll. Contr. Fed. 46: 921
Lee, D. R. 1977. A device for measuring seepage flux in lakes and estuaries.
Limnol. Oceanogr. 22: 140-147.
113
-------
Maslitij P. E. 1969. Population dynamics and productivity of zooplankton in
two sandhills lakes. Ph.D. Dissertation. University of Florida,
Gainesville.
Mattraw, H. C. and C. B. Sherwood. 1977. Quality of storm-water runoff from
a residential area, Broward County, Florida. Jour. Research U. S. Geol.
Survey 5_: 823-834.
Nail, L. E., M. J. Mahler, and J. Schardt. 1977. Aquatic macrophyte samp-
ling in Lake Conway. U. S. Army Engineer Waterways Experiment Station.
M. P. A-77-3:113-122.
, J. D. Schardt, and A. P. Burkhalter. 1978. The effect of the
white amur on the aquatic vegetation of Lake Conway, Florida. First
Annual Report, Baseline Data. Florida Department of Natural Resources.
Tallahassee, Florida. 161 p.
NOAA. 1972-1978. U. S. Climatological reports. Florida section. National
Oceanographic and Atmospheric Agency, Washington, D. C.
Nordlie, F. G. 1976. Plankton communities of three central Florida lakes.
Hydrobiologia 48: 65.
and A. Salki. 1973. Crustacean plankton and the eutrophication
of lakes in the Okanagan Valley, British Columbia. J. Fish. Res. Bd.
Can. 30: 519.
Pennak, R. N. 1962. Species composition of limnetic zooplankton communities,
Limnol. Oceanogr. 2^:222.
Sawicki, A. T. 1977. Background water quality analysis of Lake Conway.
U. S. Army Engineer Waterways Experiment Station. M. P. A-77-3: 146-
158.
Sompongse, D. 1978. A study of the nitrogen cycle in the Lake Conway,
Florida, ecosystem. M. S. Thesis. College of Engineering, University
of Florida, Gainesville.
1976. Methods for chemical analysis of water and wastes.
Environmental monitoring and support laboratory, Cincinnati, Ohio. 298
Weibel, S. R. 1969. Urban drainage as a factor in eutrophication. In:
Eutrophication: Causes, consequences and correctives. Nat!. Acad. of
Sci. Washington, D. C. 383 p.
114
-------
THE DISTRIBUTION OF SEDIMENTS AND PARTICULATE
CONTAMINANTS FROM COMBINED SEWER AND STORM DRAIN
OVERFLOWS IN SEATTLE'S NEARSHORE WATERS
R.D. Tomlinson, B,.N. Bebee, and R.G. Swartz
Municipality of Metropolitan Seattle, Seattle, Washington
S. Lazoff and.D.E. Spyridakis .31 r
University of Washington, Seattle, Washington
ABSTRACT
The distributions of particulates and the associated
contaminants emanating from combined sewer outfalls and storm
drains in Lake Washington were examined. In a preliminary
study sediment samples were collected by SCUBA divers from the
area of greatest apparent contamination at each of 20 outfalls.
The sample content of organics, heavy metals, oils and greases,
and pesticides was used to select stations for more intensive
study.. At each of those sites quality and quantity analyses of
the overflows provided loading estimates for the wastewater
particulates entering the nearshore waters. Light transmission
measurements of storm-induced wastewater plumes helped to define
the nature of plume dispersion and its area of significant
influence. Organic carbon and heavy metals analyses of settling
particulates and surface sediments near the outfalls supplied
further detail and confirmation of the fate of the wastewater
particulates.
ACKNOWLEDGEMENTS
This work was done by personnel of the Municipality of'
Metropolitan Seattle and the University of Washington
Department of Civil Engineering under a grant from the U.S.
Environmental Protection Agency. The authors gratefully
acknowledge this support. .
INTRODUCTION
There are about 11Q overflow locations in the Seattle area,
of which 30 are the Municipality of Metropolitan Seattle's
(Metro' s) and the remainder are the responsibility of the City
of Seattle. A brief, intensive rainstorm can gorge city
collector sewers, causing basement backups and discharges to
Lake Washington. In contrast, a milder storm of longer duration
will tend to flow more smoothly to the larger trunk sewers
where a buildup of flow beyond treatment and interceptor trunk
capacity will lead to programmed overflows to the lower Duwamish
River, Elliott Bay and Lake Union. These longer storms, typical
of the winter season, iticiy result in less than 50% of sewer
115
-------
wastewater actually reaching the West Point Treatment Plant
during a given storm period. Operation of Metro's Computer
Augmented Treatment and Disposal (CATAD) System, which provides
in-line storage in some of the larger sewers, permits some
flexibility as to discharge location, with highest priority
given to protection of freshwater bodies (Lake Washington? Lake
Union) \vhere possible. Frequency of overflow at present is
estimated to average about 40 occurrences per outfall per year,
of which perhaps five or six are summer occurrences.
Combined sewer and storm drain outfalls are scattered
throughout the Seattle area, discharging varying volumes to all
of the major water bodies around the city. Some of these
discharges are over shellfish beds, or as in Lake Washington,
near public bathing beaches or spawning areas for anadromous
fish; others are in the harbor area where water contact sports
or potential public health hazards are minimal.
Summer season dye studies done in 1976 indicated that the
nearshore surface circulation at two lake stations was compara-
tively sluggish, providing a maximum dilution of 2.7:1 (Lake
Washington) and 75:1 (Lake Union) during the first three hours
following release. Such conditions might be expected to favor
the nearshore settling of much of the sediment injected into
the environment by combined sewer overflows and storm drains.
Very little information has been available concerning the
dispersion of these materials.
To help fill the gaps in our knowledge of the distribution
and fate of wastewater particulates in Seattles's nearshore
waters, EPA and Seattle Metro cosponsored a research project
with the following objectives:
1. To determine the fate and distribution patterns of
suspended particulates emanating from representa-
tive combined sewer overflows and storm drains in
the Seattle area through the correlation of
quantitative in-situ observations with suspended
solids loading and current patterns.
2. To determine the distribution patterns and ultimate
fate of selected particulate wastewater contami-
nants, including copper, lead, zinc, carbon and
phosphorus.
METHODS AND MATERIALS
EFFLUENT
Quantity and Quality
For selected rainstorms the overflow effluents at a
116
-------
representative combined sewer outfall and storm drain (CSO 023
and SD 7 - refer to Figure 2 for locations) were sampled for
loading analysis of suspended solids, metals, total organic
carbon, total phosphorus, oils and greases, and total chlorina-
ted hydrocarbons. Automatic sequential samplers triggered. by
pressure switches were used in conjunction with Arkon pressure-
driven flow recorders to provide 2-liter samples at regular
intervals during the overflows. The sampling intervals were
chosen according to the speed of flow response in each system,
so that samples would be taken that were representative of all
the major high and low points on the overflow hydrographs .
Individual analyses were performed as follows:
Total suspended solids - Determined in conjunction
with the total chlorinated hydrocarbons analyses,
as specified below.
Copper, lead, zinc, aluminum and mercury - Split
sample, acidified one part with cone. HNOj . Filtered
one half of sample through a ,45y membrane filter
prerinsed with 1% HNOj and deionized water. Acidified
filtrate and unfiltered aliquot sample with cone.
HN03; digested the latter. Both portions analyzed for
Cu, Pb , Zn, Al by AAS , and for Hg by cold vapor flame-
less AAS.
Total phosphorus - Split sample, filtered one part
through Whatman #40 filter. Acidified samples were digested
as for Cu, Pb , and Zn . P04 measured as orthophosphate
by the ascorbic acid method (APHA, 1975) .
Chlorinated hydrocarbons - Analyzed particulates only.
Filtered sample through muffled, tare-weighed Whatman
GF/A glass fiber filters. Filter dried at 105°C,
cooled, weighed and Soxhlet extracted with acetone for
eight hours. Extracts dried, infused with petro-
leum ether and concentrated. Cleanup and fractiona-
tion of dissolved components done in Florisil columns
by elution with various mixtures of ethyl and petro-
leum ether. Analyses done by GC, with a Tracer Model
222 GC fitted with a ^Ni, high temperature, electron
capture detector.
Oils and greases - Hexane/Soxhlet extraction method
(APHA, 1975) .
Total organic carbon - K2S20g/H3P04 digestion of
blended samples and standards in nitrogen-bubbled,
sealed ampoules. Resultant C02 quantified using a
Model 865 Beckmann Infrared Analyzer.
117
-------
Estimates of total loading for each measured parameter for
each outfall for each storm were made with the aid of a com-
puter program designed to calculate a time integral product of
concentration times flow for each storm hydrograph. To this
end the flow data were entered as values measured at regular
internals throughout. The quality analyses were less abundant,
being limited specifically to the significant features of the
hydrograph, i.e., the highs, the lows and the initial and final
ascending and descending limbs. Between these points, values
were interpolated to complement the flows given at regular in-
tervals. A storm was defined as a minimum total of .03 in. of
rain separated from preceding and succeeding measurable rain-
fall by a minimum of three hours.
Single quality samples contained insufficient particulate
material for chlorinated hydrocarbon analyses. Therefore,
sequential samples were composited to represent specific
features on the hydrographs. With the exception of the final
storm monitored at each of the two principal outfalls, the
sediment loading values were likewise calculated from storm
segment composites.
For the storm drain, where a dry weather base flow is
discharged, pollutant mass loadings were calculated for both
the base flow and the storm washoff—the latter being
determined as the difference between the total storm load and
the base load estimated from preceding flows and concentra-
tions. Total rainfall volume on each service area was also
determined.
Plume Distributions
During a given storm a grid of up to 24 stations was covered at
each sample site; at each grid point a Martek Model XAS
Transmissometer was lowered to the bottom, giving a real-time
plot of percent light transmission versus depth on a shipboard
XY plotter. Each complete grid coverage required about one
hour, and was repeated whenever possible in order to measure
various stages of the plume development and dispersion.
For grid-site location, a one-time survey was carried out
wherein onshore-offshore transects were established visually
using landmarks. Rangefinder distances were then determined
along these transects using buoy markers when needed. Depth
readings from the shipboard depth-finder were recorded for each
point and used for all subsequent outings to determine station
locations along the transect. The estimated maximum positional
error for the stations farthest offshore was 15 m and was some-
what less for inshore stations.
Mounted on the transmissometer sensor, temperature and
oxygen probes were used to periodically check the stratification
118
-------
of the water column. These sensors were interfaced to a Martek
Mark V Water Quality Monitor and an XY plotter. Both units were
lab-calibrated prior to use in the field.
The resultant vertical profiles of light transmission
versus depth were digitized and keypunched for computer --contour-
ing. For contouring purposes an interpolative contouring
routine was used to fit a bivariate elastic surface to
the initially coarse, raw data grid, resulting in smooth contour
output. This program was a modified version of a program
described in Numerical Plotting System User's Manual No. W00053
(University of Washington; August, 1977). , - • .
SETTLING PARTICULATES :
Settling particulate matter was sampled continuously from
1/30/78 to 1/29/79 near the wastewater outfalls at CSO 023 and
3D 7, and from 10 to 60 m offshore at the Control 3 site
(Figure 2). The sampling devices, or sediment traps, consisted
of 30 cm X 30 cm polyvinyl chloride platforms, which held four
10-cm diameter funnels joined to 50-ml centrifuge tubes
(Figure 1). All parts of the platforms (including screws) were
made of plastic to limit metals contamination. The collection
surface for each trap was suspended 2 m above the lake bottom.
For each collection period the traps were serviced and cleaned,
with retrieval and reset being done by SCUBA divers- to minimize
agitation of bottom sediments.
Arrays of six sediment traps were moored at each station,
with each array covering approximately .5 acre (Figure 4) .
Sediment trap placement was determined by visual and photo-
graphic observations of the divers. The rather distinct line
of transition between the effluent debris and the ambient
sediments was denoted the plume boundary.
The sediment trap samples were centrifuged at 9000 rpm for
7 minutes, decanted and dried at 60°C for 36 hours. The samples
were then weighed individually and the four samples for each
trap were combined and comminuted prior to chemical analysis.
One hundred mg subsamples were then digested, using
HF/HN03/BClG4 (Bortleson and Lee, 1972) in a 10-ml teflon
crucible (Birch, 1976). An aliquot of this digestrate was used
for determination of P, Cu, Pb, Zn and Al.
The analyses were done as follows:
Phosphorus - Ascorbic acid molybdenum blue method
(APHA, 1975). The mean of duplicates was reported
for each sample.
119
-------
POLYURETHANE
FLOATS
CENTRAL LINE
SUPPORT LINE
PVC LATCH
PVC PLATFORM
FUNNEL
TYGON SLEEVE
BO ml CENTRIFUGE
TUBE-
BRASS SNAP LOOP
EXCESS LINE
CONCRETE ANCHOR
SEDIMENT-WATER
INTERFACE-
Figure 1. Details of sediment trap configuration.
120
-------
PUGETSOUND
PUGET SOUND
I I FINAL STUDY SITE
• COMBINED SEWER STUDY SITE
0 STORM DRAIN STUDY SITE
O CONTROL STUDY SITE
-COMBINED SEWER OUTFALL
STORM DRAIN OUTFALL
1 SANITARY SEWER
EMERGENCY OUTFALL
D}- PUMP STATION-
LOCAL JURISDICTION
*!•»• METRO PUMP STATION
SEWER DISTRICT BOUNDARY
Area not mapped
Figure 2. Locations of sampling sites.
121
-------
Copper, lead, zinc and aluminum - Digestrate
analyzed by AAS. Al also analyzed by neutron
activation.
Total carbon — Analyzed 20 to 100 mg of dried sediment
using a Leco induction furnace and carbon analyzer.
Results reported as the mean of duplicate analyses.
SEDIMENTS
Surface Samples
Sampling for preliminary analyses of the freshwater sedi-
ments was carried out by SCUBA divers. Each station was
initially surveyed as to depth and bottom contours, nature and
condition of the outfall pipe, extent of apparent debris
accumulation, and proximity and nature of intefering structures.
At each CSO and SD, duplicate samples of the surface 3 cm of
the (apparently) most highly contaminated sediments were
collected in specially prepared llO-cnr*, glass cylinders closed
at each end with rubber stoppers0..The stoppers of one of each
set were covered with aluminum foil to facilitate pesticide/PCB
analyses. For the aluminum analyses the stoppers of the other
were left bare. ;
The excess water was drained off each of the surface
sediment samples collected for the preliminary study, and the
sediment was transferred to glass containers for PCB and total
chlorinated hydrocarbon analysis and polypropylene containers
for all other analyses. The samples were weighed wet and then
oven-dried at 105°C. Aliquots taken for Hg analysis were
dried at 50°C to minimize loss through vaporization. All
samples were weighed dry when cool.
Individual analyses were performed as follows:
Copper, lead and zinc - Analysis as for settling
particulates.
Mercury - Dried sediments digested for one hour in
cone. H7SO4/conc. HNCUmixture, infused with a
KMnO3/K^S_Op-mixture and digested for an additional
two hours. Analyses were done by cold vapor,
flameless AAS.
Total phosphorus - Digestion and analysis as for
effluent.
PCB/chlorinated hydrocarbons - Wet sediments
extracted and analyzed as for effluent.
122
-------
Oils and greases - Hexane/Soxhlet extraction method
(APHA, 1975).
Total organic carbon - Preliminary study:
inorganic carbon driven off as CO- by acidification
of samples to pH = 2.0 with HC1. Residual (organic)
carbon measured as combustion C02 by a Leco Model
WR-11 Total Carbon Analyzer. Intensive studies: as
for settling particulates.
Cores
Lake sediment cores for analysis of sediment profiles of
metals and carbon were also collected by divers at the trans-
missometer grid sites, using 30 cm lengths of clear tenite
butyrate tubing ("3.5 cm. ID). Cores from depths greater than
15 m were taken with a 7-kg modified Phleger gravity corer.
The cores were examined for gross physical characteristics
and then extruded and sliced at 0.5 cm and at 1-cm intervals
to 12 cm depth. The slices were weighed, dried at 103°C for
24 hours, reweighed and ground in a Diamonite mortar. Diges-
tion and analysis were done as described for settling particu-
lates on the 0-0.5 cm, 0.5-1 cm, 3-4 cm and 7-8 cm sections.
Analysis for chlorinated hydrocarbons was as for effluent.
Some cores were stored at 5°C for up to one week before
processing.
RESULTS AND DISCUSSION
PRELIMINARY STUDIES
Selection of Sampling Sites
Prior to the present study, no comprehensive map was
available giving the locations and dimensions of the sewer
outfalls in Lake Washington. Drawing information from a
number of sources, we compiled as much of the available out-
fall information as feasible to provide a comprehensive choice
of study sites. A base map constructed from subbasin"drainage
charts (RIBCO, 1974) was augmented with further details obtained
through visits to the files of agencies representing the nine
sewer districts surrounding the lake. The resultant map
(Figure 2) included 23 emergency outfalls (activated only by
power failure), 34 CSOs, 56 pump stations and 240 SDs. For
Lake Washington's 71.5 miles of shoreline, this gave a conserva-
tive average of five outfalls per mile. The map, although in-
complete, was felt to include all of the major drainage systems.
123
-------
On the basis of the compiled information and further
details obtained through field outings, sampling sites (10 CSOs,
10 SDs and 10 control locations) were selected for the pre-
liminary study (Figure 2). In this, we were guided by three
principal criteria: 1) that the locations be as free as
possible of contamination from other outfalls; 2) that potential
difficulties with monitoring equipment installation, servicing
and protection be minimal; arid 3) that the selected facilities
have a high probability of large and frequent storm overflows
or direct drainage responses.
Outfall interferences were judged from map information.
Data on estimated (modelled) overflow frequency, volume and
duration for each of the CSOs were obtained from the City of
Seattle Engineering Department; because similar information was
unavailable from the SDs, those facilities having the largest
drainage basins were chosen for study.
Relative Levels of Sediment Enrichment
At each of the CSO, SD and control sites selected, samples of
the surface (3 cm) layer of sediments were collected by divers
at the near-outfall locations thought to have the maximum
potential for contamination by the wastewater effluents.
These samples were analyzed for Cu, Pb, Zn, Hg, total P, TOC,
oils and greases, PCBs and chlorinated hydrocarbons.
Figure 3 represents an attempt to composite the resultant
information. The "sediment enrichment ^factors" are given for
three categories of parameters: l)-org*anics (TOC plus oils and
greases) and total phosphorus, 2) metals (Cu, Pb, Zn, Hg), and
3) chlorinated hydrocarbons and P.CBs. To calculate these
relationships, mean values were derived for each parameter at
each station and the resultant figures were normalized by
parameter, i.e., the lowest value found for each parameter was
divided into each of the corresponding values from the other
28 stations. These figures were then added for each station
for each category, and the sums were in turn normalized within
the categories. Thus, the total length of the three histogram
bars at each station, divided by 3.0, gives the overall enrich-
ment of that station compared to a hypothetical composite of
the most pristine sediments found in the lake. The sediments
from CSO 046, which does not appear to overflow, approach this
quality overall. The background concentration standards used
for the Figure 3 calculations were all comparable to or lower
than those determined for pre-1900 core segments collected in
the deep portion of the lake by Birch (1976) and Spyridakis and
Barnes (1976).
Sediments around the sewer outfalls on the eastern shore
of Lake Washington (where there are no combined sewers)
generally appeared to be only one-third to one-fourth as con-
124
-------
t- at to Q i
O O O (0 !
S 5
3 O
,1 .5
s If**
o
S3 _
5 •
O1
rf
-o
1
(tj
™ O
S CT
s s
(
I J3
<
CM"
£0- •£ <
8 A
o
Q.
11
'+
•Esl
5 =
IS
-§
M- -2 'N .£
s 2|-rf|
P- y} (o c
aj "D o> UJ
El. "" >
_ o ^-- c
JH *^- O3 '~
o .2 *- S.-5
" >- X Q. —
C 0) o °>
, gS2
cc i E <
co'
0)
D>
in
125
-------
taminated overall as those between the two floating bridges and
south of Seward Park (denoted by C 4 and C 5) on the western
shore. Sediment pollutants at several stations in this area
were enriched more than 16 times over background levels. The
area between the bridges had particularly high concentrations
of metals, which are typically associated with both CSOs and
SDs.
INTENSIVE STUDIES
Description of Study Sites
The locations ultimately chosen were the best available
but did not fully meet all of our specified criteria. For
either of the two outfalls selected there was a substantial
potential for effluent interference from other, nearby
systems. This is evident in Figure 4, which shows the site
configurations. In each case a large storm drain was found
within several hundred feet of the sample outfall. However,
the approximate dimensions of the zones of debris accumulation
observed by the divers indicated that the separation distances
might be sufficient to minimize interferences; also, probable
current patterns deduced from dye studies in the general
vicinity of the CSO (CH2M/Hill, 1974) indicated that the waste-
water plumes there were likely to be parallel in most instances
and not convergent.
The combined sewer outfall (designated CSO 023) was found
to have a scoured pit at the end, with a built-up pile of debris
somewhat further offshore. The pit was approximately half a
meter deep and one meter wide, whereas the mound was about 1.5 m
high and 14 m across. The discharge point for the principal
storm drain system (SD 7) was nearshore in localized riprap and
opened onto much steeper terrain than did that of the CSO, thus
minimizing scouring and debris buildup. The outfall structures,
which both had pipe diameters of 24 inches, were also found to
have transient contiguous areas of visible debris accumulation
as large as .5 acre, as indicated in Figure 4. Such evidence
of scouring and deposition was found to be common at the many
outfalls investigated for the preliminary studies.
The locations of the sediment traps and transmissometer
grid sites at CSO 023, SD 7 and the control station (C 3) are
also shown in Figure 4. The typical substrate in the various
areas was: CSO 023 - fine sand, silt, light debris; SD 7 - fine
sand, silt, some clay, light debris; control site (C 3) - sand.
126
-------
I(a) Combined Sewer Outfall 023
30m W
OUTFALL
LINE
I'30m E
84m E
OUTFALL
107m E
8m W
SED. TRAP
ARRAY
23m E
53m E
84m E
T SEDIMENT TRAP
© OUTFALL TERMINUS
ZONE OF VISIBLE EFFLUENT DEBRIS
TRANSMISSOMETER SAMPLING LOCATION
Figure 4. Principal set of Intensive sampling sites in Lake Washington.
127
-------
Overflow Monitoring
Effluent Loading Estimates for Single Events—
The wastewater effluents generated by several storms were
sampled quantitatively and qualitatively at CSO 023 and SD 7
to determine typical particulate inputs to the freshwater
environment by a representative CSO and SD. For the five
storms monitored the rainfall range was 0.17 to 1.05 in., and
the discharge volumes varied from 51 to 951 m3 at CSO 023 and
213 to 1849 m3 at SD 7. The months represented include the
end of the rainy season and most of the dry season. Because
the drainage basins for the two stations sampled were only
1.5 miles apart, with similar orientation and no intervening
ridges, a single rainguage was used for both.
The similarities in relative size of the two drainage
basins (CSO 023 - 117 acres, SD 7 - 125 acres), and in incident
rainfall for any given storm, invited comparisons of their
pollutant loading and runoff characteristics. The major
distinction affecting the relative quality of runoff for these
two areas is land-use distribution; the CSO 023 basin is
comprised of approximately 75% single-family residences with
the rest multiple-family residences, whereas the SD 7 basin has
a significantly higher fraction o£ single residences (normally
associated with lower pollutant accumulation rates than multiple
residences). For three of four storms an appreciably lower
percentage of the total rainfall on its drainage basin was
discharged by CSO 023 than by SD 7; this was to be expected
since a CSO discharge is typically intermittent, whereas that of
an SD is continuous. On the other hatid, the relative magnitudes
of the mean storm discharge concentrations for the various
pollutants varied from one storm to the next, as documented
below. A statistical summary of the contaminant mass discharges
is offered in Table 1. Mean ratios (SD 7:CSO 023) of per-
storm mean concentrations and total loadings are also presented
here for each parameter, in support of the ensuing discussion
(Table 2).
Particulate effluent—With reference to Table 2 the
SD 7:CSO 023 ratios of mean concentrations in particulates were
appreciably less than 1.0 for suspended solids, Cu, total P,
oils and greases, and chlorinated HC; the particulate ratios
were near unity for Zn, Al and total organic C. Only Pb had a
significantly higher solids concentration at SD 7, this being
evidently due to dilution by low-lead sanitary wastes at the
CSO. However, the mean discharge volume for SD 7 was over
three times that of CSO 023; consequently the mean storm
particulate loadings for all parameters except total P were
greater at SD 7 than CSO 023. The mean loading values for
total P were similar for the two sites, giving a SD:CSO ratio
near 1.0.
128
-------
TABLE 1. STATISTICAL SUMMARY OF ESTIMATED POLLUTANT LOADS AND CON-
CENTRATIONS IN STORM DISCHARGES MONITORED AT STORM DRAIN
7 AND COMBINED SEWER OUTFALL 023, MARCH-SEPTEMBER, 1978
CSO 023
SD 7
MASS RANGE (kq)
Suspended Solids
Total Cu
Total Hg
Total Pb
Total Zn
Total Al
Total Org C
Total P
Total O & G
Partic Chlor HC
PARTICULATE
MASS (% of Total)
Suspended Solids
Cu
Hg
Pb
Zn
Al
Org C
P
O & G
Chlor EC
DISCHARGE
CONC ( ppm)
Suspended Solids
Total Cu
Total Kg
Total Pb
Total Zn
Total Al
Total Org C
Total P
Total O & G
Partic Chlor HC
6.40 -
.001 -
.0000 -
.001 -
.003 -
.059 -
.865 -
.059 -
.584 -
.316 -
x
100
78.1 +
88.4 +
69.1 +
68.6 +
96.0 +
42.1 +
28.5 +
NA
ND
X
153
.052
.0010
.057
.125
2.58
17.9
1.76
12.9
018 ppb
No.
Storms
142
.039
.0008
.041
.113
2.56
18.0
1.28
9.41
100 mg
(4)
(4)
(4)
(4)
(4)
(3)
(2)
(4)
(4)
(4)
No.
+ la Storms
.0
10.2
6.9
14.2
19.6
2.0
22.9
10.3
Range
90.8 -
.026 -
.0003 -
.025 -
.061 -
1.18 -
16.6 -
1.05 -
6.71 -
.004 -
(4)
(4)
(2)
(4)
(4)
(3)
(2)
(4)
of la
258
.105
.002C
.130
.258
5.65
19.3
2.94
25.0
.081
27.4 -
.020 -
.0000*-
.098 -
.044 -
.762 -
11.0 -
.120 -
.740 -
1.02 -
x +
100
64.0 +
ND
88.1 +
64.1 +
96.6 +
56.0 +
52.7 +
NA
ND
x
80.8
.042
) ND*
.218
.099
3.51
10.8
.315
4.08
.014 ppb
142
,049
.0004
.316
.156
6.62
16.2
.417
8.06
86.1
la
.0
8.7
*
2.7
5.0
0.9
8.1
13.1
Range
43.8 -
.026 -
ND
.127 -
.066 -
2.83 -
8.09 -
.153 -
2.94 -
.006 -
No.
Storms
(5)
(5)
(5)
(5)
(5)
(4)
(2)
(4)
(5)
(5)
No.
Storms
(5)
(5)
(5)
(5)
(5)
(4)
(2)
(4)
of la
149
.068
.372
.150
4.36
14.4
.648
5.66
.034
(1) Arithmetic mean
(2) Geometric mean
NA Not applicable
ND Not determined
* Below detection limits
129
-------
TABLE 2. INTERSTATION COMPARISON FOR STORM DRAIN 7 AND COMBINED
SEWER OUTFALL 023 OF MEAN EFFLUENT CONCENTRATIONS AND
POLLUTANT LOADINGS ESTIMATED FOR FOUR STORMS H)
MONITORED AT BOTH SITES
Geom Mean of SD 7:CSO 023 Ratios'2^
Parameter Fraction Mean Cone Pollutant Load
Susp Solids
Cu
Hg
Pb
Zn
Al
Total Organic C
Total P
Oils & Greases
Chlorinated HC
Total
Partic
Total
Partic
Total
Partic
Total
Partic
Total
Partic
Total
Partic
Total
Partic
Total
Partic
' Total
Partic
Total
0.6
0.6
0.8
*
0.8
5.1
3.9
0.8
0.8
1.2
1.2
0.9
0.6
0.4
0.2
NA
0.3
0.7
ND
1.8
2.0
2.7
*
0.6
16.1
12.5
2.6
2.7
6.3 O f L6
6.3
5 . 0 ' ? ^ ;
3 . 4 -£iiz
'" r, , ,
1.1
0.6 •••"•"•
NA
0.9
2.4
ND
Discharge Volume "• 3.3
(1) Storms of 3/23/78, 3/24/78, 4/15/78, -8/31/78
(2) Ratios initially calculated for each storm
* Below detection limits
NA Not applicable
ND Not determined
130
-------
There were consistently higher mass loadings of Al from
SD 7 than from; CSO 023,. indicating a greater input of inorganic
terrigenous materials to the lake by the storm drain. There
was a large variation from one storm to the next in the mean
concentrations arid total mass loading of chlorinated hydro-
carbons at both sites (Table 1). Trends at the two stations
were similar in that DDT accounted for a very large portion
(approximately 75%-95%) of the total loading of particulate
pesticides during the spring storms, and only about half of the
August-September storm total, which was itself appreciably lower
than that of the other storms. Lindane was the only other
pesticide consistently present at both stations. The persistent
occurrence of DDT and lindane conforms to the observations of
Brenner et al. (1978) for runoff from the Juanita Creek watershed
in the northeastern segment of the Lake Washington basin. The
occurrence of these contaminants in both the CSO and SD-systems
indicates that they were present in the storm drainage, rather
than in the sanitary wastes.
Total effluent—The largest concentration discrepancies
between the two stations were for total suspended solids, Pb
and those parameters related to domestic wastes, i.e., total
organic carbon, total phosphorus,and oils and greases; SD 7 had
substantially lower mean total effluent concentrations than CSO
023 for all of these except Pb, which was appreciably higher
there than at the CSO. Only Hg and total P had lower mean total
pollutant loadings at SD 7 than at CSO 023.
Annual Rainfall, Flow and Loading Summary—
The total rainfall, on the Seattle central business district
for the March, 1978-February, 1979 study period was 38.9 in.,
a value near the annual average of 34.1 in. (Phillips, 1968).
However, the total rainfall, recorded for the study area on^ the
western shore of Lake Washington was only 23.2 in., a vaiue°t;
confirmed by three separate gauges. It is probable that this
difference represents a shielding effect, or rain-shadow, with
the hills of the city blocking part of the precipitation from
the storm fronts, which typically come from the southwest.
Lacking long-term, confirming data, it was assumed that this
effect is typical for the study area, and that the rainfall
measured for the study period was near the corresponding-annual
average. The 1978 spring and summer were somewhat wetter than
usual, but the 1978-1979 winter was drier, compensating for most
of the observed difference. For a total of 62 periods of
measurable rainfall (separated by one or more dry days) the
fraction of those storms resulting in increased discharge at one
or both outfall stations was 76%; the average total rainfall for
those storms that failed to generate runoff effluent was .04 in.
The total discharge volumes for the 12-month monitoring period
were: CSO 023 - 25,600 m3 C6.76 MG); SD 7 - 27,700 m3
(7.34 MG) , + 50,200 m3 (13.3 MG) base flow. These values, were
used with the mean discharge concentrations given in Table 1
131
-------
and with concentrations measured for nonstorm discharge at
SD 7 to calculate total annual loads for the various parameters
The results are presented in Table 3.
TABLE 3. ESTIMATED TOTAL MASS OF SELECTED CONSTITUENTS IN EFFLUENT
DISCHARGED BY COMBINED SEWER OUTFALL 023 AND STORM DRAIN
7 MARCH 3,1978 TO FEBRUARY 28,1979
Suspended Solids
Cu
Hg
Pb
Zn
Al
Organic C
Total P
Oils and Greases
Chlorinated HC
CSO 023
Storm
Total Particulate
Mass (kg) Mass (kg)
3920 3920
1.33 1.04
.026 .023
1.46 1.01
3.20 .2.20
66.0 63.4
458 193
45.1 12.8
330 NA
ND .460g
SD 7
Storm
Total Particulate
Mass (kg) Mass (kg)
2238 2238
1.16 .745
ND ND
6.04 5.32
2.74 1.76
97.2 93.9
299 168
8.73 4.60
113 NA
ND .388c
Non-Storm
Total Particulate
Mass (kg) Mass (kg)
666 666
.904 .663
ND ND
1.00 .502
.502 .050
10.0 ND"r '
151 100 . ,
59 "7 ~ ^ fl i 6 ,
* I") ri f J )
65.3 NA " -
• ND vi50g
NA: Not applicable
ND: Not determined
For all constituents except Pb and Al, the annual storm
loading of particulates was less from the storm drain than from
the combined sewer. However, with the nonstorm loading of the
storm drain added to its storm inputs, the totals for Cu, organic
C and chlorinated HC also exceeded those of the combined system,
and the Zn loading was nearly as great. The parameter relation-
ships were similar for the total loading (particulate + soluble),
except that the C and P inputs were considerably more augmented
for the combined sewer than for the storm drain due to the great-
er quantities of dissolved organics in the sanitary wastes of the
combined system.
Receiving Water Monitoring
Turbidity Distributions of Effluent Plumes--
The distributions and movements of wastewater effluent plumes
were monitored at CSO 023 and SD 7 during four different storms.
132
-------
Background particulate distributions were also assessed at the
control site during three storms. These measurements were made
using a submersible light transmissometer system of the type
originally described by Petzold and Austin (1968). This instru-
ment provides a measurement of the volume attenuation coefficient,
a, which is equivalent to the sum of the light scattering and ab-
sorption coefficients. The volume attenuation coefficient may in
turn be related to common turbidity units using methods described
by Austin (1973), who determined the relationship between a and
Jackson Turbidity Units (JTU) to be approximately linear for nat-
ural wateri'of moderate to good clarity.
For plume tracing, the light transmission (turbidity) data
were reduced to sectional contour plots. These included aerial
perspectives at various depths and onshore-offshore and long-
shore transects showing turbidity patterns from the perspective
of a diver looking across the lake bottom.
Contrary to intentions, we were unable to obtain light trans-
mission data, for periods of lake stratification. The data pre-
sented here were collected during March and April of 1978 and
February of 1979, and in all instances show the wastewater plumes
to have been intersecting the surface. At the combined sewer
outfall site this was due predominantly to the comparatively high
wastewater temperatures; e.g., a difference of approximately 6°C
was^measured for a comparison of effluent and ambient water at
C&C)5023 on 4/4/78. For the storm drain the shallow discharge
depth was the foremost determinant.
aG^-fhe various characteristics of the effluent plume dispersion
observed at CSO 023 are summarized in Figures 5 and 6. Figure 5
shows the residual turbidity distributions 21.5 hr. from the time
an overflow began on 3/7/78. The plume was seen as a lens of
turbid surface water 150 m NE of the outfall, 2.7m thick (Figure
6 (c)) and covering an area of approximately 11 acres. The ex-
tremely muddy water SW of the outfall (Figure 6 (a)) was tempo-
rarily uncontrolled drainage from a nearby building site. Inter-
ference from a large, nearshore storm drain was also observed,
being particularly obvious during the storm of 2/6/79, for which
contour plots of onshore-offshore transects clearly indicated a
tongue of turbid wastewater moving offshore from the drain site
80 m SW of the CSO.
Considerable effluent interference was also observed in
the turbidity distributions measured for storm discharges at SD 7.
Substantial wastewater inputs from a 24-inch storm drain 345 m
north of the SD 7 outfall were recorded for the storm of 2/6/79.
The relevant areas of influence for the two outfalls are strik-
ingly apparent in Figures 7 and 8, which give aerial turbidity
distributions at one and four hours following a 10-hour storm
discharge at SD 7.
133
-------
.;i .COMBINED SEWER OUTFALL 023
- ~f-.'$* • **• *ssss^ *a*m™&m*&mtiii£t.$!mm*.MS8£
10 feet (3.0 meters)
Figure 5. Aerial perspectives of contours of percent light transmission at Combined
Sewer Outfall 023 -1533-1645 hrs., 3/7/78. The lettered symbols indicate
surface positions of the wastewater plume core at time of overflow initiation
plus (a) 5 hr., 3/6/78; (b) 21.5 hr., 3/7/78 (distribution shown); (c) 3.5 hr.,
3/24/78 - 3/25/78; and (d) 3hr., 4/4/78.
134
-------
DEPTH
Figure 6. Longitudinal sections contours of percent light transmission at Combined
Sewer Outfall 023 -1533-1645 hrs., 3/7/78. The perspective is that of a
diver facing shore.
135
-------
10 feet (3.0 meters)
Figure 7. Aerial perspectives of contours of percent light transmission at
Storm Drain 7 -1915-2014 hrs., 2/6/78.
During the three-hour interim period the SD 7 plume began to
dissipate and the effluent from the other outfall became more pre-
valent. The small cove between the two outfalls appears to have
encouraged a local eddy in the nearshore circulation, resulting in
a southerly movement of effluent offshore. There were also mani-
festations of this clockwise circulation pattern, and ov particu-
lates discharged by the second storm drain, in storm data collec-
ted 3/6-3/7/78.
The surface position of an effluent lens as a function of
time gives an incomplete indication of particulate circulation
patterns. Much of the solid material present in these features
apparently remains suspended for long periods, as evidenced by
the lack of turbidity at mid-depths in Figure 6, and is
ultimately advected out of the area. Figure 9, however, shows
turbidity distributions for the onshore-offshore transect
61 m N ("downstream") of the SD 7 outfall, and indicates that
a substantial local fallout and settling of particulates occur
in an outfall area during an overflow; the settling pattern
denoted by the figure shows a downslope movement of materials
near the bottom typical of that observed near both the SD and
the CSO. These observations correlate well with the measured
distributions of sediment surface contaminants presented below.
136
-------
107m E
10 feet (3.0 meters)
Figure 8. Aerial perspectives of contours of percent light transmission at Storm Drain 7 -
The control area (C 3} was also found to be inundated with
The control area (C 3) was also found to be inundated with
suspended particulate matter. Light transmission contours drawn
for storm data collected there indicate the presence of a deep,
turbid layer believed to have been a longshore flow of settling
particulates moving from the northeast. The most likely sources
of this material were: 1) a series of 22 storm drains (including
eight > 24 in. diameter) carrying runoff from an area of active
construction 0.7-3.8 km to the north, and 2) a creek which dis-
charges highly turbid runoff from a 7,740-acre basin another 0.4
km beyond that (Figure 2).
The nature of turbid longshore advection through the C 3
area seemed to be quite variable - an extensive particulate
layer observed entering from the north on 3/25/78 was on the
lake surface, whereas that mentioned above was at depth. Re-
gardless, the sediment quality analyses below implied only light
settling of contaminated particles in this area, and fhe mea-
sured settling rate was substantially lower than for the CSO 023
and SD 7 areas.
Quantification of Settling Particulates--
Particulates captured by arrays of six sediment traps at
each station were dried, weighed and analyzed for C, P, Pb, Cu,
Zn and Al. The results are given in Table 4.
137
-------
UI
E
to
o
DEPTH
ft JB.
, (a) One hour after cessation of
storm discharge by SD 7
• (1915-2014 hrs.)
-5 1.5
i'10 3.0
15 4.6
i 20 6.1
< 25 7.6
' 30 9.1
35 10.7
40 12.2
45 13.7
(b) Four hours after cessation
of storm discharge by SD 7
(2244-2348 hrs,)
50 15.2
0
5 1.5
.10 3.0
' 15 4.6
20 6.1
25 7.6
30 9.1
35 10.7
40 12.2
45 13.7
50 15.2
Figure 9. Transverse sections through a point 61m N of Storm Drain 7, showing contours
of percent light transmission - 2/6/79. The perspective is that of a diver with
the shore to his left.
The mean loading rates in g/mz/day from 1/30/78 to 2/2/79 were
.752 for CSO 023, .805 for SD 7 and .268 for C 3. A relative
value for the adjacent profundal area is .661, which was deter-
mined by correcting a rate measured at 60 m depth (Birch, 1976)
for the offshore movement of particulates entering the lake in
the littoral zone. The offshore sedimentation rate was there-
fore less than nearshore values measured in the vicinity of the
wastewater outfalls, but greater than that determined for the
control area. A significant fraction of the suspended particu-
lates in the study area was transient material which would even-
tually contribute to the sedimentation rate in the profundal
zone.
138
-------
TABLE 4. MEAN SEDIMENTATION RATES AND DRY WEIGHT CONCENTRATIONS OF
SELECTED CONSTITUENTS ANALYZED IN SEDIMENT TRAP SOLIDS COLLECTED
AT COMBINED SEWER OUTFALL 023, STORM DRAIN 7 AND CONTROL SITE 3,
1/30/78-2/2/79
CSO 023
SD 7
C 3
X
Sx
X
X
.752
9.58
.252
366
126
349
4.26
.714
3.85
.192
104
108
152
1.07
.805
9.29
.212
521
94.8
286
4.70
.696
4.53
.211
355
44.8
127
1.13
.268
9.34
.273
245
70.8
244
3.72
.199
6.08
.205
97.1
41.8
125
1.09
Sed. Rate (g/m2/day)
Org. C (%)
Total P (%)
Pb (ppm)
Cu (ppm)
Zn (ppm)
Al (%)
2
The mean loading rates in g/m /day from 1/30/78 to 2/2/79 were
.752 for CSO 023, .805 for SD 7 and .268 for C 3. A relative
value for the adjacent profundal area is .661, which was
determined by correcting a rate measured at 60 m depth
(Birch, 1976) for the offshore movement of particulates entering
the lake in the littoral zone. The offshore sedimentation
rate was therefore less than nearshore values measured in
the vicinity of the wastewater outfalls, but greater than that
determined for the control area. A significant fraction of the
suspended particulates in the study area was transient material
which would eventually contribute to the sedimentation rate in
the profundal zone.
The flux (or quantity settling through a given cross-
sectional area per unit time) of particulate C, P, Pb, Cu, Zn or
Al was computed as the product of the constituent concentration
(wt./total wt. of solids) and the sedimentation rate. Compared
to CSO 023 and SD 7, the control site had similar or lower
concentrations of each settling contaminant and a significantly
(at the 99% confidence level) lower sedimentation rate; the
settling flux of all solid elements at C 3 was therefore
significantly lower than at CSO 023 or SD 7.
The average concentrations of Pb in the sediment trap
solids were significantly greater at SD 7 than at CSO 023—a
relationship in keeping with the relative Pb discharge
evaluations discussed previously (Tables 1 and 2). Cu and Zn
had similar concentration hierarchies for settling particulates:
CSO 023>SD 7>C 3. The mean Al concentrations determined for
solids settling near both outfalls were significantly higher
than for solids from the control site, implying a greater input
of erosional material.
139
-------
Core Analyses of Sediment Contaminants—
To help further define trends of movement and accumulation
of wastewater particulates in the receiving waters, sediment
cores were collected at the transmissometer grid sites at
Stations CSO 023, SD 7 and C 3. Fifty-six cores were dried,
sectioned and analyzed for Pb, Zn, Cu and TOC. Their
contour distributions in the 1-cm sediment surface layer are
presented here in Figures 10 and 11.
w
LEAD (ppm)
30m W
OUTFALL
LINE
30m E
84m E
ZINC (ppm)
/V>_40
/// 60
^221 80
COPPER (ppm)
4.0
CARBON
Figure 10. Dry weight distribution of lead, zinc, copper and carbon in the surface
centimeter of sediments collected near Combined Sewer Outfall 023.
140
-------
30m W
OUTFALL LINE
i 15m E
46m E
107m E
LEAD (ppm)
ZINC (ppm)
COPPER (ppm)
% CARBON
Figure 11. Dry weight distributions of lead, zinc, copper and carbon in the surface
centemeter of sediments collected near Storm Drain 7.
The contours show enrichment of the sediments near the two
outfalls, with apparent distribution modification from current
action and near-bottom downslope streaming. The isopleths for
Pb and Zn around CSO 023 show the effects of near-bottom
turbidity plumes moving downslope to the southeast (refer to
141
-------
Figure 4 for areal bathymetry). For Cu the concentration
isopleths imply the additional influence of advective transport
toward the northwest, where the particulates settled due to flow
disruption by condominium support pilings. Although the
prevalent direction of flow of subsurface currents near the CSO
was not clearly defined by the light transmission studies, the
northwesterly advection implied by the sediment distributions is
logical as the most direct route to the flow outlet for Lake
Washington.
The considerable differences in sediment distributions for
Pb and Zn vs. Cu near CSO 023 support the hypothesis that the
Cu-bearing effluent particulates were smaller and/or lighter,
and thus more easily influenced by water motion. Carbon was
found to be more widely dispersed than the metals, in keeping
with the nature of the large, light carbon-bearing particulates.
Correlation coefficients calculated for concentrations of Cu,
Pb and Zn vs. C in CSO 023 sediments were 0.65, 0.42 and Q.24,
respectively, implying that the relatively light Cu and C
particulates have the most similar transport characteristics,
and further confirming the Cu>Pb>Zn dispersion hierarchy
suggested by Figure 10.
At SD 1, substantial advective losses of wastewater
particulates from the sampling area result in a considerably
different pattern of local deposition compared to that seen at
CSO 023. Near the SD 7 outfall the sediment distributions of
PB, Zn, Cu and C were all found to be quite similar (Figure 11),
indicating the dominant influence of near-bottom downslope
streaming just north of the outfall. Based on the strength of
correlations between C and metal concentrations
(Pb: r = .92, Zn: r = .89, Cu: r = .44) the dispersion
hierarchy for SD 7 was determined to be Pb>Zn>Cu.
Circulation Analysis for Wastewater Solids—
Table 5 is a compilation of metals data for all types of
particulates sampled at the three principal study sites0 These
numbers provide a useful aid for tracing the movements of
wastewater solids.
142
-------
TABLE 5. MEAN CONCENTRATIONS (1) AND CONCENTRATION RATIOS FOR SELECTED METALS
IN PARTICULATES SAMPLED AT COMBINED SEWER OUTFALL 023, STORM DRAIN 7 AND
CONTROL SITE 3
Mean Cone.
Location
CSO 023
SD 7
C 3
Source
Effluent*2^
Sed. Traps
0-.5cm Sed.
7-8cm Sed.,
Effluent*2)
Sed. Traps
0-.5cm Sed.
7-8cm Sed.
Sed. Traps
0-.5cm Sed.
7-8cm Sed..
Pb
257
366
129
51
Zn
560
349
168
92
Cu
(ppm)
Pb: Cu
Zn:Cu Zn:Pb
265
126
39
40
2377
521
89
46
785
286
61
35
333
95
47
42
245 244
86 131
21 51
Profundal*3) 0-lcm Sed.,
Pre-1900 Sed.
192
12
192
59
71
20
7
46
17
1.0
2.9
3.3
1.3
7.1
5.5
1.9
1.1
3.5
4.4
2.9
4.2
1.0
2.1
2.8
4.3
2.3
2.4
3.0
1.3
0.8
3.4
6.6
7.1
4.2
3.5
2.2
0.9
1.3
1.8
0,
0,
0.7
1.4
1.0
1.5
2.5
1.0
3.5
(1) Dry weight.
(2) Calculated from data in Table 1, as ppm of suspended
solids in the effluent.
(3) From cores collected in the deep central basin of
Lake Washington, between CSO 023 and SD 7
(Spyridakis and Barnes, 1976).
The concentrations of metals in the sediment trap collections
were generally 2 to 5 times those of the sediment surface
(0-0.5 cm) layer, implying selective removal to deeper areas,
of the finer particulates with which the highest metal concen-
trations are typically associated (Guy and Chakrabarti, 1976).
Sediment cores collected by Spyridakis and Barnes (1976) along
an across-lake transect contiguous to CSO 023 and SD 7 clearly
show an increase of Pb, Zn and Cu concentrations with lake depth.
As is evident from a comparison of the CSO 023 and SD 7 sediment
trap data with values for profundal (deep offshore) surface
sediments (Table 5), however, the latter materials had been
appreciably diluted by relatively uncontaminated secondary
sources, including diatoms and river detritus. Spyridakis and
Barnes have estimated that contemporary fluvial inputs
(including wastewater discharges) are responsible for 26%, 90%
and 53%, respectively, of the Pb, Cu and Zn entering the lake.
The selective transport tendencies of Cu, Pb and Zn at
CSO 023 are also delineated by the concentration ratios given
in Table 5. The Zn:Cu ratio increases as the particulates move
from the outfall to the traps, and again, from the traps to the
sediments; this is because of the higher losses of Cu through
143
-------
advective transport. The decrease of the Zn:Pb ratio
represents an external input to this system - particulate Pb
contributions from aeolian transport (Spyridakis and Barnes,
1976). The subsequent increase in the Zn:Pb ratio between the
traps and the sediments implies a shorter local residence time
for Pb. All of the sediments above 8 cm depth at CSO 023 have
been deposited since the addition of tetraethyl lead to gasoline
in the late 1920s, as is evident from their substantial Pb
enrichment relative to pre-1900 (background) concentrations.
Estimates of median effluent particle diameter show that
the wastewater particulates emitted by SD 7 are typically much
smaller than those from CSO 023, facilitating advective losses;
this helps to account for the comparatively precipitous drop of
particulate metals concentrations between the SD 7 effluent and
the sediment traps; the concentrations of all three metals in
the effluent particulates were substantially higher at SD 7 than
at CSO 023, but due to greater advective exports, sediment trap
concentrations were higher only for Pb. The steeper bathymetry
of the SD 7 sampling site also helps aggravate the offshore,
near-bottom movement of effluent particulates.
Typical enrichment factors for total metals contributions
from all sources to the nearshore surface sediments may also be
derived from Table 5. For CSO 023, SD 7 and C 3, respectively,
they are: Pb - 11 times, 7 times, 7 times; Zn - 3 times, none,
2 times; and Cu - 3 times, 3 times, none. Overall, the SD 7
area was not found to be much different than the control site
in this sense, a conclusion statistically confirmed by an
analysis of variance for mean station concentrations of Pb, Zn,
Cu and C in the 0.5 cm sediment surface layer. This is due
predominantly to the rapid movement of effluent particulates
out of the SD sampling area and to the contamination of the
control site by longshore advection.
SUMMARY
The analyses of sediments collected near ten combined
sewer outfalls (CSOs) and ten storm drains (SDs) discharging
into Lake Washington revealed their widespread enrichment with
heavy metals, chlorinated hydrocarbons and organic wastes. This
was particularly true of the western nearshore region, where all
of the CSOs were located. Intensive comparative studies were
carried out at one representative CSO and one SD having similar
drainage basins and rainfall. Pollutant loading estimates
indicated that, whereas mean storm concentrations for most
parameters were greater for the CSO, the total annual loading
was greater in most instances for the SD, due to its continuous
and therefore greater volume of discharge. Effluent turbidity
patterns- and metals concentration ratios for effluent solids,
144
-------
settling particulates and sediments demonstrated significant
local deposition of wastewater particles. The effects of
near-bottom, offshore transport and longshore advection were also
indicated, with different dispersion patterns seen at the SD
than at the CSO.
REFERENCES
AP.EA. 1975. Standard Methods :for the Examination of Water and
Wastewa'ter. 14th Edition. Amer. Pub. Health Assoc.,
Amer. Water Works Assoc., Water Poll. Cont. Fed., Wash.,
D.C. 1193 pp.
Austin, R.W. 1973. Problems in measuring turbidity as a water
quality parameter. Paper presented at Environmental
Protection Agency Seminar on Methodology for Monitoring
the Marine Environment. October 16-18, 1973. Seattle,
Wash.
Birch, P.B. 1976. The Relationship of Sedimentation and
Nutrient Cycling to the Trophic Status of Four Lakes in
the Lake Washington Drainage Basin. Ph.D. dissertation,
Univ. of Washington, Seattle.
Bortleson, G.C. and G.F. Lee. 1972. Recent sedimentary history
of Lake Mendota. Wis. Env. Sci. and Tech. 6^:799-808.
Brenner, R.M., R.J. Morrice and R.G. Swartz. 1978. Effects of
Stormwater Runoff on the Juanita Creek Drainage Basin.
Report for the Municipality of Metropolitan Seattle.
59 pp.
CH2M/H111. 1974. Storm Water Outfall Study of Lake Washington.
Report for the City of Seattle. 50 pp.
Guy, R.D. and Chakrabarti. 1976. Studies of metal-organic
interactions in model systems pertaining to natural waters.
Can. J. Chem. 54:2600-2611.
Petzdld, T.J. and R.W. Austin. 1968. An Underwater Trans-
missometer for Ocean Survey Work. SIO Ref. 68-9, Scripps
Inst. of Oceanography, Visibility Laboratory. Univ. of
Calif., San Diego.
Phillips, E.L. 1968. Washington Climate. Coop. Ext. Serv.,
College of Agriculture, Wash. State Univ., Pullman.
65 pp.
RIBCO. 1974. Environmental Management for the Metropolitan
Area, Part T. Water Resources Appendices. Compiled by
CE2M/ffill for the River Basin Coordinating Comm., Seattle.
145
-------
Spyridakis, D.E. and R.S. Barnes, 1976. The Effects of Waste
Water Diversion on Heavy Metal Levels in the Sediments of
a Large Urban Lake. Compl. Rpt. to OWRT (Proj. No.
A-070-WASH) . Dept. of. Civil Engineering, Univ. of Wash.,
Seattle.
146
-------
THE ECOLOGICAL EFFECTS OF
URBAN RUNOFF ON STREAM COMMUNITIES
Donald B. Porcella
Tetra Tech, Inc.
Lafayette, California
ABSTRACT
A literature review of the ecological effects of urban runoff on
streams indicates a dearth of principles for evaluating the impact of urban
runoff or any pollutant on streams and a serious lack of studies that deter-
mine urban runoff impacts on communities of stream ecosystems. The uniqueness
of urban runoff impacts results from the nature of pollutant input to streams
—large magnitued variation/stochastic occurrence. Most stormwater research
has been directed towards determining runoff quantity and quality including
the fates of specific pollutants such as heavy metals, nutrients, salts, toxic
substances, and bacteria. Demonstration of the impacts of these materials
on stream communities is necessary to justify costs of managing stormwater
effects.
Several approaches for analyzing ecosystems are suggested, such as
stream community analysis and biogeochemical cycling of elements (carbon,
nitrogen, phosphorus). Three published case studies of stream ecological
impacts typify the effects of urban runoff inputs resulting in a hypothesis
that large scale variations and instabilities of stream pollutant inputaand
concentration would result in greater impact to stream communities than steady
inputs. This concept is discussed and applied to urban runoff analysis and
management.
INTRODUCTION
Materials in urban runoff impact stream ecosystems in several specific
and unique ways that distinguish those impacts from other linked systems.
Published literature was used in an attempt to determine whether those unique
features result in ecological differences that can be distinguished by appro-
priate ecosystem variables.
THE URBAN NPS/STREAM SYSTEM
By definition, materials in urban runoff originate from nonpoint sources
(NPS). These NPS materials are transported by water generated as a stochastic
(random, but characterized by time of the year) hydro!ogic event, are affected
significantly by temporal and spatial factors that have preceded the hydro-
147
-------
logic event, and impact aquatic ecosystems as discrete, largely receiving-
system-independent input. Urban runoff impacts are similar to other aquatic
ecosystem impacts with respect to the relationships between activities and
materials produced, the materials themselves, and interactions within the
aquatic ecosystem that are not a function of the input. Control of urban
runoff occurs in the watershed but ecosystem impacts are the basis for assess-
ing the need for such controls.
The system watershed can be visualized as consisting of separate but
somewhat interrelated phases or processes: 1) the production of waterborne
materials in the watershed which can enter the aquatic ecosystem, 2) the
hydro!ogic event, 3) the transport system, and 4) the aqueous environment
and its community (Figure 1). Urban.runoff impacts on the aquatic ecosystem
are characterized by large magnitude, random inputs of materials that affect
the ecosystem. Thus, they are similar to "spill" events and other "cata-
strophic" types of events.
Most studies of urban runoff are concerned with the first three phases
of the stream ecosystem and with water quality in the receiving environment.
In this paper it is postulated that the urban runoff impact on the stream
community is different from other types of inputs and that, therefore, the
community should be evaluated differently. Although methods of study would
be similar, interpretation of timing effects is different. The need to
regulate or treat urban runoff depends on accurate assessment of impacts on
the aquatic community. Without this assessment, costly and complex NFS con-
trols may not be justified. Only stream ecosystems are discussed in this
paper. Lakes generally serve as at least partial sinks for input materials.
Lakes may not differentiate between urban NFS loadings and continuous load-
ings such as wastewaters, while streams are impacted and then have an oppor-
tunity to recover as they are flushed by upstream water.
CASE HISTORIES OF RANDOM EVENT AND
MAGNITUDE IMPULSE IMPACTS ON STREAMS
The selection of case studies was not intended to be exhaustive nor
restricted to urban runoff events. Generally, the approach used was to
locate stream impact analyses and use those which illustrate the following
assumptions about urban runoff:
• Urban runoff often alters stream flow patterns.
• Pollutant inputs exert impacts related to accumulation
by the biota which in turn is based on pollutant
concentration.
• Pollutant concentrations vary extensively and
relatively rapidly because of the mode and timing
of runoff input.
Justification for this approach is the lack of published studies relat-
ing urban runoff impacts to stream communities and the very few related to
impulse iihputs to stream ecosystems. Approximately 1,200 reports and papers
on urban runoff have been critically evaluated (Porcella and Sorensen, 1979).
148
-------
1. Watershed Phase
Watershed Activities
Land Uses
Watershed
Characteristics
Geology
Geography
Climate
Random hydrologic
event that produces
adequate runoff to
reach aquatic receiving
system. The variables
include temporal and
spatial factors such as
time since last storm,
intensity of storm in
time and space
Material load =
FLOW *
CONCENTRATION
2. Hydrologic Phase
3. Transport
Phase
Effects on
ecosystem
variables
f\
4. Ecosystem
Phase
Fate in
stream
7\
In stream
flows
Figure 1. The hierarchial stream/watershed system that produces
material flow to aquatic ecosystems.
149
-------
It was estimated that 2% related to ecolocrical impacts, 40% concerned model -
inq, hvdroloqy and hydraulics, 35% concerned urban runoff measurements and
23% applied to water quality measurements in the stream.
Althouqh BOD/DO models have been used to assess urban runoff and other
NPS impacts on streams (Rickert et al_., 1975), these models do not reflect
response of higher organisms. Until stream ecological models can be de-
veloped to assess nonpoint source impacts on the stream community, simpler
methods comparing upstream/downstream, before/after, and between similar
environment community variables must be used. As a part of the review, use-
ful target variables were identified that would reflect stream biotic re-
sponse to pollutant inputs (Table 1).
Table 1. Variables identified with properties that describe the
state or condition of ecosystems
Properties
Variables* (measured per standard unit area)
I. Environmental
II. Structure
Biomass
Energy
Materials
Volume, pressure, abiotic mass flow rates,
system attributes (homeo stasis-feedback,
storage, loss), physical attributes (top-
ography, geology, . . .), habitat.
Diversity, number, mass, chlorophyll-a_
Calories
C, N, P, ..., total mass
III. Function
Biomass
Energy
Materials
Change in diversity, number, weight,
competition, resilience, resistance
Energy transfer, efficiency, P/R ratios
Elemental turnover times, P/R ratios,
limiting nutrient
*Can be applied at any or all levels: community, trophic, population,
organism, dominance, keystone, guild (functional groups) or habitat.
150
-------
FLOODING EFFECTS ON FISH COMMUNITIES
Harrell (1978) analyzed effects of flooding on fish diversity for Devil's
River in Texas. A fortuitous flood event (ninth largest on record) occurred
midway during Harrell's study to assess the fish community of the river.
Marked changes in habitat occurred as a result of the flooding; distinctness
of spring, riffle, pool, channel and intermediate habitats generally were
blurred and riffle-like habitat-types were more frequent. Fish community
diversity declined. Barrel! suggests that the ecological plasticity of the
communities allowed shifts in the roles and associations of fish species and
thus allowed the community to be maintained. In this respect, the community
is stable; naturally stressful environments (Slobodkin and Sanders, 1969;
Orians, 1975) would result in dominance and maintenance of a few species.
EFFECTS OF TOXICANT SPILLS
Although there are many spill impact studies on marine ecosystems, few
similar types of studies have been done in freshwater systems. Spill impacts
are random impulse type events and therefore are analogous to urban NFS run-
off impacts on streams, although they occur less frequently.
Cairns and Dickson (1977) describe the Clinch River (Virginia) as impacted
in this manner. On June 10, 1967, caustic wastewater (pH = 12) from a fly
ash pond associated with a power plant adjacent to the Clinch River were
released through a collapsed dike and caused extensive damage to the aquatic
community. The pond inflow was 40 percent of the normal Clinch River flow
at that time (Cairns et aJL, 1970). In 1971 an acid spill occurred. Mean-
while during this period, day-to-day operations of the power plant and
periodic flooding had short term effects on the aquatic biota. The authors •
and co-workers assessed benthic macroinvertebrates at 21 ecologically sim-
ilar stations during the years 1969, 1970, and 1971. The data analysis .was
based on numbers, biomass, diversity, and presence/absence (cluster analysis).
Although approximately 200,000 fish were killed by the spill, Cairns and
co-workers focused on benthic invertebrates because they
1) are relatively sessile and cannot avoid stress;
2) have long and complex life cycles;
3) are important members of the food web and affect related
organisms;
4) are sampled by techniques that are more reliable than for
other organisms;
5) yield more biological information obtained per dollar
invested.
151
-------
The initial measurements (June 20, 1967) were made by State of
Virginia employees and Cairns and associates analyzed these data to show
that the number of taxa and organisms per unit area were severely impacted by
the spill (Table 2).
The authors noted that the benthic community was 8 percent midges at
Station 1 but increased to more than 60 percent at damaged stations, illus-
trating tolerance to stress. By Station 10, the midges were less than 30
percent. The pattern at these lower stations was that of lesser impact (due
to dilution and physico-chemical reactions).
Recovery was not rapid and was still ongoing for the July-December, 1969
sampling (Table 3). The biotic density was considerably depressed but diver-
sity had reached control levels. Unfortunately no data on diversity were
presented for the 1967 data so it was not possible to analyze diversity as a
function of time (e.g., Cornell eial_., 1976). The longer life cycle and
non-motile organisms had not returned by 1969; thus recovery is a function
of recolonization time which is controlled by life cycle and motility. Re-
covery was essentially complete by 1971.
STREAM COMPARISONS - MICHIGAN
Ball el; al_. (1973) studied three streams in Michigan to show pollutant
impacts on stream communities: 1) the Jordan River was least developed with
the downstream sitd differing from the upstream site because of a fish
hatchery effluent; 2) the Au Sable River was affected chiefly by recreational
use but also by a previously thriving lumber industry and forest fires; 3)
the Red Cedar River was in a more urban/industrial region and, although
having received waste effluents directly in past years, urban runoff and
some industrial inputs were more common. Generally water quality at up-
stream stations was better than downstream.
The authors and their colleagues had performed a variety of measure-
ments to compare with material inputs to the rivers. Although they con-
cluded that the rivers were geochemically similar, the slope of the Red
Cedar River was steeper. Generally, total organic carbon, non-carbonate
hardness and total phosphorus were greater at the lower station of the Red
Cedar River than the other stations. Suspended matter was considerably
greater for both stations on the Red Cedar River than for the other rivers.
Generally toxicants (metals, pesticides and PCB's), inorganic nitrogen com-
pounds and chloride were greater on the Red Cedar. Although data on macro-
phytes and fish were discussed, those data were not directly comparable to
other measures because the same species were not always present at all the
stations.
The most important variables of comparison are listed in Tables 4 and
5. The Au Sable River seemed most productive based on diel DO fluctuations
although the gross productivity of the Red Cedar River was greater. A com-
parison of average P/R ratios indicated that the better quality rivers
(Jordan and Au Sable) had fairly typical values for clean midwestern rivers
152
-------
Table 2. Ash pond spill (June 10, 1967) effects on Clinch River (Virginia)
benthic invertebrates (adapted from Cairns et al_., 1970)
Station
No.*
1
2
3
4
5
6
7
8
9
10
11
12
Description
Upstream (reference)
Upstream but backwater
effects
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
RM
+21.0
+ 0.5
- 0.3
3.5
9.5
12.5
17.2
30.5
40.2
56.5
65.5
77.5
Taxa
24
9
0
5
11
15
8
11
22
18
22
32
Total 2
Benthos/m
5,600
320
0
2,900
3,060
7,000
6,900
1,900
8,100
860
2,800
11,500
*Sampled June 20, 1967; RM = river mile from spill.
153
-------
tO
E
en
s-
•i —
^
t—
•r-
Q.
I/)
•o
1
in
fO
CU
,dz
•*->
t^— ,
o
(/)
]
> 0} \
•p—
rv^ t/t
c.
o •«-
C res
•r- O
0 E
p
O H—
5- 0)
O) -{->
> 0-
O fl3
U T3
cu (0
ce: - —
•
co
cu
is
^
>,CM
•r^ ^3)
0,0
(U "~
Q re
CM
e
ns t/)
•*-> 0
0 J=
I— 4->
c
cu
CO
to
X
tO
H-
§
c^
o
Q.
•r—
s-
o
to
cu
o
J-
o *
•r— •
•t-5 O
CO ~"^-
to
CO r-f LO ,-1 CO 1^.
O CM CO O O •— 1
,.,_ ^. ,_).,_,. ^ ,_,.
1 1 1 1 1 1
[**•• CM K^t" CO CTi i~* S
cy> cri CD co «-i co
CM CM CO CM CO CO
o
CO
t— 1 O CD CO CD O
1 i— 1 CM <3- CO O
O LO CM
CO
CM
«^t-
LO
1 CO CO CD CM CO
co >3- co co -^ •vt-
«^J-
LO LO CM CT> CTl LO LO
• ^5 ••••»•
t— i 4-> co o o co en co
CM i-l
^ — ^
r-
O T-
<^ (~>
eC to
«^J-
i i^. co en o >— i
r— 1 t— 1 I— 1
•
r-«
r—
Q.
to
E
o
S-
(I
T~"
cu
•r~
E
S-
cu
^>
•1—
II
^>"
a:
• A
en
co
I— 1
n
S-
cu
0
cu
f~"l
1
>>
r—
•-D
-a
cu
'a.
E
to
oo
154
-------
E
to
CO
•r~*
u
•r-
s:
E
•r—
CO
S-
OJ
•r—
•P
E
(D
S_
cu
•r—
T3
S-
o
q-
co
CO
1 —
JO
ro
(O CO
> r**N
CD
^s <~~
-p
i — •
(O p™™
H3 fO
CT*
-PI
E O)
to
CU r—
S- r—
-P (0
to OQ
tl , ^^
O O
S-
E q-
o
to -a
•t- CD
S- -P
to Q.
a. re
E "0
O to
CJ - -
•
c^J-
cu
r~i
to
1—
i,
(O
S- T3
O) CO
3 O
O
—I T3
CU
Ct:
s-
tO
s- -a
a> co
Q-CJ
a. '
ZD T3
01
ry
CD
S- JD
0) ro
S 00
O
—I 3
0)
r—
s- .a
CD (O
Q.OO
a.
~~i ^3
•=c
c~
S- to
cu -o
2 S-
o o
— J •-D
c~
S- tO
CU TD
Q. i-
Q. 0
ZD ""D
CO
CU
o
to
., —
J_
rO
CM O «5j- CD
• • • •
LO CT> LO O
i—
l^~
LO LO cn o
• • • •
LO O i — CO
1 — LO
0 0 CM 0
• • • a
cn (^o CM r--
r— CM
LO o r-~. o
• • • •
•=d- CM i — r^
r— CM
LO O LO O
• • • •
CM CO LO UD
LO Ln co o
• c • •
CM Cn i — CO
^-~
, — »
to
cu
fO
4_>
co
JD
3
co
, ^
c>^ ~^f r—
\ CO fO
CD > CJ -i—
E S- <=? CU (J
CU ~~^ to •!—
•> to CD ~-^ q-
CU -O £r CO -P *r-
O D " E E -P
E 0) S-
CO E E " O to
S- 3 3 S S-
CU E E O CO
q- -r- •!-- r— Q. to
q- x x u_ E
•i- ro ro •> O
-a E E E > -P
ro O to
O O CU T-
Q Q s: Q
CM O
o cn r^ o o LO
• • • • • • •
O i — i — P"» Cn O
•si- CM
«^J*
r—
O h~. O co
• • • •
CD r— 1 CD i—
0 i—
1 — '
LO LO
o co r-~ o o cn
• • • • • •
O i — CM i — CM O
CO CO r—
co
o r~~ o cn
i
Of— (~^ 0
r^ i—
LO r~^
O r— CO O O CO
• * * * • . •
o CM co co r-~ o
CO r—
CM
-CD CO
CD i — CO CD O CO
t • • • • •
O CM -=d- LO CO O
CM
i —
>^
(O
-a >>
+j
CM -i- >,
E to to
O i- -O
-•-, cu
CJ > ^-~~
•r— -E tO
CD Q -P O
-a E -a T-
CU >> to T- -P
> « +->^-^ CU -P S to
S- E T- CM to -r- E — ' Q;
cu o to cu o cu
tO-r- S-CD -PCUO E C£
JD.-P cuo roa.s-\ ^.
OO >r— S-OOQJ CJ D.
3 T— JD O. O)
E -o o " cu o cu
3O = -P T- •> " CT
ES- toa: s-j^^> -P to
•r-D- 3 — CU-PO q- s-
X E > E T- CU
rO CO E CU S- >
S! CD i— i CO Q eC
•
c™
O
• r—
•P
ro
•r—
^.
ro
>
q-
o
•p
E
CO
•r" •
O
•r-
q-
q-
0)
o
o
u
^>
CJ
155
-------
•
^— *,
CO
2
t\
•
•Pi
tO
CQ
£
O
q-
"O
CD
•p
o.
to
-o
(0
10
S-
O)
•J_
S-
CD
O)
-C
•p
o>
-p
S-
o
-p
CD
cn
•o
3
^^
cn
(U
c
CD
C
•
LO
o>
f>
(0
f—
S-
(O
5- T3
CD CD
SO
O
CD
Q:
s-
10
S-T3
CD O)
O.O
Q.
O)
a:
(U
1 1 i-
s- .0
CD ro
3 oo
o
_1 3
(1)
r—
S- J3
CD IO
O.OO
a.
ZD 3
£=
i- 10
aj-a
3 s-
0 0
£=
t— (&
*
cn
s-
O)
c:
UI
CO
o
*
0
LO
1 o
•
o
LO
0
0
oo
0
o
00
0
o
oo
CD
o
^
o
•r—
•p
o
3
•0
o
S-
Q-
sz
0
>.
JZ
Q.
"il
(U
D-
f— -
o
•
o
1
cn
CM
r
0
1
p_
0
o
CO
CD
CD
c
o
'•p
o
3
ID
o
S-
0-
CD
>>
r"
CL
O
i.
0
to
s:
cn
0
•
O
*
o
•
0
LO
oo
•
o
*
co
o
•
o
cn
o
•
0
*«Q
0
o
-o
to
C"
O
>J
c- a)
p> -P
•r— >>
i- <"*
0) Q.
Q. O
S- 0
O IO
s:
E
3
oo
CD
OO
CM
0
0
*
^f
O
LO
•
OJ
o
CM
•
r—
O
CM
•
^—
0
*
0
to
•j—
tO
cu
c™
4_>
C
C/)
Q
+J
0
JS
a.
to
to
o
S-
U3
0
0
•
LO
0
•
CO
o
•
CM
0
CO
•
p-~
O
{&
•
r—
0
r-.
•
0
c*
o
•r-
•P
to
i-
•r-
Q.
to
CD
rv^
>^
4->
•r—
C
3
^
F
O
O
LO CD
C*3 O
CM LO
n
cn
00
r- o
cn o
r — cn
! *»
LO
P~~
00 O
CO O
CM cn
«\
o
1 —
r^- o
LO O
CM r->
*
CD
P—
CM 0
LO O
r~- CM
*v
CO
1— •
oo o
CM O
oo cn
n
CO
o
•p—
c
to
cn
* £-
* O
P
3 a>
Q. -P
£•• tO
1 — 1 r—
3
r— O "^
n3 -i- *
•r- -P S-
s- s- a;
•P 10 -P
CO Q- P
CD (O
S- 0) S
S- S=
a> •!-
H- U-
'CD
0,
*» ''.
LO
•
CM
o
o
00
LO
LO
0
o
CM
•>
CM
oo
o
o
oo
n
CM
CO
0
o
CO
A
CM
CM
O
O
CM
r,
OO
*
*
t.
(U
-p
-p
IO
2:
o
•I—
C"
to
cn
S-
0
•a
cu
>
r—
O
LO
to
•p—
Q
'•*
\
•
•o
OJ
•P
to
e
•r™
•p
to
a>
•P
o
c
to
rt)
3
c:
o
•r~
+J
O
3
"O
O
^
CL
a>
P
^^
jj^
a.
o •
s- >>
O IO
to ^3
E
(U '
LO E
3 ~~x
to O
o cn
QJ
J3 «
to
3 *p
O 3
i— 0.
CO T-
•r-
r—
O) IO
3 _J
•X *
*
156
-------
while the Red Cedar River was the most variable and had values outside of
the normal range of P/R ratios. This may have been associated with the con-
siderable input of organic matter in the lower Red Cedar, possibly from
urban runoff.
Community structure as measured by diversity varied considerably. Al-
though diatom diversity was varied considerably. Although diatom diversity
was greater for the lower Red Cedar River than the Au Sable, and for inver-
tebrates the lower Au Sable was greater than the lower Red Cedar, both sta-
tions on the Jordan were considerably greater than those of both the other
rivers. Also, the error associated with the estimated diversity for the Au
Sable and the Red Cedar Rivers was greater than for the Jordan. This obser-
vation relates closely to the theoretical development of diversity change as
proposed by Cornell ejt a]_. (1976). Based on these data (Tables 4 and 5) one
could conclude that all of the variables, DO, diversity (periphyton, inver-
tebrates), and productivity could be used to assess urban NFS runoff impacts
in freshwater streams.
NPA RUNOFF IMPACTS ON STREAM COMMUNITIES
Chisholm and Downs (1978) studied the benthic invertebrate community
of a small West Virginia stream which received sediment from construction
of a super-highway. In comparison with a nearby control stream, the diver-
sity index, generic count, and total count of invertebrates indicated severe
reduction or destruction of the benthos in the impacted stream. The greatest
degradation occurred in areas of highest sediment movement. When construc-
tion ceased, the benthic population of the impacted stream recovered to a
comparable level with the control stream. Colonization was from upstream,
unimpacted tributaries.
Except in streams subject to combined sewer and sanitary overflow dis-
charges, Ragan and Dieteman (1975) found streams in areas of Maryland,
urbanized during and after the 1950's, to be of good quality. This was not
expected but the authors attributed this to the lack of sanitary discharges
in the area, and to the large (>10 square miles) size of the drainage basins
being studied. Even though traditional water quality parameters did not
show dramatic changes, pollution sensitive fish species (e.g., Rosyside
Dace) were no longer found in urbanized areas. A nearby unurbanized stream
had the same 21 species of fish that were present in 1912.
DiGiano et al_. (1975, 1976) directed their study in Greenfield, Massa-
chusetts, toward both short-term impacts on water quality and the longer
term disruptions of the benthic macroinvertebrate population caused by urban
runoff. Using weekly grab samples from the Green River, they found that
total P, turbidity, chloride, and total col iform concentrations increased
with runoff events. They were unable to show conclusively that BOD, TOC,
and oil and grease concentrations increased with runoff. It was suggested
that more detailed analyses of both discharge and concentration be used in
determining the mass loading for relatively short (3.2 mile) river reaches.
In an analysis of three specific storm events sampled at 15 or 30 minute
157
-------
intervals, they found total pollutant loading variations of 50 to 100 times
from storm to storm. Distinguishing between combined sewer overflows and
stormwater runoff events is also important, particularly for bacteriological
studies.
The benthic diversity (Brillouin's) of the stream macroinvertebrate
community remained constant upstream. However, downstream in the urbanized
areas of the stream, reduced diversity and a community shift to more pollu-
tion tolerant (polysaprobic) species occurred. In addition, the seasonal
variation in diversity increased at downstream stations. The physical
character of the stream bottom and availability of substrate deteriorated
due to sedimentation. Sedimentation did not explain all of the changes in
benthos. Evaluating the effect of contaminants from natural and man-made
sources, bioaccumulation of metals through the detritus-based food chain,
and stress from these heavy metals on benthic invertebrates required further
investigation to determine cause and effect relationships.
CASE HISTORY SUMMARY
The major conclusions from these case histories are:
• macro invertebrates and periphyton seem to be
effective for monitoring impulse impacts on streams;
• diversity and biomass are often used for the analysis
of monitoring data but they are not always adequately
sensitive;
• water quality variables do not always reflect stream
impacts directly. The biota must be assessed by
measuring the quantitative effects of impulse inputs
and contaminants.
A scenario based on the cited case histories and other knowledge of the
system illustrates the variables (hydrologic, physical, chemical, biological,
ecological) for determining urban NPS runoff impacts on stream ecosystems.
A storm occurs at a specific time for a specific interval and intensity.
Water begins to run off depending on antecedent storms, soil moisture and
impervious surfaces. The pickup of contaminants varies with antecedent
"washings", velocity of runoff and natural and cultural activities in the
runoff basin. Water containing contaminant materials enters the stream
diffusely, or at a point or several points, and over an interval and pattern
controlled by the transport system and the storm event. This pattern pro-
vides the random event/magnitude impulse of contaminants.
The impulse of contaminants is diluted within the stream and has vari-
able impact depending on flow, peak concentration of contaminants, duration
of exposure and life cycle period of the organisms making up the community.
Survival of specific organisms/taxa/niche is controlled by such events.
158
-------
Eliminated community components can be replaced by recolonization but the
interval depends on out-of-stream or upstream sources and physiological
variables (length of reproductive cycle, motility). Continual contaminant
inputs may prevent recolonization. Where storms are more frequent, inputs
approach the limiting condition of point sources such as wastewater discharges
from domestic and industrial sources.
Other niches may be opened by short-term infrequent inputs but generally
communities become simpler, nutrient cycling speeds up (shorter turnover
times), and ecological efficiency decreases. The variables that seem most
useful in reflecting these processes are not necessarily biomass and diversity
but include those processes associated with community functions; energy flow
and material cycling should also be measured to determine their value as an
assessment tool.
A PROBABILITY HYPOTHESIS
FOR ECOLOGICAL RESPONSE VARIABLES
Throughout this paper, it has been asserted that random event and mag-
nitude impulses to stream ecosystems would result in changes in ecological
variables related to community structure and function. In this section, an
approach for assessing such impacts is proposed. Because the spatial dis-
tribution (Slocomb and Dickson, 1978) and the probability distribution of
runoff events can be determined, a stochastic approach is suggested. Some
application of stochastic models to water quality data illustrate the
feasibility of this approach.
Krishnan ejt aj_. (1974), Padgett ejt al_. (1977), and Pamanabhan and
Delleur (1978) applied this approach to BOD modeling in streams and thus
were able to calculate the mean DO in a stream, the error about the mean,
and the probability of violating a specific criterion or standard. In all
three reports, data generated from storm runoff models were used because
actual data were insufficient. Temperature modeling has been used to assess
the effects of catastrophic events such as climatic and flow variations
which result in extreme water temperatures. The model provided a reasonable
duplication of actual data (Morse, 1978). It is simple to see how such
events could result in large-scale community changes but it is difficult to
assess the magnitude of the ecological response and even more difficult at
this time to develop the relationships between the impulse and the response.
For example, Slatkin (1978) used a model study to show that community
stability existed only when the time scale for environmental change is
roughly comparable to .the average response time due to those changes. Means
and variances did not determine extinction tendencies.
Material (nutrient, toxicant) and energy (allochthonous reduced carbon,
light, temperature) inputs to streams as impulse loadings vary seasonally.
Such impulses are exemplified for a hypothetical seasonal cycle for a fish in
Figure 2. Two sensitive stages are illustrated: 1) egg growth and larval
development requires dissolved oxygen (the DO is utilized in degrading
organic matter in urban NPS runoff); 2) fry development requires lower
159
-------
r
5.1
_ X)
CO
D-
i-
3
CD
O
.i a>
O
c
o>
o.
i Q
o>
c
'c
o
a.
to
.
O)
o
ui
o
< <
5 UJ
2
3
Q.
E
O
•a
-P
o
QJ
q- •
M- co
O) CD
O) (O
•M -r-
W i-
>> 03
CO >
O
O i—
QJ CO
to O5
O) O
S- •—
4-> O
CO O
(O
o
4-> CO
a> cr>
o -o
O. (C
>> o
CM
O>
3
160
-------
stream velocities to catch food and maintain safety from predation by large
fish. If impulse loadings drive DO and flow variables outside acceptable
ranges for a sufficient time and magnitude, a year class could be eliminated.
Continued annual cycles where violations occur could eliminate the species.
The probability that a species or individual will not survive is equal
to the probability that at least one event will occur that will eliminate
the species or individual. For example, where flow is the event that
threatens the species:
prob (species eliminated) = 1 - prob (Q £ q ) = 1 - Fn (q )
C \{ C*
(1)
in which Q
= flow
q = critical value of flow above which
species is eliminated
FQ(q ) = cumulative probability distribution
^ of flow evaluated at q
\f
However, flow has a deterministic seasonal component and therefore, FQ and
low
Q
can be written as functions of time. Also, the critical value of flow varies
with the stage of development of the species and is also a function of time.
Therefore, Equation 1 can be rewritten as follows:
prob (species eliminated) = 1 - prob (Q(t) £ (q (t))
= 1 - FQ (qc(t); t)
Similar relationships can be derived for other response variables. This
approach is only a flow magnitude event, but this basic procedure can
be extended to the bivariate case of a flow magnitude-duration event system.
With the relationship developed above, survival versus probability curves
could be obtained that relate to inputs having a certain probability
(Figure 3). Similarly, the approach could be extended to variables that
represent ecosystem processes, cycling and energy flow.
The effects of a certain urban NFS runoff input on a specific ecologi-
cal variable would not be a single number but would vary within the proba-
.bility of response depending on the space wherein the impact occurs
(Figure 4)).
Because response is concentration related and high concentrations have
a low probability of occurring, the probability of a species surviving or not
surviving may be difficult to assess by modeling and calculation. Microeco-
systems could be used to develop a model and then be field tested. One means
of testing this approach is to carefully measure the pattern of urban NPS in-
puts to a stream, determine the immediate response, relate the response to
average (peak, range) concentration, and determine if the community structure
161
-------
clOO
-10
o:
V)
UJ
o
o:
tf
-1.0
10 yr.
Event
1 Recurrence
f Interval for
Imput of Urban
Runoff Carried
Materials.
0.1
TIME , YEARS
Figure 3. Percent survival decreases with more severe
runoff events.
urban
162
-------
UJ
3E
UJ
a.
UJ
_i
CD
<
a:
I
Experimental
Error
INPUT
Space B
INPUT
Space A
Figure 4. There is experimental error for system variables which adds
to the variance caused by response to stochastic inputs.
163
-------
and function varies with time (before/after) or space (upstream/downstream)
as would be expected. The hypothesis for testing is: the community struc-
ture aft
-------
Cornell, H., I.E. Hurd, and V.A. Lotn'ch. 1976. A measure of response
to perturbation used to assess structural change in some polluted and
unpolluted stream fish corvnunities. oecoiogia (Beri.) 23:335-342.
DiGiano, F.A., R.A. Coler, R.C. Dahiga, and B.B. Berger. 1975. A
projection of pollutional effects of urban runoff in the Green River,
Massachusetts. In: Urbanization and Water Quality Control. Annual
Symp. of the Amer. Wat. Res. Assoc. Proc. 20. pp. 28-37.
DiGiano, F.A., R.A. Coler, R. Dahiga, and B.B. Berger. 1976. characteri-
zation of urban runoff - Greenfield, Massachusetts. Phase II. Wat.
Res. Res. Cent., Univ. of Mass. Amherst, MA. 137 p.
Barrel!, H.L. 1978. Response of the Devil's River (Texas) fish community
to flooding, copia. 1:60-68.
Krishnan, K., P. Radha, J.J. Lizcano, I.E. Erickson, and L.T. Fan. 1974.
Evaluation of methods for estimating stream water quality parameters
in a transient model for stochastic data. wat. Res. Bull. 10:899-913.
Morse, W.L. 1978. The dishonest method in stream temperature modeling.
Wat. Res. Res. 14:45-51.
Orians, G.H. 1975. Diversity, stability and maturity in natural ecosystems.
\ unifying concepts in Ecology. W.H. Van Dobben and R.H. Lowe-McConnell,
Eds. pp. 139-150.
Padgett, W.J., G. Schultz, and C.P. Tsokos. 1977. A stochastic model for
BOD and DO in streams when pollutants are discharged over a continuous
Stretch. 'Intern. J. Environ. Studies. 11:45-55.
Padmanabhan* G. and J.W. Delleur. 1978. Statistical and stochastic analyses
of synthetically generated urban drainage quantity and quality data.
Tech. Rept. No. 108. Wat. Res. Cen., Purdue Univ. West Lafayette, IN.
Ill p.
Porcella, D.B. and D.L. Sorensen. 1979. Characteristics of nonpoint source
urban runoff and its effects on stream ecosystems. Report to U.S. EPA,
Con/all is, OR 97330. 99.p.
Ragan, R.M. and A.J. Dieteman. 1975. Impact of urban stormwater runoff on
Stream quality. In: Urbanization and Water Quality Control, Annual
Symposium of the'Amer. Wat. Res. Assoc., Proc. 20. pp. 55-61.
Rickert, D.A., W.G. Hines, and S.W. McKenzie. 1975. Planning implication
of dissolved oxygen depletion in the Willamette River, Oregon, in:
Urbanization and Water Quality Control, Annual Symposium of the Amer.
Wat. Res. Assoc. Proc. 20. pp. 70-84.
165
-------
Simberloff, D. 1978. Use of rarefaction and related methods in ecology.
In: Biological Data in Water Pollution Assessment: Quantitative and
statistical Analyses. K.L. Dickson, J. Cairns, Jr., and R.J. Livingston,
Eds. ASTM. Philadelphia, PA 19103. pp. 150-165.
Slatkin, M. 1978. The dynamics of a population in a Markovian environment.
Ecol. 59:249-256.
Slobodkin, L.B. and H.L. Sanders. 1969. On the contribution of environmental
predictability to species diversity. Brookhaven symp. Bioi. 22:82-95.
Slocomb, J. and K.L. Dickson. 1978. Estimating the total number of species
in a biological community. In: Biological Data in Water Pollution
Assessment: Quantitative and Statistical Analyses. K.L. Dickson,
0. Cairns, Jr., and R.J. Livingston, Eds. STP 652. ASTM, Philadelphia,
PA 19103. pp. 38-52.
166
-------
Fourth Session
IMPACTS ON RIVERS
Moderator: Robert Shubinsky
Water Resources Engineers
Springfield, Virginia
167
-------
ANALYSIS OF RECEIVING STREAM IMPACTS
ON THE MILWAUKEE RIVER
Thomas L. Meinholz
EcolSciences, Inc.
Milwaukee, Wisconsin
ABSTRACT
The Milwaudee Metropolitan area contains approximately 15,000
acres of combined sewers which discharge to the three rivers of
the area or directly to Lake Michigan. As part of the Facilities
Planning efforts of the Milwaukee Metropolitan Sewerage District
and research for EPA's Storm and Combined Sewer Section, extensive
monitoring and modeling efforts were performed to quantify the re-
ceiving water impacts of the combined sewers.
Over 100 overflow points within the combined system were
modeled using the EPA SWMM model and Corpos of Engineers STORM
model. The output of these models was used to load Harper's re-
ceiving water model. Major difficulties in modeling the river
systems were the variable influence of Lake Michigan in the lower
reaches. The relatively clean, high DO and lower temperature lake
inflows could not be easily quantified because of the dependence
of the inflows on wind speed, direction, lake level and other
variable factors. Final model calibration using a linearly de-
creasing flux was found to match the continuous DO data generated
during two years of record.
The Milwaukee river exhibits tremendous DO sags in the lower
reaches of the river following runoff events. The modeling tasks
of the CSO project could not duplicate these sags using the in-
stream concentrations found in the monitoring program. Extensive
field monitoring was then conducted to quantify the source and
mechanism of these sags. After numerous investigations, the
bottom sediments in the lower reaches were found to be the source
of the rapid DO declines. The mechanism was related to the scour-
ing action of submerged combined sewer outfalls.
In order to model the response of the river to discharge
events, the receiving water model was modified to include an ex-
pression which would predict the extent and duration of the scour
action from the submerged outfalls. Long term simulations of DO
and other parameters were calibrated and verified using the re-
sponse of the river to a multitude of rainfall events.
The use of this model network in the evaluation of alterna-
tives for abating combined sewer overflows produced magnitudes of
DO and fecal coliform impacts for each alternative using 20 years
of rainfall record. The results for the following alternatives
are presented.
168
-------
* existing conditions
* partial separation
* complete separation
* out-of-basin (storage-conveyance-treatment)
* en'd*of-pipe (EOF)
* 100% CSO removal
Discussions of how these results were quantified including
the model development will be discussed. The cost-benefit analy-
ses that were used to satisfy the EPA PG-61 requirements will
also be detailed to provide the reader with a methodology that
has been successfully applied and approved in the Milwaukee area,
STUDY AREA
The Milwaukee area contains approximately 15,000 acres (6060
hectares) of combined sewers that discharge to the three rivers
of the area which flow into Lake Michigan as shown in Figure 1.
In 1975 the Milwaukee Metropolitan Sewerage District initiated
a project to identify alternatives to abate combined sewer over-
flow (CSO) and select the alternative which was most cost-
effective in terms of the stated water quality goals for the
area. The monitoring and modeling investigations of this CSO
project identified significant receiving water impacts as a re-
sult of wet weather discharges.
The preliminary findings of the CSO project led EPA to select
Milwaukee as one site for another project to quantify the mag-
nitude of the receiving water impacts associated with CSO. The
Milwaukee area was selected as one of the sites for the impact
evaluation project because of its large combined sewer area, the
availability of previous and ongoing projects relating to CSO, and
the significant water quality impacts which occur as a result of
wet weather discharges. Each of the rivers has an associated
combined and storm sewer area which discharges during wet weather.
For purposes of this study, the investigations into the water
quality impacts were limited to the Milwaukee River.
The most dominant feature of the river is the inflow from Lake
Michigan in to the lower reaches of the river which influences
flow conditions as far upstream as the North Avenue Dam (Figure
2). The inflow of lake water results in the slowing or reversal
of flow in these reaches throughout the year. The lake influence
has been monitored at a few downstream locations with a sensitive
current speed and direction meter. The findings of these surveys
have shown the river to be flowing upstream at one depth in the
water column and downstream at another depth with a complete re-
versal of this trend a few moments later. Attempts to relate
the inflows to wind speed, river flow and lake level have proved
fruitless.
169
-------
0 4000 8000
(0) (1200) (2.400)
SCALE, feet (meters)
LINCOLN
CREEK ,
MILWAUKEE
LINCOLN R1VER
CREEK
LAKE MICHIGAN
KINNICKINT
RIVER
LAKE MICHIGAN'*'
Figure 1. Illustration of the CSO drainage areas contributing
to the various receiving waters in Milwaukee, Wisconsin.
170
-------
£ USGS FLOW GAUGING STATION
A RAIN GAUGE
• CONTINUOUS DO MONITOR
MENOMONEE RIVER
PORT WASHINGTON ROAD
ESTABROOK PARK
CAPITOL D
MILWAUKEE RIVER
HOLTOM STREET -
WALNUT STREET
CHERRY STREET
WELLS STREET
ST. PAUL AVE.
BARTLETT AVENUE
NORTH AVENUE DAM
LAKE MICHIGAN
BROADWAY STREET
Figure 2. Monitoring locations in the study area of the Milwaukee River.
171
-------
The mean, annual discharge for the river as measured by the USGS
gauge at Estabrook Park is 400 cfs (11.3 m^/sec). During the
coldest portion of the winter months, the river is ice covered
in both the lower and upper reaches. The CSO area that dis-
charges to the river extends along both sides of the river with
the first outfall located at Capitol Drive as shown in Figure 2.
A majority of the outfalls in the lower portions of the river
are completely or partially submerged. Typical cross-sections of
the river at a downstream location and a photograph of the lower
reach site at St. Paul Ave. is shown in Figure 3.
Combined Sewer Area
The combined sewer area that contributes to the Milwaukee River
is approximately 6000 ac (2428 ha) with a total of 52 outfalls
ranging in size from 12 in. (30.5 cm) to a double 10 by 7.5 ft
(3 by 2.3m) box outfall. The drainage areas for individual
outfalls range in size from 5 to 702 ac (2 to 284 ha). The en-
tire combined sewer area has been modeled in a previous project
using the EPA Storm Water Management Model (SWMM) and the Army
Corps of Engineers Storage Treatment Overflow Runoff Model (STORM)
(II). The land use percentages of the CSO area from the model
data are as follows:
Land use
Single family residential
Multi family residential
Commercial
Industrial
Parkland
Storm Sewer Area
Percent of CSO area
57.6
12.6
17.8
11.3
0.7
Those areas within Milwaukee County that drain to the Milwaukee
River and are served by storm sewers comprise approximately
27,000 ac (10,900 ha) of drainage area. These'areas contain the
following land use breakdown:
Land use
Single family residential
Multi family residential
Commercial
Industrial
Park - open area
Percent of storm sewer area
50.1
2.7
12.9
4.4
29.9
This area has also been modeled with the SWMM and STORM models.
Figure 4 illustrates a comparison of the storm sewer and CSO areas.
Some of the storm sewers within this area contain cross connections
with the sanitary sewers so that during wet weather the surcharged
sanitary system may be relieved through nearby storm sewers.
172
-------
WEST
EAST
« o (o)
^_e
4,10 (3)
(U
a)
H-
= 20 (6)
i-
o_
LU
0
(o)
TOO
(30)
200
(60)
WIDTH, feet (meters)
Figure 3. Photograph and cross section of the Milwaukee River
at St. Paul Avenue
173
-------
1
MILES
km = mi x 0.62
OZAUKEE COUNTY
MILWAUKEE COUNTY
GREENFIELD AVENUE
Figure 4 Illustration of CSO and storm sewer areas tributary
to the Milwaukee River.
174
-------
Upstream Areas
The Milwaukee River watershed is comprised of 644 nd.2 (1668
of drainage area that lies to the north of the Milwaukee metro
area. The predominant land use within this area is agricultural
in nature which comprises approximately 66 percent of the up-
stream watershed. Urban land use accounts for 1.5 percent while
open areas and woodlands make up the remaining area. In 1970,
an estimated 530,000 persons resided in the entire watershed in-
cluding the area within Milwaukee County.
FIELD INVESTIGATIONS
The water quality conditions within the Milwaukee River exhibit
extreme variations between wet and dry weather periods. Figure 5
presents data from a continuous DO monitoring device located at
St. Paul Avenue showing the variation in DO. Figure 6 presents
fecal coliform data from the same site during two intensive
monitoring surveys. The dramatic loss of dissolved oxygen and
rapid increases in fecal coliform concentrations were investi-
gated in the field monitoring portions of the projects. These
consisted of comprehensive field surveys, the analysis of two
years of data taken from three continuous DO and temperature
monitors, dye studies to define the rivers hydraulic and mixing
characteristics, and sediment investigations.
The term water quality impact for this study was limited to the
violations of the DO and fecal coliform standards that have been
set by the Wisconsin Department of Natural Resources for the
Milwaukee River. The standards apply for both dry and wet weather
flow conditions in the river and are as follows:
Upstream of North Avenue Dam
1. Preservation and enhancement of fish and other aquatic
life. The DO content shall not be lowered to less than
5 mg/1 at any time.
2. Full body contact recreational use; the membrane filter
fecal coliform count (MFFCC) shall not exceed 200 per
100 ml as a geometric mean based on not less .than five
samples per month.
Downstream of North Avenue Dam
1. Marginal conditions for fish and aquatic life. The DO
shall not be lowered to less than 2 mg/1 at any time.
2. Partial body contact recreational use; the MFFCC shall
not exceed 1000 per 100 ml as a geometric mean based on
not less than five samples per month.
175
-------
RAINFALL
0)
o
o 3
in ->
CD
in
Q
0.36 in. (0.91 cm)
DRY WEATHER
1200 2400 1200
TIME, hours
2400
Figure 5. Illustration of dry weather (July 2S~3\, 1977) and wet weather
(August 3-5, 1977) DO levels at St. Paul Avenue in the Milwaukee River.
o
o
CJ
o
U-
O
o
o
UJ
1200 2400 1200 2400 1200
TIME, hours
Figure 6., Illustration of dry weather (September 22-24, 1976) and wet weather
(August 4-6, 1977) fecal coliforms levels at St. Paul Avenue in the Milwaukee
River.
176
-------
The difference in standards between the upper and lower portions
of the river reflect the physical and biological characteristics
of these two reaches. The upper portions above the dam are free
flowing and not under the influence of Lake Michigan while the
lower portions are those reaches which exhibit some degree of in-
flow from the lake.
Source of Impacts
The data collected in the monitoring portions of both projects
were used in Harper's Multi-Parametric Model to attempt a DO
calibration to determine the source of these impacts. An accept-
able wet weather calibration for DO could not be made, however.
Extensive work in modifying the model provided no better results.
The magnitude and the duration of the predicted sags as well as
the location was consistantly underestimated. At this time the
bottom sediments in the lower reaches of the river were thought
to be the possible source of these sags.
Sediment Oxygen Demand Tests
In order to determine the significance of the sediments on the
dissolved oxygen balance in the Milwaukee River, both ill situ
measurements and laboratory testing was conducted. A respirometer
consisting of a plexiglass chamber with a DO probe attached in-
side and a small 12 v submersible pump attached outside for
circulation of water was used to determine the decline in DO over
time after the chamber was sealed on the sediment surface. The
in situ SOD rate was calculated using the following formula:
SOD = (Ci-Cf)V
tA
Where:
SOD = Sediment uptake rate in gm 02/m2-day
V = Volume of confined water in m3.
A = Bottom area within chamber in m2.
t = Test period in days.
Ci = Initial measured DO in chamber in mg/1.
Cf = Final measured DO in chamber in mg/1.
This formula provides the SOD rate on an areal basis and was con-
verted to a volumetric rate in the river by division with the mean
depth at the sampling location.
The results of the in situ SOD measurements showed the rates to
vary from zero at the upper limits of the study area to an average
of 5.0 gms/m2/day in the lake influenced reaches. The new values
were utilized in the DO sink equations of Harper's model with the
output as follows;
177
-------
1) accurate DO predictions during dry weather
at all stations
2) accurate DO predictions during wet
weather at upstream (non-lake influenced
locations
3) poor prediction of DO magnitudes during
wet weather- at downstream locations.
These findings led the project team to further investigate the
oxygen demand of the bottom materials in the laboratory to
determine the mechanism of the dramatic DO sags.
Samples of the bottom materials were collected and placed in
plexiglass chambers in the laboratory. River water was placed
over the sediments and allowed to settle. After lightly aerating
the river water, the rate of decline of the DO was compared with
the corresponding in situ measurements.
In most cases the laboratory measurements were slightly lower than
the field SOD's which indicated the bottom materials may have been
slightly disturbed when the SOD chamber was lowered into the sedi-
ment. This phenomenon led the project team to investigate the
disturbed demand of these sediments to quantify the change in SOD
magnitude. The results of these investigations are shown in Table
1 and Figure 7. Note that the disturbed demand shows more than a
thousand fold increase in SOD over undisturbed conditions. This
disturbed or scoured demand was thus, identified as the source of
the extensive DO sags.
Mechanism of Impact
The determination of the disturbed SOD as the source of the DO
impact next led to the identification of the mechanism of the im-
pact. Questions such as what causes the sediments to be disturbed
(scoured) needed to be answered so that the receiving stream model
could be properly calibrated. Settling tests of these sediments
indicated the scouring velocity must exceed 0.1 foot per second
(.03 m/sec). A program of velocity monitoring at the sediment
river interface was then undertaken in the river during different
flow conditions. The locations selected for the monitoring were
those that had shown significant DO sags during previous storms.
A sensitive velocity meter was lowered to approximately 1 foot
(.3m) of the bottom of the twenty foot water column. Velocities
were continuously monitored to determine if scouring velocities
were present.
The scour of bottom materials within the river by discharges from
submerged CSO outfalls was verified with the velocity studies and
with visual observations. The velocities measured in the vicinity
of the submerged outfalls were extremely large and capable of
scouring considerable quantities of sediment. The effect of the
178
-------
TABLE 1 . BENCH SCALE DETERMINATIONS OF SEDIMENT OXYGEN DEMAND
UNDER UNDISTURBED AND DISTURBED CONDITIONS1.
Map location
no.
Descri ption
Undisturbed SOD, Disturbed SOD,
gm/m^-day
1
2
3
*
5
6
Kinnickinnic River
at First Street
Mooring Basin
in 1 nner Harbor
Mi Iwaukee River
at RR Bridge
Menomonee River
at Great Lakes Coal
Mi Iwaukee River
at Highland Avenue
Mi Iwaukee River
at Hubbard Park
0.65
1 .^0
2.10
1.70
1 .^0
0.33
^30
1,370
800
270
360
66
The laboratory sediment oxygen demand determinations were carried
out at a temperature of 20 + 0.5 C.
scouring can be illustrated with the rainfall, DO and CSO velo-
city data for a storm occurring on August 3, 1977. Figure 8
depicts the response of the river at St. Paul Avenue in terms of
DO to this rainfall event and the subsequent scouring of sedi-
ments due to CSO discharges. An instream CSO velocity of nearly
12.0 ft/sec (3.7 m/sec) was measured 1.0 ft (0.3 m) above the
sediment surface at a distance of approximately 30 ft (9.1 m)
from a submerged CSO outfall at St. Paul Avenue. This velocity
was measured just after discharges from this CSO outfall began
and the direction of the velocity was perpendicular to the river.
After discharge from the outfall ceased, the river velocity was
variable in direction and very small in magnitude. The rapid de-
cline of DO shown in Figure 8 following the rainfall event is the
result of the scouring or agitation of bottom sediments through-
out the lower Milwaukee River from discharges from submerged CSO
outfalls. This CSO velocity measurement, as well as other mea-
surements indicate that sediment scouring from submerged CSO dis-
charges does occur in the lower portion of the river.
Visual observations in the lower river have also indicated the
significance of sediment scouring. During severe storm events,
the plumes from the numerous outfalls and the resulting scour of
sediments were observed. The overall significance of scouring
from CSO discharges is evident when one considers that there are
179
-------
approximately 40 outfalls discharging into the Milwaukee River
below the North Avenue Dam. Although some of these outfalls have
irregular shapes, the majority are circular and range from 36 in.
(0.9 m) to 96 in. (2.7 m) in diameter. Also, most of them are
submerged. Averaging the number of outfalls evenly over the
length of this portion of the river, approximately one CSO out-
fall exists every 300 ft (95 m) of river. Considering that many
of the outfalls enter the river in pairs from opposite banks, and
considering frequency of the outfall occurrence, the scouring
potential from these outfalls can be quite significant.
Long Term Modeling
In order to evaluate the water quality effects of various CSO
abatement alternatives, a long term simulation of water quality
was devised to quantify the water quality improvement associated
with each alternative. Harper's model was used to evaluate years
of rainfall-runoff loads from the contributing areas as predicted
by the STORM model. The output of the model was then used to
determine the frequency or magnitude of standards violations in
the river as a function of the CSO alternatives.
In order to predict the impact of sediment scouring by CSO out-
falls during rainfall events an emperical expression was developed
for the model using multiple regression analyses. Observed de-
creases in DO at the continuous monitoring locations in the river
following 20 rainfall events of varying volume and intensity were
analyzed versus the rainfall, runoff and river conditions for
each event. The change in dissolved oxygen concentration (decline)
was correlated with the following variables to determint the most
significant relationship:
Rainfall volume
Peak rainfall invensity
Average rainfall intensity
Volume of CSO discharge
Peak river flow
Change in river flow
Time since previous runoff
event
The results of this analysis indicated that rainfall volume, vol-
ume of CSO discharge and time since the previous events were
significant at the 95 percent confidence limits using a linear
equation.
A similar relationship has been developed to predict the duration
of the decline in DO since this parameter is important in quanti-
fying the frequency of DO violations. The duration is much more
difficult to predict since the lower reaches of the river are in-
fluenced by the flux of water from Lake Michigan. Despite this
difficulty, a linear relationship of DO sag duration to rainfall
volume, CSO volume and time since the previous storm was found to
be significant at the 90% confidence limits.
182
-------
Within the long term simulation of water quality an expression has
been added to the model to provide for a time varying sediment
oxygen demand CSOD) rate. This SOD rate is modeled as an immed-
iate high demand (distiirbed] at the beginning of a CSO event which
is reduced at an exponential rate to the dry weather (undisturbed)
demand between storms. The magnitude and duration of the increase
in SOD rates are determined through the use of previously des-
cribed empirical equations. The equations, for time varying SOD
rates in the model are controlled by three coefficients:
Magnitude (A)
Decline (B)
Time of Decline (T)
The A Coefficient is used to adjust the magnitude of the initial
increase in the SOD rate due to a particular storm event while the
B coefficient is used to adjust the exponentional decline in this
high SOD rate to the lower dry weather rate. Coefficient T is the
total duration of the increased SOD rate as predicted with the
regression equation.
Harper's model is quite sensitive to the A and B coefficients
in terms of the rate of dissolved oxygen decline following a
storm event. These coefficients can be adjusted in a manner
that nearly all of the decline in dissolved oxygen occurs dur-
ing the first or second 3-hour time step following a storm
event. In order to match observed changes in the DO concen-
tration in the lower Milwaukee River, the A and B coefficients
are set so that the ratio of the peak SOD rate to average SOD
rate following a storm ranges between 1.1 and 1.5 depending on
the size of the storm event. Although this method of setting
the coefficients is arbitrary, the predicted dissolved oxygen
decline matches the observed decline at St. Paul with good accur-
acy.
The major assumption in the development of this procedure for
predicting the DO impact of scoured sediments is that the ob-
served rapid decline in DO is primarily due to the sediments.
Some of the decline is due to the organic loading from CSO,
storm sewers and upstream loads, however, the rates of DO de-
cline far exceed the rates which could be due to the measured
BOD5 and TOG concentrations in the river. For this reason, the
previous assumptions are justified and when implemented in the
model, provide an excellent prediction of the monitored dissolved
oxygen concentrations.
183
-------
EVALUATION OF CSO IMPACT
In order to evaluate the contribution of CSO to the DO impacts
within the Milwaukee River, a sensitivity analysis using the in-
stream model network with various CSO loads was conducted. The
STORM and instream model were run for one year of rainfall for the
period April through October when non-frozen ground and non-ice
cover conditions are present. In order to measure the differences
in the DO curves that are computed by Harper's model, the tech-
nique shown in Figure 9 was utilized. The shaded area below 5.0
mg/1 standard represents the magnitude of the dissolved oxygen
impact which has the units of mg-day/1. The magnitude measure-
ment represents the difference in water quality between a given
standard and the predicted values. Another measurement which is
used to quantify the DO impact is the number of days of violation
that occur within the simulated time period. Thus, any part of
a day which contains a violation of the standard will be listed
as such. Figure 9 presents the data where there are eight DO
values listed per day in the model output. The 5.0 mg/1 standard
is violated during day 2 and 3 of the period with day 2 having
four values below the 5.0 mg/1 standard and day 3 having one.
Two days of violation will then be listed along with the magni-
tude of the violations for this period.
X
o
o
LLl
o
01
V)
1Q
8
6
*»
/\
AT NORTH AVENUE DAM
STANDARDS
VIOLATION
1 1 i II 1 1
DAY 1
II 1 1 1 1 1
DAY 2
TIME OF DAY
1 1 1 1 1 1
DAY 3
Figure 9. Typical model output for determining frequency
and magnitude of dissolved oxygen violations.
184
-------
In order to quantify the sensitivity of the river to the loads
generated by CSO, an extreme year of rainfall was selected from
the 38 years of record available for the Milwaukee area. The
predicted flow and quality from STORM for each storm event were
input to Harper's model to generate the frequency and magnitude
of DO violations at each site for the April through October per-
iod using mean monthly low flows. The DO standards for determining
the magnitude the frequency of violations were 5.0 rog/1 for sites
upstream of the North Avenue Dam and 2.0 mg/1 for sites downstream
of the dam as discussed previously.
Next, the predicted pounds of BOD, suspended solids and numbers of
fecal coliforms for each overflow event during the year in the CSO
area were doubled and loaded into Harper's model with the same
storm sewer and upstream boundary loads as the previous model runs.
This technique was used to determine the change in DO quality with
this 200 percent increase in CSO pollutant load. Similarly, the
CSO load and flow was also completely removed to simulate the com-
plete elimination of CSO. These three simulations were run using
the CSO sediment scour potential for each runoff event and they
were run without the scour mechanism in order to quantify the dif-
ferences between these conditions. The results of these model runs
shown in Figure 10 for both the with and without scour mechanism.
The data of Figure 11 provide a means of evaluating the contribution
of CSO to the impacts on DO at various locations within the
Milwaukee River. For example, the North Avenue data shows that
the doubling of the CSO load results in an increase of approxi-
mately 20 percent in the magnitude of the DQ violations of the
5.0 mg/1 standard. Removal of all CSO at this site results in a
21 percent decrease in the magnitude. Sediment scouring from sub-
merged CSO outfalls does not occur in the portions of the river
upstream of. the North Avenue Dam. Therefore, the DO impacts at
North Avenue are strictly a function of the wet weather loads and
the low flow conditions that were used for this analysis. At
Walnut Street, where sediment scouring does occur, the differences
are more pronounced. Doubling the CSO load results in a 78 per-
cent increase in the magnitude of the standard violations of
2.0 mg/1. Complete removal of the CSO load results in zero magni-
tude below the standard. The contribution of sediment scour to
these violations, as determined using the without scour results,
is approximately 69 percent. The existing magnitude of DO vio-
lations is 17.2 mg-days/1 with scour. This decreases to 5.3
mg-days/1 (69 percent decrease) when the scour mechanism is re-
moved. At higher CSO loads (200 percent), the sediment contribution
to the magnitude of DO violations is reduced to approximately 37
percent at Walnut Street. This reduction in influence of sedi-
ments is due to the higher dissolved and suspended loads utilized
in the 200 percent CSO load simulation.
185
-------
r
a
o
to
O
O
to
UJ
s:
o
o
o
UJ
o
a
o
OC
a.
0)
13
t/J
r~-
r-
CTl
CM
I
CO
CD
c
a
c
a)
ZJ
ro
a.
< —
c
o
(0
o
i_
l/Btu 'N39AXO QBATOSSIQ
136
-------
150
ro
-o
H-
CJ
LlJ
CJ
X
o
o
to
to
100
50
1954
WITH SCOUR
n
EX 1 ST
200% i
ZERO,
_' NORTH AVENUE
,EX I ST
200%
WALNUT
ZERO
EXIST! 2002
T!
I
ZERO
ST. PAUL
* Figure 11. Sens PtTvi ty'df'di ssol Ved oxygen results in CSO: loads';
:•' \ . "'•
The St. Paut Avenue site has the^most drastic changes in water-;
quality when, comparing t'ne sensitivity of various loads. By
doubling the CSOMoad, the DO magnitude below 2.0 mg/1 increases
by 22 percent, while removing 'the CSO load reduces the impact by
96 percent. Removal of the,.scour; mechanism from the model net-
work reduces the impact of the existing CSO load by 91 percent
when compared to the.with scour alternative. This result indi-
cates the significance!pf the scouring of sediments on DO conditions
in the lower portions of%the*Milwaukee River. In comparison, the
CSO loads have, only a minor.influence on DO levels during the low
flow conditions used for this s,imulation.
''' " ••-1 ',' *
Evaluation of CSO Abatement Alternatives
A total of three general CSO abatement alternatives"were investi-
gated using various pollutant removal efficiencies and design
criteria. The three alternatives and their combinations are
listed as: ,,..;.,,
Sewer separation - partial and complete
End of Pipe - 35% solids removal, 20% BOD removal,
98% fecal coliform removal
187
-------
Out of Basin - conveyance, storage and treatment of
the overflows in deep tunnels with the
treated effluent discharged directly
to Lake Michigan
- designed for a 1/2, 1,2,5 year level
of protection Cdesign storm)
Each of these alternatives were simulated using the STORM model
with hourly rainfall for one year of record. The pollutant load-
ings for each rainfall event within the April through October
period were input to Harper's model to determine the resulting DO
and fecal coliform concentrations within the river. The storm
sewer and upstream loads that resulted from the rainfall events
were also input to the receiving water model on an hourly basis.
In addition to the simulation of CSO abatement alternatives, the
model network was also run for two additional cases: the exist-
ing, unabated CSO system the total elimination of all CSO flow and
quality. These two schemes were run to determine the sensitivity
of the receiving water to the CSO loads and to determine the "do-
nothing" alternative or present day results. The results of these
simulations are shown in Figure 11 for three sites within the
Milwaukee river study area.
Figure 12 represents the DO and fecal coliform results at the
North Avenue dam site which is only slightly affected by the
sediment mechanisms which are present in the lower reaches of
the river. As can be seen in this figure, the DO results show
little difference between the existing CSO conditions and any
of the alternative CSO control techniques including 100% removal
of the CSO. This pattern is a result of the DO impact at this
site being related to the dissolved load that is present in the
river from upstream sources which is not changed by the removal
of CSO. The North Avenue site has approximately 50% of the CSO
drainage area discharging upstream of this location but the im-
pact on DO from these discharges is not expressed at this site.
The fecal coliform results shown in this figure reflect the re-
moval of CSO in a much more dramatic manner than the DO results.
Thus, the zero CSO alternative shows one day of violation which
is a result of the upstream loads of fecal coliform while this
site contains 40 days of violation when the existing CSO load
is present. The sewer separation alternative also shows the dif-
ference in complete removal of the sanitary component in the
combined system and the partial removal of this element.
Figures 13 and 14 represent the DO and fecal coliform results for
two downstream sites on the Milwaukee river. Here, the differ-
ences in DO magnitudes are much more evident than at the upstream
site because of the presence of the sediments and the influence
of the scour mechanism is more dominant. Each of the DO results
for an alternative can be evaluated by relating the degree of
188
-------
SEPARATION^
MAGNITUDE OF DO
VIOLATIONS PER YEAR
mg - days/1
(STANDARD =5 mg/1
5O-
4O-
3O-
20-
10-
n_
EXISTW&
rM
*• EOP
OUT- OF- BA&IN
COM.
'/a
1 2 5
!§?
CSC
NUMBER OF DAYS PER,
'YEAR OF FECAL • •
COLIFORM VIOLATIONS
OF 200/100'ml '"' :
(days) ^ ' ;
50-
4O-
-3O-
2O-
10-
O-
EXISTIN6 ,
SEPARATION
, PAR.
OUT-OF-eASIN
IA - "' zgRp.-
r P-:- | -> 5 EOF cso
Figure 12. Comparison of the average dissolved oxygen and fecal coliform
results during,the typical and extreme runoff years for the alternatives
at North Avenue.
189
-------
MAGNITUDE OF DO
VIOLATIONS PER YEAR
mg - days/I
(STANDARD = 2 mg/1)
io-
EXISTING
SEPARATION
PAFt EOF
COM.
CUT-OF-BASIN
\ 1 2 5
ZERO
CSO
NUMBER OF DAYS PER
YEAR OF FECAL
COLIFORM VIOLATIONS
OF 1000/100 ml
(days)
50-
40-
20-
20-
10-
n.
EXIStlNG
•
• SEPARATION
.
PAR.
OUT- OF- BASIN
COM. ^125
. rn . . .
ZERO
EOF rsn
tilt
Figure 13. Comparison of the average dissolved oxygen and fecal coliform
results during the typical and extreme runoff years for the alternatives
at Walnut Street.
190
-------
EXISTING
MAGNITUDE OF DO
VIOLATIONS PER YEAR
mg - days/1
(STANDARD = 2 mg/1)
I2O=,
iOO-
30-
6O-
4O-
2O-
n_
SEPARATION
1
EOP
PAR.
COM.
OUT- OF- BASIN
Vi
|~| ' 2 5
1 I 1 1 1
ZERO
CSO
i 1
NUMBER OF DAYS PER
YEAR OF FECAL
COLIFORM VIOLATIONS
OF 1000/100 ml
(days)
30-
20-
IQ-
O-
EXISTINS
SEPAFiATION
PAR.
OUT-OF-BASIN
COM.
2 5
ZERO
EOP CSO
J L
J-
Figure 14. Comparison of the average dissolved oxygen and fecal coliform
results during the typical and extreme runoff years for the alternatives
at St. Paul Avenue.
191
-------
scour associated with, that alternative. For example, partial and
complete separation do not remove the submerged outfalls, and as
a result the scour potential of this alternative is the same as
for existing conditions. The out-of-basin alternative removes
the scour mechanism by routing the overflows to tunnels which
eventually discharge outside the river basin. None of the al-
ternatives provide the improvement in DO violations which could
be achieved by complete elimination of CSO. .However, the one,
two and five year sizes of the out-of-basin concept provide DO
improvements which approach this level. Partial separation pro-
vides the least DO improvement.
192
-------
COST-EFFECTIVENESS OF ALTERNATIVES
The evaluation of each alternative in terms of water quality
improvements cannot be completed until the cost associated with
each is compared. The detailed costs required for this analysis
are continuously being updated for the Milwaukee Metropolitan
Sewerage District. It must be emphasized that the costs shown in
the following pages are previous estimates and have been updated
a number of times since this project was completed.
In order to provide an indication of the shape of the cost versus
water quality improvement curve, cost data developed for the en-
tire CSO area for the out-of-basin and sewer separation concepts
was utilized. These cost data are presented on an annual basis
in order to determine where the "knee" of the curve .will occur.
The total costs for all three river basins will be utilized
throughout this analysis with the water quality data1 from the
Milwaukee River to insure uniformity between in and out-of-basin
alternatives.
Water Quality Improvement Versus Cost
The water quality improvement parameter for use in this analysis
was quantified utilizing the percent improvement in DO magnitude
or days of fecal coliform violations when compared to existing
conditions. For the purposes of the following discussion, the ex-
treme year of 1941 will be used for the examples in the text. The
data to be used in developing the curve of cost versus water qual-
ity as measured for DO is shown in Figure 15. As the figure indi-
cates, the four levels of protection are bunched at or near the
zero CSO discharge line while the existing CSO conditions are at
the other extreme with no points between. In order to place a line
between the one-half year point and the existing conditions loca-
tion a few assumptions must be made until further data is available.
The first and most critical assumption is that a significant number
of dollars on an alternative would have to be spent before any no-
ticeable water quality improvements would,,occur. This means the line
connecting the two points will start out'ffrom the existing condi-
tions location (zero dollars/year) in an-almost vertical trajectory
before leveling off as it approaches the mid-point of the axis.
Figure 16 illustrates the DO and fecal coliform measurements of
water quality improvement versus cost for the alternatives at St.
Paul Avenue. The locations of the partial and complete sewer sep-
193
-------
OUT-OF-BASIN
106|>/YEAR
50-i
40-
30-
20-
10-
n-
U C V
EXISTING CONDITIONS
S
CL vjr t
r
2
1
\i
'2
"*- ZE
DIS
0 20 40 60 80 100
PERCENT WATER QUALITY IMPROVEMENT IN
IN THE MAGNITUDE OF D O VIOLATIONS
IMPROVING WATER QUALITY
Figure 15- Cost versus water quality Improvement and the magnitude of DO
violations for the 19^1 extreme runoff year at St. Paul Avenue.
aration points are also included in Figure 16 as well as lines
indicating a possible shape of the curve. The problem of insuf-
ficient data to generate the cost versus water quality improvement
curve is also evident with the fecal coliform results. Despite
these problems, it is evident in the figures that the "knee" of
the curve would occur at a level of protection of one-half year
or less for both the DO and fecal coliform results.
Conclusion
This paper has attempted to briefly summarize the receiving water
investigations that have been undertaken in conjunction with the
CSO problem in Milwaukee. To date, the selection of the final
solution for this problem involves a combination of sewer separ-
ation and storage treatment of the overflows. The final level
of protection of the storage facilities will be dependent upon
litigation presently being reviewed by the U.S. Supreme Court.
If this litigation is overthrown, then the storage facilities
would be designed for a two year level of protection.
194
-------
10*$/YEAR
LH COMPLETE SEPARATION
A PARTIAL SEPARATION
5CH
40-
OUT-OF-BASIN
LEVEL OF PROTECTION
ZERO CSO
DISCHARGE
80
100
PERCENT WATER QUALITY IMPROVEMENT
IN THE MAGNITUDE OF D O VIOLATIONS
IMPROVING WATER QUALJTY
OUT-OF—BASIN
LEVEL OF PROTECTION
ZERO CSO
DISCHARGE
0 20 40 60 80 100
PERCENT WATER QUALITY IMPROVEMENT IN
THE DAYS OF FECAL COLIFORM VIOLATIONS
V
IMPROVING WATER QUALITY
Figure 16. Cost versus water quality improvement for the 195^ extreme
runoff year at St. Paul Aven-ue. .
195
-------
ACKNOWLEDGEMENT
This paper is based upon the results of two separate projects
related to Milwaukee's Combined Sewer Overflow Study. The
project final reports are entitled;
"Verification of the Water Quality Impacts
of Combined Sewer Overflow Using Real Time
Data"
for the Storm and Combined Sewer Section - USEPA Edison, N.J.;
and
"Water Quality Analysis of the Milwaukee
River to Meet PRM 75-34 (PG 61) Requirements'1
for the Milwaukee Metropolitan/Sewerage District. :
Both of these project reports were co-authored by
Thomas L. Meinholz
William A. Kreutzberger
Martin E. Harper
presently of EcolSciences, inc.
Milwaukee
Rexnord Env. Res. Center,
MiIwaukee
Harper-Owes, Seattle
The principal author,would also like to acknowledge Messrs Richard
Field and John English - USEPA and Messrs James Meinholz and
James Ibach - Milwaukee Metropolitan Sewerage District;
3CE
.«oJ
IB
UL
196
-------
URBAN STORMWATER IMPACTS ON RECEIVING
STREAMS IN NORTH CAROLINA
by
E. Ryland Brown, Env. Scientist
. Ross S. Green/*Env. Biologist
N.C. Department,of Natural Resources
and Community Development
Diyision of Environmental Management
•' Raleigh, North' Carolina' 27611
ABSTRACT '
Studies were conducted in North. Carolina by the. Division .of Environmen-
tal Management as. part of "the statewide 208 program to determine* the extent
of water quality degradation in streams* receiving urban stormwater runoff.
Selected streams were monitored in three of the larger urban areas within.the
state; Asheville, Raleigh, and Winston-Salem. The monitoring program in-
volved physical/chemical sampling .under both high (storm) and low streamflow.
conditions. Parameters sampled include pH, temperature, D.O., BOD^, COD,
fecal coliform, ammonia-nitrogen, total Kjeldahl nitrogen, nitrite plus
nitrate nitrogen, total phosphorous,'iron, mercury, lead, zinc, copper,
cadmium, chromium, nickel and solids. Biological sampling of aquatic benthic
macroinvertebrates was conducted concurrently to further document water
quality conditions. .'..-,... ...
•In Asheville, a stream draining a mixed land use urban area was studied.
-In the Raleigh area, three streams were monitored, including one that drains,
a highly impervious watershed comprised primarily of a shopping center.: The
most intensive sampling was conducted in the Winston-Salem area. In this
locality, streams draining both residential and Central Business District
watersheds were monitored to investigate water quality characteristics
associated with these different land use types. For comparison purposes,
control stations upstream of urban inputs were also established and monitored.
All of the urban streams were found to exhibit extensive water quality
problems. Physical/chemical sampling consistently revealed high pollutant
concentrations for several parameters under high flow conditions. Notable
problematic parameters included suspended solids, lead, and some nutrients.
Several parameters were found to frequently be present in high concentrations
under low flow conditions also. Variations in pollutant concentrations from
197
-------
the different land use types were also observed.
The Division of Environmental Management's biological monitoring group
has developed an index of the macroinvertebrate organisms found in North
Carolina streams. As part of this index, each taxa is rated according to its
ability to withstand pollutional stressed conditions. All of the urban
streams monitored were found to be extensively biologically degraded. Popu-
lations of Diptera and Oligochaeta, pollution-tolerant organisms, averaged
over 90% of the existing fauna. In contrast, control station population
percentages of these organisms were a maximum of 15% of the fauna in the
mountain stream, and 35% in Piedmont streams. Many intolerant groups were
not just reduced in numbers in the urban streams; they were usually complete-
ly absent.
These studies have shown that, under present conditions, almost all
urban streams will be unable to meet the 1983 water quality goals. Improve-
ment of water quality conditions in these streams is dependent upon the
development of programs to effectively manage the entry of pollutants from
urban land surfaces and unrecorded or intermittent point sources.
.INTRODUCTION
Federal legislation enacted during the early 1970's provided the impetus
for controlling the entry of pollutants from municipal and industrial dis-
charges to the watercourses of the country. Since then, many advances have
been made in the treatment of these discharges. These advances have greatly
reduced the loads of various types of pollutants entering water bodies from
these point sources.
In conjunction with the national commitment to limit point source dis-
charges came the realization that significant inputs of pollutants were also
contributed by non-point sources. Public Law 92-500, Section 208 provided
the mechanism for addressing these non-point pollution sources. In North
Carolina, one of the non-point pollution sources investigated was urban
stormwater runoff. As part of this work, several streams were monitored in
three of the larger urban areas within the State. The purpose of these
studies was to determine the impa'cts associated with urban runoff and the
management programs needed to address any observed impacts.
The objective of this paper is to present the results of monitoring
activities conducted on these streams. A methodology is presented that
utilizes the monitoring of aquatic benthic macroinvertebrate communities to
complement traditional physical/chemical sampling techniques. The results
of both sampling approaches are used to define the water quality status of
the streams.
.THE ENVIRONMENTAL SETTING
The southeast United^S'tates, extending from Maryland to Alabama, is
characterized by three dis,finct physiographic regions: the Appalachian
mountains, the Piedmont Plateau, and the Coastal Plain. Urbanization is
greatest in the Piedmont, ,^due largely to the abundant supply of water that
has attracted industry. This urbanization is characterized by numerous
198
-------
medium size cities (populations of 30,000 to 150,000) located primarily along
the major watercourses.
North Carolina is primarily a rural state. In 1970, only 23.9 percent
of the State's population lived in designated urbanized areas, in contrast to
58.3 percent nationwide (1). The Piedmont is the most populated region, and
the trend towards urbanization in this part of the State is increasing. It
has been estimated that in 1980 the Piedmont will contain 52 percent of the
State's population while 38 percent will live in the Coastal Plain and 10
percent in the mountains (2). The locations of the physiographic regions and
larger cities in North Carolina are shown in Figure 1.
, APPALACHIAN MOUNTAINS.
| ' | PIEDMONT PLATEAU
lill COASTAL PLAIN
••i
.'id
iC.
Figure 1.
Carolina.
Location of the Physiographic Regions and Larger Cities in North
:The majority of the municipalities in the State have.separate sewer
systems for sewage and stormwater. The older Central Business Districts are
usually served by underground storm sewers, with construction ranging from
brick in the older areas to concrete in the newer,lines. These.inner-city
areas 'are normally surrounded by well-established: old residential areas with
open channeled streams. Beyond this are generally rapidly developing
suburban areas.
URBAN STREAM SAMPLING TECHNIQUES
' T1
Characterization of the magnitude and importance of urban stormwater
runoff is a difficult task. Sources of pollutants are diverse and hard to
characterize. In most cases, the pollutants entering streams via urban
drainage originate from a mixture of sources: washoff from urban land sur-
faces; atmospheric fallout and washout; garbage deposited'in or near streams;
polluted groundwater; leaking or broken sanitary^sewers; cross connections
between sanitary and storm sewers; sanitary sewer•"• overflow; small,
199
-------
unpermitted or intermittent point sources; or benthic resuspension of mate-
rial settled out in pipes or stream pools. All of these sources must be
considered as to their relative importance when determining a management
program for a specific area.
Urban stream sampling is likewise not an easy task. Periodic grab
sampling under low and high flow conditions can provide considerable infor-
mation concerning the range of pollutant concentrations that are found in
the streams. Concentrations can then be compared to established water
quality standards to provide a mechanism for determining relative water
quality conditions. Automatic sampling equipment, when available, can give
more complete data records for impact assessment. However, this approach
can be very expensive when applied on a large scale, and instrument malfunc-
tions can result in incomplete data records upon which to base total pollut-
ant loadings.
The costs of chemical analyses often limit the number of samples and
parameters that can be analyzed in urban stream monitoring programs.
Therefore, certain pollutant parameters or possible high concentration
pollutant slugs that can cause aquatic biological degradation may not be
detected in normal monitoring efforts.
Sampling for aquatic benthic macroinvertebrate organisms can provide a
valuable mechanism for quantifying the extent of stream degradation. When
used in conjunction with physical/chemical sampling techniques, a complete
picture of problematic pollutants and resultant in-stream effects may be
acquired.
STUDY LOCATIONS
Urban streams in the cities of Asheville, Raleigh, and Winston-Salem
were sampled between fall, 1977 and early 1979. Names, predominant land use,
location and drainage area of these streams are included in Table 1.
TABLE 1. CHARACTERISTICS OF STREAMS STUDIED
Stream
Sweeten Creek
Pigeonhouse Branch
Tar Branch
Salem Creek
Silas Creek Trib. 1
Silas Creek Trib. 2
Drainage Predominant
Area (mi ) Land Use
5.2 Commercial/Industrial
2.4 CBD/Commercial
0.4 CBD
31.7 Mixed Urban
0.5 Residential
0.4 Residential
City
Asheville
Raleigh
Winston-Salem
Winston-Salem
Winston-Salem
Winston-Salem
Additionally, study stations were located upstream of the urbanized
areas along Sweeten Creek and Salem Creek. Another Piedmont province
station was located on Fourmile Creek, a minimally polluted stream. These
stations served as control localities for comparisons of water quality
200
-------
characteristics between areas receiving and not receiving urban drainage.
Sweeten Creek receives considerable urban drainage from the city of
Asheville (population 60,000). The stream runs parallel to many businesses
and small industries. Recent studies have also indicated that broken sewer
lines contribute drainage to the creek on an intermittent basis. The
monitoring station is located about one stream mile into the commercial
section. The drainage area is about 5.2 square miles at this point, and no
significant point source dischargers are known to exist. The control station
upstream of the urbanized area is located at the headwaters of Sweeten Creek,
approximately % mile upstream of Busbee Reservoir.
Pigeonhouse Branch, an urban stream in Raleigh (population 150,000),
drains a watershed consisting of a large shopping center, high density
residential areas, commercial areas and small industries. Two monitoring
stations were located on the stream. The first station is located at a
point where the drainage area is only about 100 acres, and is primarily
made up of the shopping center. Approximately 65 percent of the watershed
is impervious due to the extensive parking lots, street surfaces, and
rooftops. Resultant high volume and high velocity stormwater runoff has
caused extensive downstream streambank degradation. The second station is
located near West Street. At this point, the watershed is approximately 2.4
square miles in size and the stream has received drainage from the various
residential and commercial areas in addition to the shopping center.
• Tar Branch drains the Central Business District (CBD) of Winston-Salem
(population 150,000). The watershed is highly impervious and is approxi-
mately 0.4 square miles in size. Much of the stream runs through pipes and
culverts. It becomes open channel flow approximately 25 feet upstream of
the established sampling station. The stream is approximately two meters
wide and averages seven centimeters deep there.
The station located on Salem Creek downstream of Winston-Salem is near
Silas Creek Parkway. At this point, the stream has received drainage from
the Central Business District via Tar Branch plus drainage from various
residential, commercial and industrial land use areas. The control station
on Salem Creek is located approximately 1 mile downstream of Salem Lake.
The stream is four meters wide and has an average depth of 30 centimeters at
this location.
The two Silas Creek tributaries in Winston-Salem drain residential
areas. Tributary 1 drains a watershed consisting mainly of medium density,
high income housing. The second tributary drains an area that is mainly
medium density, middle income housing. The'watersheds were chosen because
they are both well established, have little ongoing construction, and well
represent typical residential areas in the larger cities within the state.
PHYSICAL/CHEMICAL SAMPLING
Physical/chemical sampling was conducted on a nonperiodic basis at the
monitoring stations beginning in spring, 1977 and running through 1978.
Samples were collected by hand using a DH-48 depth-integrated sampler. Moni-
toring was conducted under both normal (low flow) conditions and storm (high
flow ) conditions. Parameters sampled include pH, temperature, D.O., BOD5,
201
-------
COD, fecal coliform, ammonia-nitrogen, total Kjeldahl nitrogen, nitrite plus
nitrate nitrogen, total phosphorous, iron, mercury, lead, zinc, copper,
cadmium, chromium, nickel, and solids.
Data collected were examined based on a comparison of the observed
pollutant concentrations to proposed North Carolina water quality standards.
Those samples having concentrations in excess of the proposed standards were
termed "violations". This is a useful expression, since it is envisioned
that when non-point source compliance monitoring programs are established,
it will be necessary to express allowable' limits for various pollutants on a
probablistic or frequency basis (3). It would be impractical if not impos-
sible to comply with pollutant limitations 100 percent of the time.
The data collected demonstrated some obvious urban-related water quality
problems. High concentrations of several parameters and "violations" of
proposed water quality standards were found in all of the urban streams under
storm flow conditions. Additionally, monitoring showed that water quality
problems also exist during low flow conditions. Several of the streams were
reclassified in 1977 to support fish and wildlife. The streams had been
previously classified only for urban and industrial conveyance usage. In'
many cases, the degradation of water is to the point that it cannot support
fish life, and the poor quality exists during normal as well as storm flow
conditions.
Distinct differences in pollutant concentrations were observed for
several parameters in samples from the CBD arid residential areas in Winston-
Salem. Concentrations of most pollutants were found to consistently be
higher in the CBD samples. This observation is consistent with other
investigations of urban streams (4,5,6), and is important in determining ways
of abating urban water pollution problems.
Nutrient Concentrations; Nutrient parameters monitored included NH3,
TKN, NC>2 + NOs/ anc^ 2 + NO3 were found to be even more visibly different
between the CBD, residential, and control stations under low flow conditions.
Mean quarterly concentrations were found to be several times higher in the
CBD watershed for all quarters. Control station concentrations were consis-
tently lower than those in both the CBD and residential samples. Average
202
-------
values observed, at. the downstream Salem Greek station-were between the CBD
and residential samples, again possibly reflecting the mixed land use
drainage to the stream.
TABLE 2. MEAN LOW-FLOW NUTRIENT CONCENTRATIONS
IN WINSTON-SALEM STREAMS
Stream '
Salem Creek Control
Silas Creek Trib. 1
Silas Creek Trib. 2 , . _
Tar Branch
Salem Creek Downstream
Watershed
Type
Rural
Residential
Residential
CBD
Mixed Urban ,
Mean Low-Flow
Nutrient Concentration (mg/1) *
NH3 ,
.08
.04
.05,
.56
.20,;
N02+NO3 TKN
.27 .29
1.00 / .19
-.. .52 .18
4.48 "1.56
"; .39 .45
. rpp
.07
.16
.20
.24
.09
*Average of 20 samples taken during 1978. ; " "
Average total phosphorous concentrations did not vary greatly between
samples- collected at the different stations.-; Average values were consis-
tently low, ranging from 0.07 mg/1 at the control station to 0.24 mg/1 at
the CBD station. . • ,,, : . ..•;;.„<•.,. ; ;•...... ...... ..>... ,../,•
Concentrations of" each of these parameters generally tended to be
greater during high 'flow conditions", as compared to low flow conditions, "al-
though this was not always the case. These'elevated concentrations, when
they occurred, were not as pronounced as those of some of the heavy metal
parameters. No visible seasonal:concentration differences were evident" in ,
this study. : ^ • • .•--'/, :; •• . • • :,- -•.• -• - ; •• - ' •
Heavy Metal Concentrations: Table 3 shows the monitoring results for
the heavy metal parameters of cadmium, iron> lead, mercury, and nickel'for
data collected over the first three quarters of 1978 in the'Winston-Salem
area. It is important to note that all metals did not show the'same re^
lationships. For instance, under low flow conditions, iron "violations"
were 8 percent in the residential area, and 13 percent in the CBD. At the
downstream station which receives flow from both of these land use types,
"violations" occurred in 25 percent of the samples. Average iron concen-
trations appeared higher in the CBD watershed than in the residential areas.
However, the upstream control station was found to have the highest "percent
violation" for iron. This may reflect the natural high iron content of the
soils in this area. Mercury concentrations were similar throughout the
different land uses, including the upstream control area. "Violations"
averaged 45 percent. Under low flow conditions, lead "violations" appeared
highest in the residential areas; no "violations"were found in the upstream
control samples. Storm flow samples were found to have much higher lead
concentrations than low flow samples in the CBD. "Violations" occurred 100
percent of the time in samples collected during storm events.
203
-------
Zinc concentrations appeared higher in the CBD than in the residential
areas. Percent "violations" are not expressed in this work since proposed
North Carolina standards are based on a 96-hour LC50 figure rather than in
terms of concentration for this parameter. There were no concentrations of
cadmium and chromium, and very few concentrations of nickel high enough to
constitute "violations" of the proposed standards under either low or high
flow conditions.
TABLE 3. HEAVY METALS "VIOLATIONS"
Standards for Five Toxics Investigated
.0211 (a)(4) General Toxic Standards (7)
Cadmium: not greater than 4.0 ug/1 for non-trout waters
Iron: not greater than 1.0 mg/1 for non-trout waters
Lead: not greater than 0.03 mg/1 for non-trout waters
Mercury: not greater than 0.05 ug/1 for non-trout waters
Nickel: not greater than 0.1 mg/1 for non-trout waters
"Percent Violations"
Cadmium
Iron
Upstream (Salem Creek Control)
Residential (Silas Creek Tribs)
CBD (Tar Branch)
Urban (Salem Creek Downstream)
Avg. (All Stations except Control)
Lead*
Upstream
Residential
CBD
Urban
Avg.
Mercury
Upstream
Residential
CBD
Urban
Avg.
Nickel
Upstream
Residential
CBD
Urban
Avg.
Low Flow Avg. Storm Flow Avg.
0% for all samples taken
37%
8%
13%
25%
15%
20%
8%
10%
13%
40%
45%
50%
40%
45%
92%
100%
13%
0%
3%
8% 9%
for all samples taken
4%
*The proposed standard for lead is below the laboratory detection limit of
204
-------
100 ug/1. The reported "violations" were samples with concentrations of
100 ug/1 or greater. Some of the samples with concentrations less than 100
ug/1 may actually also be greater than the proposed 30 ug/1 standard.
Mercury concentrations appeared to be higher under low flow conditions
than under high flow conditions. Stations in the CBD evidenced proposed
standard "violations" 50 percent of the time under low flow conditions; high
flow "violations" there were observed 13 percent of the time. Concentrations
of lead, iron, and zinc, however, were found to be many times higher under
wet than under dry flow conditions (Figure 2). iron, for example, was found
in concentrations in excess of the proposed standards 92 percent of the time
under high flow conditions in the CBD; dry flow "violations" averaged 13
percent.
o
o-i
VD
O
lO-
o
in-
Lead
(ug/1)
o
O-i
o
o
CO
o
o
O'
CM
o
s-
Iron
(ug/1)
o
on
00
o
Or-
(0
O
o-
Zinc
(ug/1)
High
Flow
Low
Flow
High Low
Flow Flow
High Low
Flow Flow
Figure 2. Average Concentrations of Heavy Metals in Streams Draining the CBD
Under Different Flow Conditions.
Samples taken from Pigeonhouse Branch in Raleigh were also found to
frequently have quite high concentrations of certain heavy metals under both
high and low flow conditions. Lead concentrations were found to be extremely
high under storm flow conditions. Figure 3 shows lead concentrations ob-
served during the storm of May 23, 1977. In all cases, concentrations of
lead were well above 30 ug/1 proposed standard. From the graph, it appears
that this watershed exhibits a "first flush" of lead during a storm event.
The highest concentration of lead observed during the study was 6000 ug/1.
205
-------
•s
o
o-
o
n
o
o.
ID
CM
o
o.
CM
O
O
o
O"
in
'Lead
Zinc
^o i¥
ELAPSED TIME (MINUTES)
0 = initiation of runoff
O
-O
00
o
-o
(O
O is]
-O
O
-o
CM
20
Figure 3. Lead and Zinc Concentrations Observed During the Storm of I lay 23,1977
(Pigeonhouse Branch Watershed).
Zinc concentrations were also found to be relatively high in several of
the samples, both under high and low flow conditions. The concentrations of
zinc found during the storm of May 23, 1977 are also shown in Figure 3. The
peak concentration of zinc was observed later in the storm than that for
lead, i.e. this pollutant did not appear to be washed off in a first flush
manner as pronounced as that of lead. Other samples collected during the
study registered concentrations as high as 2100 ug/1.
One factor leading to the extremely large concentrations of these metals
observed during runoff events may be the extensive impervious areas within
the watershed and the associated potential for pollutant buildup and washoff.
A secondary factor may be the resuspension of pollutants that have settled
out in pipes or in pools of open channels during previous periods of dry
weather.
206
-------
(^gen-Demanding Material Concentrations; Both five day-biochemical
oxygen demand (6005), and chemical oxygen demand (COD) were monitored during
the course of this study. Quite wide concentration ranges were evident for
both of these parameters at both the CBD and residential area stations in
Winston-Salem under low flow conditions. BODs values were consistently
greater in the CBD samples than those from the residential areas. BODs
values reached a maximum of 18.6 mg/1 in the CBD; the highest residential
area value was 5.94 mg/1. Figure 4 shows a plot of BOD5 concentrations
found at one of the CBD stations and the upstream control station below
Salem Lake under low flow conditions. As can be seen, 6005 concentrations
were always greater at the urban stream station, often being many times
higher. This clearly shows the type of conditions that can be associated
with urban-affected streams.
18-
16-
14-
12-
10-
8-
6-
4-
2-
•CBD Sariples
°Control Sanples
2 4
Quarter 1
FT
Quarter 2
fiannle N-jmber
Quarter 3
Figure 4. BOD5 Concentrations observed at Winston-Salem CBD and Control
Stations.
BOD5 concentrations were found to be generally much higher under storm
flow conditions than under low flow conditions in the CBD watershed. Storm
flow concentrations as high as 27.9 mg/1 were observed. The average BODs
value for six samples taken at the CBD station during the storm of March 8,
1978 was 24.1 mg/1. In comparison, the average of 17 low flow samples taken
at this station during 1978 was 7.0 mg/1.
207
-------
COD concentrations were accordingly higher in the CBD area than in the
residential areas of Winston-Salem also. Mean quarterly COD sample concen-
trations were greater at the CBD station in all cases. Additionally, the
range of COD values was much wider at the CBD station. However, there was
not as visible a difference in concentration ranges between the two land use
types as there was for BODs concentrations. The maximum COD level observed
in the CBD watershed was 49 mg/1. The high value observed in the residential
area was 39 mg/1; most values were considerably lower.
High flow COD concentrations were found to be considerably higher than
low flow values in the CBD watershed. A maximum COD of 210 mg/1 was found
at the CBD station during the storm of March 8, 1978. The average of six
samples taken during this storm was 130 mg/1 COD. Six samples taken at the
same station during the storm of March 14, 1978 had an average COD of 73 mg/L
In corparison, 20 low flow samples taken at the station during 1978 averaged
21 mg/1.
In the Raleigh area, even larger concentrations of oxygen-demanding
materials were fourtd to enter the receiving stream under storm flow condi-
tions. Both BOD and COD levels were quite high in the majority of the storm
flow samples. BODs levels were found to be as high as 89 mg/1 compared to
the 0-4 mg/1 range found in most streams.
Figure 5 shows the BODs concentrations observed during the storm of
August 1, 1977. It can be seen that oxygen-demanding materials were washed
off in the largest concentrations in the early stages of the storm in
accordance with the first flush phenomenon. Peak concentrations were found
between five and ten minutes after the initiation of runoff. The corre-
sponding COD concentrations observed during this storm are also shown in
Figure 5. COD values as high as 520 mg/1 were observed. The wide COD/BOD5
ratios suggest the possible presence of toxic materials, inhibiting biochemi-
cal oxidation.
Fecal Coliform Concentrations: Concentrations of fecal coliform
microorganisms were also monitored. Concentrations varied greatly over the
study period, but on numerous occasions the 1000 microorganisms/100 ml pro-
posed standard was exceeded. A wide range of fecal coliform concentrations
was observed in low flow samples taken from streams draining the two dif-
ferent land uses in Winston-Salem. No readily visible difference in mean
concentrations could be detected between the two areas. Maximum concentra-
tions observed were 270,000 mo/100 ml in the CBD area and 190,000 mo/100 ml
in the residential areas. Concentrations observed during storm flow condi-
tions were also high on several occasions. The data serves to show that
fecal coliform contamination from some sources within the urban area is
occurring during both wet and dry periods. This observation is borne out by
the fact that no samples at the upstream control station had concentrations
large enough to exceed the proposed samples. At this station, the mean low
flow concentration was 77 mo/100 ml, and the maximum concentration was 340
no/100 ml.
208
-------
15 2025 3D 35 40 « 50
EIAPSED Til-IE (MINUTES)
0 = initiation o* runoff
Figure 5. BODc- and COD Concentrations Observed During the Storm of August 1,
1977 (Pigeonhouse Branch Watershed).
Low flow fecal ooliform concentrations averaged 2600 in Sweeten Creek in
work conducted in conjunction with the Land-of-Sky Regional Council of
Governments (8). In comparison, stormflow samples contained concentrations
averaged 8,233 mo/100 ml.
Suspended Solids; Solids were found to be present in large concentra-
tions during periods of storm flow. Samples taken during storm flow periods
exceeded the proposed 80 mg/1 standard in almost all cases. Figure 8 shows
the total solids and total suspended solids observed in Pigeonhouse Branch
during the storm of August 1, 1977. An obvious first flush response was
found to occur for this parameter, with total suspended solids values as high
as 518 mg/1 early in the storm. An interesting feature observed, for the
storm was that relatively large percentage of the solids in samples taken
early in the storm were in the dissolved state; later samples were quite
consistent in having solids mainly in the suspended state.
209
-------
10 15 20 25 30 35 40 45 50
EtAPSED THE (MB1OTES)
0 * initiation of runoff
Figure 6. Solids Concentrations Observed During the Storm of August 1, 1977
(Pigeonhouse Branch).
BIOLOGICAL SAMPLING
The biological monitoring group has devised a system of analysis through
which biological data can be evaluated to determine the degree of environ-
mental stress from nonpoint source pollutants. The symptoms of nonpoint
source pollution are reflected as changes in the biotic components of the
environment. This shift in aquatic community structure (especially the
bknthic macroinvertebrates) is used as a means of measuring environmental
inpact. Macroinvertebrates are useful biological monitors because they are
found in all aquatic habitats. Benthic invertebrates are less mobile than
many other groups of organisms and are of a size which makes them easily
collectable. Biological sampling is a useful supplement to physical and
chemical studies because chemical studies cannot detect possible fluctuations
in water quality between sampling periods. Short-term critical events may
often be missed. The biota, however, reflect both long and short term condi-
tions. Since most species in a macroinvertebrate community have life cycles
of a year or more, the effects of a short term pollution will generally be
210
-------
revealed in the next biological sample. Chemical sampling for all potential
pollutants can fail to evaluate possible synergistic and antagonistic effects.
Macroinvertebrate sampling can integrate all forms of environmental stress
and summarize their cumulative effects.
Nonpoint source pollutants can degrade water quality in a number of
ways. The various effects of these pollutants on benthic communities has
been summarized by Hocutt in Figure 7 (9). In clean, unstressed aquatic
communities both density (N) and species richness (S) will have high
numerical values (quadrant II). Toxic stress (including pesticides and
metals) will reduce both S and N (quadrant III). Nutrient enrichment tends
to increase the number of organisms (N) while selectively favoring a few
types of organisms tolerant to this form of pollution (quadrant I). Moderate
sediment stress is characterized by a decrease in density (due to habitat
destruction) with little effect on species richness (quadrant IV).
Total Number
of Organisms
(N)
I: Selective Stress
Ex: Nutrient
Enrichment
S
t
III: Toxic Stress
Ex: Pesticides
II: Healthy Stream
Ex: "Control"
Stations
IV: Nonselective
Stress or Habitat
Reduction
Ex: Moderate
Sedimentation
Total Number of Taxa (S)
Figure 7. General Effects of Pollutants on Aquatic Macroinvertebrates.
211
-------
Toxic pollutants have the most dramatic effect on stream benthos. Urban
runoff often has associated heavy metals and pesticides, both extremely toxic
to most stream benthos. Sediment along with nutrients in the form of sewage
are also degrading components of urban runoff.
BIOLOGICAL SAMPLING PROCEDURES
Samples were collected using the "kick" method. A net is placed in the
stream, while an upstream area of one square meter is physically disrupted.
This method was chosen because it gives consistent results over a variety
of substrates as compared with the surber and artificial substrates (10,11)„
At each station a minimum of two samples were collected. Three samples
in transect were taken on the larger streams. Samples were preserved in the
field with ethanol. Physical parameters including DO, temperature, pH, and
flow measurements were also taken in the field. Physical charcteristics
such as substrate composition, percent bank cover, and surrounding land use
were estimated during each sampling period. Samples were returned to the lab
where they are sorted and identified to the lowest taxonomic level possible.
All data were then evaluated using a three level analysis system.
level I; Single Number Summaries: Single number summaries include the
Shannon-Weiner diversity index, the N.C. regional biotic index, percent re-
duction of density (N), and percent reduction in species richness (S). The
Sharmon-Weiner index is a value that represents the relationship between the
species richness and the total number of each species in a sample. High
diversity is characterized by many different taxa with few numbers within
each taxa. Relatively few taxa denotes a lower diversity. .Unstressed com-
munities tend to have the highest diversity values.
Biotic Index values are obtained by assigning an index value to each
species collected (ranging from 0-5). 0 is assigned for very intolerant
organisms and 5 is assigned when high tolerance is characteristic. A biotic
index value for each sample is then obtained by the following formula,
B.I. =
aini
N
where;
aj_ is the index value for the i^1 species,
nj_ is the density for the 1th species,
N is the total density for the sample.
Level II: Species Density and Richness;
Density and species richness
This data is then used in
data is tabulated for major taxonomic groups.
comparison with control data to assess reduction and elimination of sensitive
taxa. Changes or shifts in the aquatic community may help to reveal the
nature of pollutants.
Level III; Summaries by Species: Data at the species level can often
be used to deduce the nature of pollutants. Information on the pollution
tolerance of any given taxa when taken into consideration with the distribu-
tion of that organism often parellels any chemical and physical changes in
the environment.
Guidelines shown in Table 4 were used to evaluate Level' I and II data.
212
-------
TABLE 4. LEVEL I AND II ANALYSIS CRITERIA
IN RELATION TO CONTROLS (12)*
A. Reduction in Density (N)
1. 30% : Unstressed
2. 30% - 60% : Moderate Stress
3. 60% : Severe Stress
B. Reduction in Species (or Taxa) Richness (S)
1. 20% : Unstressed
2. 20% - 50% : Moderate Stress
3. 50% ;: Severe Stress
C. Diversity ** (from 13)
1. 3 : Clean
2. 2 - 3 : Moderate Stress
3. 2 : Severe Stress
D. Biotic Index (from 14)
1. 1.8 : Excellent
2. 1.8 - 2.3 : Good
3. 2.3 - 3.0 : Fair
4. 3.0 - 3.8 : Poor
5. 3.8 ; Very Poor
* Control stations or standard control data sets.
** A low diversity value usually reflects stress, but the
converse is not always true.
In some geographical regions of North Carolina, it is virtually impossible to
find an unimpacted site upstream to use as a control. To alleviate the pro-
blem, it is necessary to use a control data set which has been developed
for each of the three geographical regions of the state (Table 5). This
data set serves in the absence of a suitable control in the study area. The
control data set is also used to verify an upstream control site. The con-
trol data set concept is a relatively new procedure which has given -consis-
tent results.
213
-------
r
TABLE 5. SELECTED DATA FROM
UNSTRESSED STREAMS .AND RIVERS
A. Average * Taxa Richness (by Group)
Group
Epheneroptera
Plecoptera
Trichoptera
Diptera
Coleoptera
Other
Total
Mountain
Streams (4)
9.8
6.8
7.9
13.8
3.1
3.9
45.3
Mountain
Rivers (4)
-14.8
6.7
9.1
19.2
3.4
1.9
Piedmont
Streams (3)
55.1
43.9
B. Density by Group (% of Total Density)
Group
Ephemeroptera
Plecoptera
Trichoptera
Diptera
Coleoptera
Other
Total
Mountain
Streams (4)
43
14
18
13
7
5
100
Mountain
Rivers (4)
47
8
26
. . . 12
6
1
100
Piedmont
Streams (2)**
23
9
13
32
20
3
100
*Data represent the expected values for a single collection date
combining all replicates. Streams with 2 replicates - Rivers
with 3 replicates.
**Highly variable.
BIOLOGICAL SAMPLING RESULTS
Level I Analysis: Data from each of the urban streams are compared with
control data in Figures 10-13. Sixty to ninety-five percent reductions in
density (N) indicating severe biotic stress.levels according to established
criteria (12) were found at all streams, with the exception of Sweeten Creek
Station 2. The stream at this station receives high levels of organic
pollutants, which are responsible for a 186% increase.in density (15).
These density reduction results are shown in Figure 8. Density varied
drastically among sampling periods. High density populations found following
the influx of organic pollutants dropped drastically soon afterwards, indi-
cating possible toxic pollutants.
214
-------
o\o
10
20
30
40
50
60
70
80
90
100
w
to
w
Sj -P
CO CO
*
: ^4
g $
O CD
QJ- '
$Q*
wfi-
us-
-H
ft
i—i
#
H
: H
CO
CO S
tils
s
H
8£
-$
CM
5-6
.ffl M
tnPQ
•rn
ft
Figure 8. Analysis of Macroinvertebrate Communities. Level I Analysis:
Density Percent Reduction (N) . . •
Reduction in taxa richness (S) is shown in Figure '9, and illustrates
severe stress at all urban stream sites. Severe stress is indicated when
there is a 50 percent or greater reduction in taxa richness. All urban
sites in this study were found to have reductions of 65-95 percent.
*denotes a 186% avg. increase in density which is highly variable (15) .
215
-------
0
10
20
30
8 40
B so
60
7O
80
90
100
CO
c/>
CD W
H W
II)
I
H
H H
CO
•y
8
iu
2JM
•U O d)
H
-------
8
w
s
M
a w
Figvire 10. Analysis of Macroinvertebrate Communities. Level I Analysis:
Average Diversity Index Values.
Average Biotic Index values were also calculated for all stations. All
stations sampled, except the Sweeten Creek control site had Biotic Index
values ranging from 3.4 to 4.3 (Figure 11). This range suggests poor to
very poor water quality due to the dominance of pollution-tolerant taxa.
217
-------
0)
H
§ H a
••6 JB-6
Figure 11. Analysis of Macroinvertebrate Communities. Level I Analysis:
Biotic Index Values.
Level II Analysis; In analyzing the data from the three study areas, it
is apparent that there is a sharp decrease in the number of sensitive macro-
invertebrate groups present. Sweeten Creek station 1 (Control) is dominated
by Ephyetreroptera, Plecoptera, Trichoptera, and Coleoptera; a typical clean
water assembledge. However, station 2, the urban station, is dominated by
Oligochaeta and Chironomidae, the most tolerant groups of organisms. Pied-
mont streams generally consist of a population in which Trichoptera, Plecop-
tera, and Coleoptera comprise approximately 40 percent of the sample at 65
percent of the density. These groups are not only reduced in all the urban
study areas, but are virtually absent. Instead,90 to 100 percent of the
organisms are Oligochaeta and Chironomidae at these stations. Tables 6 and 7
show definite commmity shifts when observed urban station communities are
cortpared to control station communities. Such extensive shifts in populations
218
-------
appear to be characteristic of all urban streams in North Carolina. There is
not only a shift in the number of types of organisms found but also a density
shift favoring the more tolerant taxa.
Ephemsrdptera
Plecoptera
Trichoptera
Coleoptera
Diptera
Oligochaeta
Other
Total
9.8
6.8
7.9
3.1
13.8
•
3.9
45.3
i
M i — 1
•P O
W S-l
.p
tS CJ
-P
c
1
4.8
3.0
5.8
3.0
9.2
1.0
3.8
30.6
QJ
o 'o
rj 4J
W C
m u
0.4
0.4
-
4.6
2.6
.0.4
8.4
Q)
CD
s
I'
CD
I
7
2
3
16
4
9
43
1
.p
i
•H
.9
.8
.4
.2
.3
.3
.9
,-1
jj
'8
0.8
0.2
0.4
0.4
8.2
2.6
2.0
14.6
i^|
0)
.§
. iH
' .-_.•
0.2
0.2
6o2
4.2
2.6
13.4
H
H
Ky|
«
s
cH
,5
0.'2
—
0.2
—
2.0
1.6
0.4
4.4
•6
I"
y
" 7.9
2.8
3.4
16.2
4.3
—
9.3
43.9
1
M i— 1
-p o
CO p
F
-H
0.3
—
—
—
11.3
2.5
0.9
15.0
0)
W H
||
•H
0.3
—
5.3
2.5
0.2
8.3
QJ
W H
cu (3
-ffcQ
ft
A.
di
A.
Table 6. Analysis of Macroinvertebrate Communities. Level II Analysis:
Taxa Richness by Group.
Group ,
Ephemeroptera
Plecoptera
Trichoptera
Coleoptera
Diptera
Oligochaeta
Other .
Total
43
14
18
7
13
—
5
100
' ,17
36
27
4
13
1
2
100
23
9
13
20
32
—
3.,
3
2
40
42
13
1
—
21
77
• 1 ,
3
. 4"
63
23
7
23
9
13
20
32
—
3"
—
—
9
90
1
—
—
62
37
, 1
10
89
1
100 100 100 100 100 100 100 100 100 100
•s
•s
fi-
&
m
Table 7. Analysis of Macroinvertebrate Communities.
Taxa Density by Group (% of Total).
Level II Analysis:
219
-------
Level III Analysis: Level III data analyses support conclusions found
in prior sections. Control urban station data from all three study areas
were evaluated at the species level. A marked shift in community composition
between the control and urban stations was evident. The control communities
were found to be dominated by relatively intolerant species from several
orders. These organisms are virtually absent at all urban impacted stations.
It is important to also note that the species that do survive at the urban
sites are very tolerant organisms to most forms of pollution. These organ-
isms typically include the oligocheates Limnodrilus spp. and Limnodrilus
hoffmeisteri and the tolerant Chironomidae (i.e. Chironomus sp. and
Polypedilum illinoense). Table 8 illustrates the ten species that are
characteristic of the urban streams studied. The presence of these organisms
in urban streams as opposed to those found at the control stations is at-
tributed to the extensive water quality pollution resulting from urban
drainage.
TABLE 8. CHARACTERISTIC FAUNA OF URBAN STREAMS
TAXA
Oligochaeta
Limnodrilus hoffmeisteri
L. udekamianus
Ilyodrilustempletoni
Nais sp. (Prob: N. Communis)
Lurobriculidae
Chironomidae
Chironomus sp.
Polypedilum illinoense
Tribelos sp.
Cricotopus spp.
C. sp 5/14
C.' sp 6
C. sp 1
Conchapelopia sp.
COMMENTS
Common
Irregular distribution, usually associated
with high organics
A species often associated with L.
hoffmeisteri
Highly sporadic distribution. High densi-
ties usually found only in spring, espe-
cially under low flow conditions.
Common, but rarely with high density
Rarely present in the high density expected
for simple organic pollution.
Common. The most tolerant species in the
genus Polypedilum.
Occasionally abundant
Common
C_. tremulus gr, heavy crenulations of
mandible
C. tremulus gr, crenulations largely absent
C_. bicinctus gr
Common
220
-------
CONCLUSIONS
1) Sources of pollutants entering urban streams are very complex and
include urban washoff, atmospheric washout and fallout, ineffective sanitary-
sewers, and small, intermittent or unpermitted point source discharges. All
of these sources must be considered as to their relative importance in the
development of management programs that will effectively improve the quality
of urban streams.
2) Physical/chemical sampling of urban streams has revelaed numerous
water quality problems. Several pollutant parameters were found to be in
excess of proposed water quality standards, and these high concentrations
occasionally occurred under low flow as well as high flow conditions.
3) The collection of stream samples on a periodic basis and a com-
parison of the concentrations observed for selected pollutants versus adopted
water quality standards may be a useful tool in expressing allowable non-
point discharge limits in conjunction with nonpoint source permitting pro-
grams if they are instituted.
4) Residential and Central Business District land uses within urban
areas have distinctly different pollutant loading characteristics, and
require different management approaches. It is apparent that restoration of
water quality in inner-city streams will require considerably more effort
than needed for streams draining residential areas.
5) Pollutants settled out in pipes or stream pools under periods of low
flow may be resuspended under subsequent storm flow conditions. Such benthic
resuspension may result in highly polluted plug flows leading to severe
downstream biological stress.
6) The sampling of benthic macroinvertebrate organisms can provide a
valuable tool to be used in conjunction with established physical/chemical
sampling methodologies for assessing the impacts of urban stontwater runoff.
7) All of the urban streams monitored exhibited extensive water quality
degradation. Observations on other urban streams in North Carolina have
shown similar water quality problems, and it is apparent that under present
conditions, almost all urban streams will be unable to meet the established
1983 water quality goals.
REFERENCES
1. Hyman, D.N. Urbanization in North Carolina. Economics Information
Report No. 38, North Carolina State University, Raleigh, North
Carolina, 1974.
2. Clay, J.W., et al. North Carolina Atlas, Portrait of a Changing
Southern State. University of North Carolina Press, Chapel Hill,
North Carolina, 1975.
3. McPherson, M.B. Challenges in Urban Runoff Control. Paper presented at
the 14th Annual Henry M. Shaw Lecture in Civil Engineering, Raleigh,
North Carolina, March 13, 1977.
221
-------
4. Grizzard, T.J., J.P. Hartigan, C.W. Randall, J.I. Kim, J.T. Smullen, and
M. Dere Wianka. Assessing Runoff Pollution Loadings for 208 Planning
Programs. In: Proceedings of the ASCE National Environmental
Engineering Conference, Nashville, Tennessee, 1977.
5. Colston, N.V., Jr. Characterization and Treatment of Urban Land Runoff.
EPA - 670/2 = 74 - 096, U.S. Environmental Protection Agency, Cincinnati,
Ohio, 1974.
6. Triangle J Regional Council of Governments. Pollution Source Analysis.
Research Triangle Park, North Carolina, 1976.
7. Division of Environmental Management, N.C. Department of Natural
Resources and Community Development. Proposed Revisions to North
Carolina's Water Quality Standards, Draft #6. Raleigh, North Carolina,
March 1, 1978.
8. Land-of-Sky Regional Council of Governments. 208 Areawide Water Quality
Management Plan, Asheville, North Carolina, 1978.
9. Hbcutt, C.H. Assessment of a Stressed Macroinvertebrate Community.
Water Resources Bulletin, 11 (4): 820-835, 1975.
10. Kinney, W.L., V.E. Pollard, and C.E. Hornig. Comparisons of Macroinver-
tebrate Samplers as They Apply to the Streams of Semi-Arid Regions.
In: Proceedings of the Joint Conference on Sensing of Environmental
Pollutants, 1978.
11. Armitage, P.O. Downstream Changes in the Composition, Numbers and
Biomass of Bottom Fauna in the Trees Below Cow Green Reservoir and in an
Unregulated Tributary Maize Beck, in the First Five Years After Impound-
ment. Hydrobiologia, 58: 145-156, 1978.
12. Lonat, D.R., D.L. Penrose, and K.W. Eagleson, Biological Evaluation of
Non-Point Source Pollutants in North Carolina Streams and Rivers.
Biological Series No. 102, North Carolina Department of Natural Resources
and Community Development, Raleigh, North Carolina, 1979.
13. Wilhrn, J.L. Range of Diversity of Index in Benthic Macroinvertebrate
Populations. Journal Water Pollution Control Federation 42: R221, 1970.
14. Hilsenhoff, W.L. Use of Arthopods to Evaluate Water Quality of Streams,
Technical Bulletin No. 100, Department of Natural Resources, Madison,
Wisconsin, 1977.
15. Duda, A.M., D.R. Lenat, and D.L. Penrose, Water Quality Degradation in
Urban Streams of the Southeast: Will Non-Point Source Controls Make
any Difference? In: Proceedings of the International Symposium on
Urban Storm Runoff, Lexington, Kentucky, 1979. pp, 151-159.
222
-------
DISSOLVED OXYGEN IMPACT
FROM URBAN STORM RUNOFF
by
Thomas N. Keefer, Robert K. Simons, and
Raul S. McQuivey
The Sutron Corporation
1925 North Lynn Street
Suite 700
Arlington, Virginia 22209,
ABSTRACT
Records from 104 water quality monitoring sites throughout
the country were.considered to determine whether,a correlation
exists between strength of dissolved oxygen (DO) deficit and the
presence of rainfall and/or storm runoff in and downstream of
urban areas. .
Sites were screened from more than 1000, mon.i tors maintained
by federal and state agencies such as the U.S. Geological Survey
(USGS), Environmental Protection Agency (EPA), Ohio River Valley
Sanitation Commission, and Wisconsin Department of Natural Re-
sources. Daily data were obtained and processed for 83 of the
104 candidate sites. Of the 83 monitors considered, 42 percent
or about one monitor in three of the 104 candidates demonstrated
a 60 percent or greater probability of a higher than average DO
deficit occurring at times -of higher-than-average streamflow or
on days with rainfall. This result was obtained by considering
daily data for the entire water years. .Mot all years at any
given station exhibited a 60 percent probability, one to three
years out of five is typical. The DO levels fell to less than
75 percent saturation at most of the sites where 60 percent of
greater probability existed. Levels of 5 mg/1 or less were not
uncommon.
Detailed hourly data analysis was made at 22 of the sites,
with high correlation between flow and DO deficit. Of the 22
monitors examined, 11 would not meet a 5.0-mg/l DO standard; six
of the 11 would not meet the EPA-suggested standard of 2.0 mg/1
for 4 hr. Streeter-Phelps analysis indicated that two addition-
al monitor sites at which hourly data were examined would not
223
-------
have met the EPA standard had they been properly located. An
additional two sites at which hourly data could not be obtained
would also not have met the EPA standard.
In general, the data examined here indicate that 19 percent
of the 104 candidate monitors might not meet a 5.0-mg/l standard
and 15 percent might not meet a 2,0-mg/l standard. Frequency
of violations was not tabulated exactly but appears to be zero
to five times per year at sites with correlations.
The data base from which this study was drawn is not geo-
graphically homogeneous. No conclusions can thus be drawn on
the national scope of either correlations or standards violations.
The frequency of EPA standards violations appears to be low, but
the causes should be identified to prevent an increase.
INTRODUCTION
The research presented here was designed to determine from
existing data bases the behavior of dissolved oxygen (DO) levels
downstream of urban areas after storms. The information pre-
sented is taken from the project final report, Reference (1).
Continuously recorded water quality information was ob-
tained from data banks such as U.S. Geological Survey (USGS) and
the U.S. Environmental Protection Agency's (EPA) STORET system.
Data from other state and local agencies such as the Ohio River
Natural Resources (WDNR), Greater Metropolitan Chicago Sanita-
tion District, and others were also considered when possible.
The data bases were reviewed to obtain a five-year (if possible)
historical picture of daily variations in DO, temperature, and
flow in the receiving waters of the United States at the loca-
tion of all reliable continuous water quality monitoring sta-
tions. Concurrent with this review, rainfall data for the areas
in which the monitoring stations are located were obtained from
the National Climatic Center in Asheville, North Carolina. The
DO and temperature records were examined to find all locations
downstream of urban areas in which DO deficits during the warm-
weather portion of the year were correlated with rainfall and/or
urban runoff events. At these locations, an analysis was made
to determine whether a correlation exists between wet-weather
discharge and DO sag on an hourly basis and how significant the
wet weather impact is.
Existing monitors are not necessarily located to study urban
runoff. The location of the monitoring stations and rain gages
relative to the urbanized areas and the relative sizes of urban
224
-------
and nonurban drainage ateas contributing to the receiving water
were considered in making judgments about urban rainfall/runoff
and DO deficits.
OBTAINING DATA
AVAILABLE DATA BASES.
A number of federal and state agencies maintain water quality
data bases. Those that were used or seriously considered for
use in this study are briefly discussed below. Idiosyncrasies
of the data bases, if any, are mentioned. After describing the
various information banks, the screening procedures used to
select monitor sites for this study will be described.
Office of Water Data Coordination (OWDC)
The OWDC is part of the Water Resources Division of the
USGS in the U.S. Department of the Interior. It compiles a
catalog of water data collection stations throughout the United
States. Water quality and quantity monitors for streams, lakes,
reservoirs, estuaries, and ground water are included. More than
350 non-federal (state and local), agencies are listed, plus all
the relevant federal agencies. The OWDC catalog was a primary
source of information for this study. The OWDC does not collect
or disperse data. It merely keeps track of those who do. Upon
request, it will supply names and addresses of all the agencies
listed as well as a personal contact and telephone.number if
known. The OWDC information, however, is not stored in computer
data banks. In order to locate the continuous monitors, it was
necessary to go through its 21 volumes one page at a time and
decipher the codes.
National Water Data Exchange (NAWDEX)
The NAWDEX is also a program of the Department of the In-
terior administered through the .Water Resources. Division of the
USGS. While OWDC is a cataloging agency, NAWDEX actually deals
in data. When fully operational, NAWDEX will be capable of
searching the complete USGS water data records as well as the
EPA STORET data base.
At the time the present study was undertaken, the NAWDEX
program was only capable of scanning the USGS records. A list
of 45 continuous water quality monitors at which flow data were
available was obtained. It was subsequently determined that
these could have been found in the OWDC catalogs.
225
-------
USGS Automatic Data Processing (ADP) Unit
The USGS ADP Unit is located at the national headquarters
in Reston, Virginia. Its purpose is to provide standardized
processing programs and procedures for the water resources data
collected by USGS. It is the most up-to-date source of infor-
mation on the contents of the USGS water quality and flow data
base. It is also the place to contact to obtain data in digital
format such as cards or magnetic tape. For very reasonable
prices ($35 to $70) any of the daily parameter values stored by
the ADP Unit can be obtained along with excellent instructions
on how to read and translate the cards or tape. The USGS ADP
Unit provided a bulk of the data analyzed in this study.
EPA STORET System
STORET is a computerized water quality data base that was
conceived and initiated in the early 1960s by the U.S. Public
Health Service. Its sole purpose is to allow storage and re-
trieval of water quality information by government users.
More than 1100 water quality monitor files are included in
STORET records. Approximately 200 of these monitors were lo-
cated on lakes, more than 100 in pipes, almost 100 beside the
ocean, and about 50 in intakes and estuaries. This left about
500 monitors located along rivers. Data from all the river sites
below urban areas were transferred to a digital tape for use in
this study.
The water quality data in STORET is in an unusual format.
Daily and hourly observations are mixed together in a continuous
string. The potential user should be aware that STORET contains
no flow information.
State and Local Agencies
For two reasons, little actual contact was made with state
and local agencies during the course of this study. First, the
OWDC catalogs are a definitive source of which state and local
agencies collect water quality data. The very great percentage
of those that do, concentrate on random observation and do not
have continuous data bases. Very few data on flow are available
from such sources. Second, those state agencies that actually
collect continuous data report the data to the STORET data base.
Thus, it was not necessary to contact many places. Two local
agencies that maintain their own data bases are the Wisconsin
Department of Natural Resources and the Metropolitan Sanitation
District of Greater Chicago.
226
-------
Wisconsin Department of Natural Resources—
The WDNR maintains a network of 11 water quality monitors.
It also records flow continuously. For a very reasonable amount,
it will copy all 11 stations of daily data onto a magnetic tape.
Hourly data are also available.
Metropolitan Sanitary District of Greater Chicago— .
The Chicago sanitation district maintains a number of flow-
water quality monitor stations in and around Chicago. Many are
located in canals. Hourly data are available, but not in digital
form. A letter from a government project,, of ficer explaining the
need for the data is required. Because of the manpower and
travel requirements, no data were used from this source in this
study.
Ohio River Valley Water Sanitation Commission (ORSANCO)—
The ORSANCO is an interstate compact agency. It was created
in 1948 by the states bordering the Ohio River_primarily to mon-
itor and prevent pollution along the .river. To achieve this,
ORSANCO maintains a network of water quality monitors, which
reports continuously to the headquarters in Cincinnati, Ohio,
where a complete picture of the quality of the river at any time
is available. L . • ...
At various times, 56 monitor sites have,,been placed on the
river. Half of these have been in continuous operation since
1961; the DO and temperature data from these stations are avail-
able on an hourly basis. However, no flow data are available.
While the data are readily available, there are serious
drawbacks, the most important of which is the data format. Be-,
cause of ORSANCO1s interest in an "overall" look at the basin,
the numbers are not grouped by station. Instead, each tape
record contains data for one day for all stations reporting that
day. The records are random length and structured in such a way
that hundreds of tape reads are required to sort the data by
station. Processing cost was so high for the ORSANCO data that
none were included in this report.
National Climatic Center
Precipitation data for the present study were obtained from
the National Climatic Center in Asheville, North Carolina. For
this study, only two forms of data were used. In the initial
phases, a digital tape containing daily precipitation and other
variables was obtained. This type of data is very costly. The
hourly data for the later phases of the study were obtained from
free copies of Local Climatologic Data, a monthly publication,
obtained from the National Weather Service (NWS) in Silver Spring,
Maryland. These publications contain the hourly precipitation
for the month of issue.
227
-------
weather station were selected for daily correlation analysis.
The number of such sites was slightly greater than 100.
Of 1500 monitors, less than 10 percent were considered.
The most common reason for excluding a monitor was that it was
not in or near an urban area. The vast majority of EPA monitors
(at least those in the STORET files) are at industrial outfalls
or in canals or lakes; this is also true of the USGS and ORSANCO
networks. A small number of monitors was rejected because no1
flow records were available. Others were rejected because an
examination of the daily monitor records indicated no quality
problems.> Finally, some sites were excluded because they were
incorrectly listed as continuous monitors.
DATA ANALYSIS
DAILY CORRELATION ANALYSIS
The daily correlation analysis was conducted in three dis-
tinct phases. First, the daily data from selected sites had to
be acquired and'converted to a standard format and stored in a
computer direct-access file. Second, the stored values had to
be retrieved and processed with suitable correlation programs.
Finally, sites for hourly and detailed analysis were selected.
Each phase will be discussed in turn.
Acquiring and Storing Data
Four major computer programming efforts were required to
complete the first phase of data processing. Each type of taped
data (USGS, NWS, STORET, and WDNR) required different recovery
and processing methods.
USGS Data—
In this study, USGS data were extremely,important. When the
site selection process was complete,. 62 of its monitors were of
interest. USGS flow data were often required to supplement DO
data from other sources. For these reasons, the USGS data were
obtained and processed first. The USGS ADP unit is highly coop-
erative and accustomed to dealing with the public. Data requests
were filled promptly and at reasonable cost. Anyone familiar
with programming can have a program to read the data operated in
one day. Daily data were transferred to direct access files in
USGS format for use in correlation programs.
National Weather Service (NWS) Data—
NWS data are provided on digital magnetic tape as card
images. No great difficulty was encountered in reading the
tapes. Data were converted to USGS format for convenience in
processing.
230
-------
STORET Data—
STOKET data are stored in integer format similar to NWS
data. Retrieval was simplified by programs provided by the
STORET users center. Daily data were sorted form the mixed
files of daily and hourly data and stored in USGS format.
WDNR Data—
The WDNR data were also supplied on digital magnetic tape.
The format, which was similar to NWS data, was converted to USGS
type as the data were transferred to direct-access files.
Data Retrieval and Correlation
The first step in the daily correlation analysis after the
required data were stored was to develop a suitable correlation
technique. The objective of the correlation program development
was to build a program that was a sensitive indicator of the
problem in question. That is, it should give a strong numerical
indication whenever a site was found at which the DO level
dropped with increasing flow.
The flow and DO signals of the type of data typical of daily
analysis are illustrated in Figure 1. They have been plotted to
illustrate the correlation of high flow and low DO at Wilson
Creek. The problem was to detect this correlation mathemati-
cally.
15-
r600
WILSON'S CREEK NEAR
SPRINGFIELD, MO.
WATER YEAR 1975
PERIODS OF LOW
DO AND HIGH
FLOW
DISSOLVED OXYGEN
-400
LU
C3
cc
o
C/3
-200 5
300
DAY OF YEAR
Figure 1. TYPICAL DAILY FLOW AND DO RECORDS ILLUSTRATING
HIGH. DEGREE OF CORRELATION BETWEEN HIGH FLOW AND LOW -DO
Standard statistical cross-correlation proved unworkable for
daily data. The DO signal is more or less random in time with
some long-term (low-frequency) components. The flow is an in-
231
-------
termittent highly skewed signal. Low correlations (0.6 or less)
were obtained at all reasonable time lags for the Wilson Creek
data. This indicated that other methods had to be developed.
The next correlation method used was based on a comparison
of the DO levels on "wet" and "dry" days. Various definitions
of "wet" and "dry" were used with both rainfall and flow data.
The different definitions of "wet" and "dry" are described as
follows. When correlating the DO sag with rainfall, a day was
considered dry when the daily precipitation was zero and wet
when the rainfall was greater than zero. When correlating the
DO sag with flow, a day was considered dry when the discharge
was less than a given percentage of mean annual discharge and
wet when the discharge was greater than the same percentage of
mean annual discharge. The percentage was varied from 50 per-
cent to 400 percent in 50 percent increments in order to deter-
mine the effect of the wet/dry cutoff. A wet/dry cutoff equal
to the mean annual discharge was found to be a reasonable value.
The method was implemented on a single year of data at a
time. First, all the DO deficits were summed for wet days and
for dry days. (DO deficit is the difference between the actual
DO level and the saturation DO level for a given water tempera-
ture.) The number of wet and dry days was also computed. The
average dry (wet) DO sag was then calculated by dividing the sum
of all dry (wet) DO sags by the number of dry (wet) days. The
average annual wet DO sag was then divided by the average annual
dry DO sag. If this ratio was greater than unity, the DO sag was
worse during wet weather. This method proved sufficiently sen-
sitive to indicate a problem at Wilsons Creek and was retained
for use in the analysis. It was not, however, considered the
best method.
The final and most sensitive correlation measure was devel-
oped by viewing the occurrence of a correlation in a probabilis-
tic sense. Two questions were asked: (1) "Is it more likely
that the DO deficit will increase or decrease when the flow
changes or a rainfall event occurs?" and (2) "How does this like-
lihood compare during periods of low and high flow?" The answers
to these questions were found by developing a computer counting
scheme. Figure 2 illustrates what may be thought of as a four-
compartment box. The compartments are labeled A through D and
numbers are assigned to the four compartments by graphing DO
deficit divided by a moving average deficit versus discharge
divided by average discharge. The averaging period was initi-
ally not specified, and various values were tried in order to
find one most suited to this study. Seven days was ultimately
chosen as the best averaging period.
The reasoning behind the graph is as follows. First, some
means was necessary to determine whether the DO deficit on any
given day was "better" or "worse" than on previous days. The
232
-------
"DRY"
"WET"
2.0
H
O
U-
i'.i
O
UJ
O
<
K.
IU
<
CJ
z
O
O
Lu
1U
O
o
h-
tu
tr
cc
r»
o
1.0
0.0
A
0 ©
©
©_
.©
©
.©'
TENDENCY FOR DO DEFICIT
TO INCREASE AS FLOW
INCREASES
"WORSE"
V
BETTER"
0.0
1.0
CURRENT DAY'S DISCHARGE/ !
MOVING AVERAGE DISCHARGE
2.0
Figure 2. BOX OR COMPARTMENT METHOD USED TO ANALYZE DAILY
DATA FOR CORRELATION BETWEEN HIGH FLOW AND LOW DO
method selected compares the value observed "today" with the
average of the values that occurred for the previous seven days.
Thus, a deficit/moving average deficit value greater than 1.0
indicates "worse" and a value less than 1.0 indicates "better."
Points on Figure 2 that lie in the upper half (compartments A
and B) represent days when the water quality was worse, than it
had been for the previous seven days. Points that lie below the
line (compartments C and D) represent days on which the water
quality improved.
Next, some means was required to determine whether the water
quality consistently decreased on days with rainfall or increased
flow. Identifying days with rainfall was easy; identifying days
with increased flow was again done by moving average. The aver-
age daily flow "today" was compared with the average for the
previous seven days. A flow/moving average flow value greater
233
-------
than 1.0 indicates "increased" flow or "wet" and a value less
than 1.0 indicates "decreased" flow or "dry." Points on Figure 2
that lie to the left of center (compartments A and C) represent
days on which the flow decreased. Points that lie to the right
of center (compartments B and D) represent days with increases.
The logic of Figure 2 can now be discussed. For any year
DO and flow or rainfall data up to 358 points (365 - 7 used for
moving average) could be plotted on a graph. By counting the
number of points that fell in each of the four compartments, it
was possible to compare the number of days during which DO levels
decreased and flow increased (or rain fell) (Compartment B) with
the number of days when DO levels increased and flow increased
(Compartment D). If the contents of Compartment B were signifi-
cantly greater than the contents of Compartment D, a correlation
was considered to be established between flow (or rainfall) and
decreased DO.
The probability method was tested on data from Wilsons Creek
near Springfield, Missouri. The method was sensitive enough to
indicate the known correlation at that site. A greater than 70
percent probability of low DO at times of higher than average
(seven-day moving) flow and on days with rainfall was found for
several years of data.
Selecting Sites for Hourly and Detailed Analysis
After determining that the probability method would identify
a site with a known correlation between flow or rainfall events
and DO deficit, it was necessary to set criteria for selecting
sites for more-detailed examination. Many sites exhibited cor-
relation between low flow and DO deficit as well as high flow
and DO deficit. Only those sites at which low DO could clearly
be identified with high flow were desired.
A cutoff level of 60 percent was finally selected. That is,
only sites at which the probability of a greater than seven-day
moving average DO deficit at times of greater than seven-day
moving average flow were chosen for further study. The 60 per-
cent cutoff level can be verified as being statistically valid.
A Chi-squared test was used to demonstrate that a probability of
greater than 60 percent or less than 40 percent is required in
order for nonrandom distributions to exist between two,categories
at the 95 percent confidence level. Thus, for a given station
year, if a 60 percent probability of low DO exists at times of
high flow or on days with rainfall, low DO is significantly
associated with these events.
Consideration was also given to the absolute DO level in
selecting sites. Even though a correlation existed between low
DO and flow or rainfall, a "problem" did not necessarily exist.
234
-------
In general, sites were selected where,DO levels less than 75
percent of saturation were present at times of high flow.
Results of Daily Correlation Analysis
Presentation of the results of the daily analysis is com-
plicated by the volume of numbers involved. Some 104 potential
locations were to be analyzed. Three primary factors reduced
the number to 83. First, data could not be obtained for some
sites; second, 'information in the OWDC catalog was not always
accurate; and finally, the list of 104 sites included a number
of ORSANCO monitors that were not analyzed because of the diffi-
culty with the data. The final list of 83 sites includes 55 US6S
monitors, 17 STORET monitors, and IT WDNR monitors. Periods of
record range from one- to five years;
Daily rainfall correlations were completed a.t most of the:
USGS monitor sites. The prohibitive expense of the National
Climatic Center data discouraged thre completion of the rainfall
correlation at the STORET and''WDNR sites.
From the 83 monitors examined on a daily basis, the follow-
ing results were obtained:
• in 19 USGS stations, low DO correlated with flow;
• in 2 EPA-STORET stations, low DO correlated with
flow; ' .-- •• • ' --- ' •-.•-.•..•,-; ' _ :
: • in 3 WDNR stations, low DO correlated with flow; '
and ' •..•••• ..-:-.
' in 18 USGS stations, low DO correlated with rain-
fall. ' . • :. • •
The stations in the above group were examined closely', and hourly
data were obtained for'further analysis when possible.
DETAILED SITE ANALYSIS / •,...-.-.. . ••• ; ,
General Considerations
Water quality monitor sites at which correlations existed
between rainfall and/or flow and low DO we're identified. The
daily data used to identify sites did not allow a determination
of specific water quality standard violations. Hourly data
analysis was required at sites that showed strong deficits (.DO
levels less than 75 percent saturation). For these reasons, it
was necessary to examine those sites with correlations in greater
detai1.
235
-------
Streeter-Phelps analysis was performed at several selected
sites to determine a given water quality monitor was in a posi-
tion to sense the maximum deficit from urban runoff and whether
observed deficits appeared reasonable based on the best esti-
mates of input loads. In addition to selecting sites for hourly
and Streeter-Phelps analysis, it was desired to form some opinion
on why a correlation might exist between flow/rainfall and low
DO at a given site. Information was needed on travel time from
the urban area to the monitor, distance from the rain gage to
the urban area, and other factors that might influence the DO
tha a monitor "sees."
Analysis Procedure
The first step in the detailed site analysis was the gath-
ering of available maps. The monitor and flow gage and the
weather station, if any, were then located precisely. An infor-
mation sheet that summarized what could be learned from USGS
records, the map, or other sources was next filled out. Infor-
mation sheet data were used to select sites with the simplest
hydraulic conditions for anlaysis by the Streeter-Phelps tech-
nique. Finally, the plots of daily average flow, DO, and rain--
fall values were examined, and periods for hourly analysis were
selected. From 10 to 30 days or more of hourly data were exam-
ined at each site at which the hourly data could be obtained.
The total of all this information was then studied carefully to
see what could be concluded as presented in the results reaction.
Obtaining and Analyzing Detailed Information—
The gathering of information at each site was systematized
by developing a standard form. The first step in filling out
the analysis forms was usually to. assemble USGS 7.5-minute topo-
graphic sheets to form a compelte picture of the particular urban
area in question. Next, the water quality monitor, the stream
gage, and the precipitation station were located by plotting the
reported latitude and longitude of the facilities. This had
been done in the early phases of the study to select monitors
for inclusion in the daily analysis. Many of the maps used,
however, were in the USGS library; more maps were purchased and
brought to Sutron for the detailed analysis work.
The assembled maps were large and cumbersome. Therefore, a
sketch of relevant features was transferred to the site form.
This procedure aided in remembering each one. Any outstanding
features, such as industry or sewage treatment plants, were
noted. Other information gathered from the maps included the
urban drainage area, stream width, the distance from monitor to
stream gage, and distances from the monitor to sewage outfalls.
The urban area was estimated as accurately as possible from the
rose-tinted areas on the maps. Distances were measured by map
scale and dividers.
236
-------
After tabulating the physical features of the site, the
stream flow was examined. The flow and DO records were gathered
and the range of values entered. The plots of DO and discharge
were examined to determine the percentage of missing records.
This examination was combined with the quality estimates in the
agency reports to rate the record. Records were "excellent" if
only one or two weeks a year were missing; a "good" rating went
to stations with one or two months missing; and "poor" rating
went to all others.
While examining the flow records, typical July-August low
flow values were selected for later use in the Streeter-Phelps
analysis (if performed). If rating curves were available, they
were used to determine the depth and velocity, which were also
required for Streeter-Phelps.
The results of the daily correlation analysis were summa-
rized to give quick recognition of the magnitude of the
"problem" at each site. This was omitted in cases where there
was too much information to fit and a note was entered referring
to a separate summary table.
The population figures were obtained from a World Almanac
from the 1970 census and no attempt was made to account for
growth increases. The maps were often more out of data- than the
census. For consistancy, no adjustments were made to "urban"
areas either.
The Streeter-Phelps portion of the site forms was completed
only if the analysis was actually performed. Development of
complex water quality models was not within the scope of work.
For this reason, sites with fairly simple hydraulic conditions
were selected for analysis. Meaningful analysis using the
Streeter-Phelps method required streams with few tributaries and
not too many waste sources. Although only five sites were re-
quired by the contract, 13 sites were used because their use
contributed a worthwhile insight into the location of monitors.
The bulk of the "Streeter-Phelps" theory used in this study was
derived from Thomann's (2) book on water quality management.
The complete procedure is described in Reference (1).
Obtaining and Analyzing Hourly Data—
The USGS, ORSANCO, and WDNR all record hourly data, but only
USGS hourly data were included in this study. Hourly informa-
tion is obtained as a biproduct of USGS's archiving of mean daily
flow. Primary computation sheets or "primaries" are computer
printouts of stage and flow (also DO and temperature) from raw
digital data tapes. The primaries are the only record of hourly
or bihourly data available. The digital tapes, from which the
primaries are made are reused, thus destroying the only digital
form of the data. Each primary sheet is carefully examined by a
hydrologist and periods of erroneous, missing, or shifted data
237
-------
are identified. A second series of programs is then run to
create a file of corrected data. The corrected files are then
averaged daily and the results stored in the USGS backfile library
of daily values.
For this study, the primary sheets were obtained for 22 of
the 30 USGS sites that showed positive correlation with rain or
flow. For various reasons hourly data could not be obtained at
the remaining eight sites. In most cases, the problem was in-
ability to obtain the primary sheets. A number of older stations
still record data on strip charts, which are not converted to
digital form. In a very small number of instances, the District
Chief of a particular state would not release the primaries be-
cuase of alleged deficiencies in the data. Emphasis was placed
on obtaining periods of hourly data during June, July, August,
and September. In addition to the primary sheets, appropriate
rating curves were obtained for use in converting the stage data
to discharge. Primary sheets report hourly stage instead of
hourly flow.
Recovering useful data from the primary sheets was simpli-
fied for this study by developing a flexible plotting and analy-
sis sequence for a desktop calculator. First, a program was
developed that created digital magnetic-tape cartridge files.
Each day of data of a particular kind (stage, DO, rainfall, or
temperature) was entered as a separate tape file. Files were
created only for periods when all four types of data were avail-
able simultaneously. Next, a processor and plotting program was
developed. This program reads the files of data, applies the
rating curve to the stage data to obtain flow, calculates the
saturation DO level, and plots the results, including rainfall.
Hourly data plots from many of the stations with storm flow/
low DO correlations were similar. A one-month period of hourly
data for the Scioto River at Chillicothe, Ohio, is illustrated
in Figure 3. A portion of the plot of 1972 daily .data that con-
tains the month is illustrated in Figure 4. These two plots are
fairly typical of the hourly analysis procedure. The correspon-
dence between decreasing DO and increased flow is evident in
Figure 4. The upper line in the figure is the observed daily
minimum DO level, the lower line is the mean daily discharge.
A decrease in the DO level can be seen in almost all instances
when the flow increases. The period from Day 215 to Day 245 was
selected as typical. Figure 3 provides a highly detailed hourly
look at this same time period. The upper graph in Figure 3
illustrates the behavior of the absolute DO level. Daily fluc-
tuations of approximately 4.0 mg/1 were common before the rain-
fall and subsequent runoff event. The upper line in the DO graph
is the saturation level based on the hourly water temperature.
Periods of supersaturation occurred on Days 3 and 11, and the
DO level fell below 2.0 mg/1 and remained there for three days
beginning at Day 18. This precisely coincides with the peak of
238
-------
20
O
Q
10
SCIOTO R. AT CH1LLICOTHE, OH.
6/1/72 TO 8/30/72
SATURATION DO LEVEL-—
DO LEVEL 1
A
:~—v—/-!
o
X
o
Q
O
> 0
^f "C"
*-- H
cc SI
-DO DEFICIT/ 10, mg/l
PRECIPITATION, inches -
2~ 4~
6.8 10 12 14 15 18 20 22 24 26 28 30
THWE FROM START OF PERIOD, days
Figure 3. ONE MONTH OF HOURLY DATA FOR THE SCIOTO RIVER
AT CHILLICOTHE, OHIO
the flow hydrograph in the lower graph. The lower graph iri
Figure 3 illustrates the behavior of the flow, rainfall, and DO
deficit. The flow has been plotted as discharge divided by
average discharge for the period. This was simply a convenience
to avoid rescaling every graph". The shape of the hydrograph is
perserved, and no useful information is lost. The DO deficit
level is the differency between the upper and lower lines in the
upper graph of Figure 3, This was included in the lower graph
to aid in seeing the coincidence between a change in the behavior
of the DO level and the change in flow. 'Rainfall amounts are
plotted as histograms. The period of greater than average DO
deficit that occurs during the time of higher-than-average flow
can clearly be seen extending from Day 18 to Day 30. There is a
clear correspondence between the rainfall events and the in-
creases in flow.
239
-------
60 -
40 -
ai
CD
O
to
O
0 -
SCIOTO FUVER AT CHILLICOTHE, OH.
1 PERIOD OF HOURLY
ANALYSIS
DO
r 15
_ 10
o
O
Q
. 5
L 0
180
270
300
I I
"TIME, day of year, 1972
Figure 4. HOURLY DATA FOR ONE-MONTH PERIOD,
SCIOTO RIVER AT CHILLICOTHE, OHIO
RESULTS AND CONCLUSIONS
SUMMARY OP RESULTS
Correlation of Low DO with Flow and Rainfall
Daily correlation analysis using USGS, STORET or WDNR flow
or rainfall or both was attempted at 104 stations. Of these/
83 - 53 USGS monitors, 17 STORET monitors, and 11 WDNR monitors -
had sufficient data to produce results.
Of the USGS monitors, 19 exhibited a 60 percent or greater
probability of low DO at times of high flow; 18 exhibited a 60
percent or greater probability of low DO on days with rainfall;
and 8 correlated with both flow and rainfall. A total of 30
stations exhibited either type correlation.
Of the 17 STORET monitors, three exhibited a 60 percent or
greater probability of low DO at times of high flow. Of the 11
WDNR monitors, three exhibited the 60 percent or greater proba-
bility.
240
-------
Out of 100 candidates for analysis, 36 gave positive results
in the correlation analysis, and 17 could not be correlated be-
cause of data problems. Thus, 42 pervent of the monitors suc-
cessfully examined or 36 percent of the likely candidates gave
positive correlation results. For discussion purposes, it is
thus concluded that one monitor in three placed near an .urban
area might indicate lower than average DO at times of storm run-
off.
Magnitude of DO Deficits
Because of the coarse time resolution of the daily data, no
direct conclusions could be drawn about the severity of the DO
deficits that sometimes accompany high flow or rainfall. Exam-
ination of hourly data at sites with strong daily correlation
indicated that water quality violations can occur. Of the 30
USGS sites, 11 clearly indicate a severe DO deficit at times of
high flow. Eleven additional sites clearly indicate an effect
from storm runoff but could not be classified as severe. Several
of these sites were too close to the urban area to detect the
maximum deficit in the Streeter-Phelps sense.
Six of the sites did not have water quality problems, at
least concerning DO. Some depression during storm events would
usually be seen.
Characteristics of DO Variations
Although the 11 sites with severe problems varied widely in
size and physical setting, all demonstrated remarkably similar
hourly data records. The period prior to a storm event is
characterized by fairly large diurnal cycles in the DO level,
with 4 to 5 mg/1 not being unusual. Periods of supersaturation
are often indicated. As the flow increases, the diurnal cycles
in the DO record disappear possibly because increased turbidity
and depth cuts off the sunlight to the aquatic plant life at
times of high flow.
At the time of peak flow, deficit levels at least equal to
and occasionally 40 to 50 percent higher than the peak diurnal
cycle value are reached.
The effect of the storm flow on the DO level lasts quite a
long time. This long-term effect gives added validity to the
daily analysis procedure. Sites at which a single hydrograph
peak was examined usually recovered in 3 to 5 days.
Comparison of Storm Flow Deficits and
Water Quality Standards
When measured against existing stream quality standards,
the DO deficits accompanying storm flow may cause water quality
violations. Hourly data at monitor sites with strong correla-
241
-------
tion consistently show deficits below 5.0 mg/1 extending over
several days. The Streeter-Phelps analysis of ten sites, in-
cluding several with strong correlations, consistently indicated
that monitors are not ideally placed to sense maximum effects.
In many cases, deficits 30 to 50 percent stronger could theoreti-
cally be found.
Storm-flow-related deficits at some sites consistently vio-
late 2.0 mg/1 4-hour standards. In fairness, it must be pointed
out that at these sites the water quality is marginal at all
times. Storm events merely push the level down further. Gen-
eral improvements in water quality at all the critical sites
would help alleviate the problem.
National Distribution of Storm Runoff
Related DO Problems
The question of whether DO deficits caused by urban runoff
is a national problem is difficult to answer. One item to con-
sider is the geographic coverage of this study. The data base
maintained by the USGS at its Reston headquarters contains
records for 150 water quality monitors. These monitors are
located in 30 of the 48 conterminous states for 63 percent
coverage. The distribution of states containing monitors is
highly nonuniform. Thirteen states are east of the Mississippi
River and 17 are west. Only 47 of the 150 monitors are in west-
ern states. , If the dividing line between East and West is con-
sidered to b"e along the. western boundaries of Louisiana, Arkan-
sas, Iowa, and Minnesota, the distribution becomes even more
unbalanced. Only 19 monitors are then located in the West.
The distribution of monitors by state is highly nonuniform
also. Ohio alone has 32 monitors, followed by New Jersey with
13 and Louisiana with 11.
The existing monitor network is probably inadequate to de-
fine a national problem. The probability is approximately one
in three of detecting a correlation between flow and DO deficit.
Many states with significant urban areas on streams have no
monitor records at all in the major data banks.
SUMMARY OF CONCLUSIONS
The following conclusions were reached from this study:
• The probability is approximately one in three
that analysis of data from a currently existing
water quality monitor in or near an urban area
will show a correlation between high flow and/
or rainfall events, and high DO deficit.
242
-------
In close examination of sites that exhibited
daily 'correlation between flow arid low DO or
both rainfall and flow and low DO, strong
visual evidence was usually found of the cor-
relation in hourly data records. Sites that
exhibited daily correlation with rainfall
seldom showed any visible evidence of correla-
tion in hourly data records. Flow was judged
the better correlation parameter because it is
a direct rather than a secondary indicator of
stream conditions.
At stations where correlation exists, maximum
deficits observed during periods of high flow
are equal to or 10 to 15 percent greater than
the maximum deficits observed during diurnal
cycles at times of steady low flow.
At locations where water quality is already
marginal (e;g.., 5-7 mg/1 DO) , a storm event
can result in DO levels less than 4-5 mg/1
for periods of several days or longer.
Occasional violations of 20 mg/1 4-hour
standards occur. , •.•••-...-
f •-•
Most water quality monitors are located too
close to the .urban areas to detect the -maximum
possible DO deficit in a, Streeter-Ehelps sense.
Flow and DO deficit are correlated in a wide
vairety of urban situations ranging J:rom towns
of 20,000 population to urban megalopolises of
greater than 1 million population.
The absence of a correlation between flow and
DO deficit does not correlate well with the
percentage of contributing urban areas in the
drainage basin. There .are exceptions in ex-
treme cases where 50 percent or more of the
contributing area is urban.
There is no reliable way to extend the results
of this study to an estimate of total effective
stream miles nationally. No government agency
currently published data on the distribution
of population along major rivers.
The probability of a high DO deficit occurring
at times of high flow is not demonstratably
greater during the summer months.
243
-------
In most instances, the exact cause of the in-
crease in DO deficit is not obvious. In some
cases, there are heavy concentrations of in-
dustry along the stream channel. In other
cases, a sewage treatment plant is close up-
stream. Reintrainment of benthic material is
a likely cause in such locations. In other
instances, there is no industry or sewage
treatment shown on the site maps. Here, the
problem may be strictly due to biochemical
oxygen demand (BOD) of urban runoff, a chemi-
cal oxygen demand, or other problems.
Violations of water quality standards appear
to be infrequent. The data examined here do
not seem to indicate a need for panic programs
to treat urban runoff. Rather, further re-
search is in order to define and alleviate
the causes of occasional problems.
SECTION 5
REFERENCES
1. Keefer, Thomas N.;. Simons, Robert K.; and McQuivey, R.S.,
"Dissolved Oxygen Impact From Urban Storm Runoff," Final
Report, EPA Contract No. 68-03-2630, March 1979.
2. Thomann, Robert V., "Systems Analysis and Water Quality
Management," Environmental Research Applications, Inc.,
New York, New York, 1972, p. 13.
244
-------
THE IMPACT OF COMBINED SEWER OVERFLOWS ON THE
DISSOLVED OXYGEN CONCENTRATION OF A SMALL STREAM
by
Thorkild Hvitved-Jacobsen
Department of Environmental Engineering
University of Aalborg, Denmark
ABSTRACT
Investigations were run in order to determine the impact of combined
sewer overflows on the dissolved oxygen concentration of a small stream. The
different oxygen consumption processes in the receiving stream were studied at
4 consecutive stations during and after the passage of the discharged polluted
volume.
Two different effects on the DO-concentrations were observed in the
stream:
1. An immediate effect caused by degradation of mainly the soluble
BOD-fractipn in the water body and by direct absorption of organic
* ,matter by the animals (invertebrates) in the bottom.
2. A delayed effect caused by degradation of the adsorbed colloidal
and outsettling particulate matter. After passage of the dis-
charged polluted volume there is still observed an effect on the
DO-concentration in the stream. The delayed degradation may in-
crease the respiration of the bottom about 100%. This delayed
effect may last 12-24 hours after the discharge event.
The investigation shows that the oxygen consumption of the organic matter
input to the stream by a storm overflow due to the delayed effect is spread
over a long period compared to time of passage of the polluted volume. This
effect appears to be essential for evaluation and design of sewage systems
within urban areas especially by discharge of run-off water with high content
of BOD.
INTRODUCTION
During the recent years municipal raw wastewater in Denmark has been in-
creasingly treated with special respect to dry-weather situations. Since 1974
these efforts have been supported by the environmental protection act, which
states the environmental considerations to maintain a comprehensive flora and
fauna.
Under these circumstances the relative effect of storm overflows from
245
-------
urban areas becoming a source of pollution of the aquatic recipients will in-
crease. In Denmark discharge of stormwater from combined sewer-systems to
streams is a special, but very important case. The problem is often that a
town of 1,000-10,000 inhabitants must discharge municipal wastewater and
stormwater to a small stream with a flow of 50-500 Si/sec. During intensive
storms this dry-weather flow may, for short periods, be increased several hun-
dred percent.
Upon this background an investigation dealing with the impact of run-off
water from combined sewers on receiving streams is being sponsored by the
Danish Association of Civil, Mechanical, Electrical and Chemical Engineers
and financially supported by the National Agency of Environmental Protection,
Copenhagen. The hydraulic, physical -chemical and biochemical parts of this
investigation are carried out by the University of Aalborg while the biologi-
cal part is executed by the Freshwater Laboratory of the Agency of Environ-
mental Protection. These investigations are taking place from 1978 to 1981 .
A provisional report was prepared /!/.
The most important part of these investigations deals with the impact of
the receiving stream. The reason why the DO-concentration is of special in-
terest is that this parameter is of great importance for the water quality of
a stream. Furthermore the discharged water from combined sewers may be a
potential source for undesirable impact on the DO-concentration in the stream.
Durin.g the last years in Denmark, as well as abroad, several investiga-
tions dealing with the generation of the stormwater together with nature and
contents of pollutants in the run-off water have been carried out. Compare
as an example survey articles by Harremoes (1977) /2/ and Johansen (1974) /3/.
i
However, it can be ascertained that only ve'ry few investigations deal
with the impact of the storm events on the receiving water bodies, Homer et
al. (1978(
The interest of the National Agency of Environmental Protection in having
a better understanding of the impact on the receiving streams during storm
events is caused by the fact that the actual practice for discharge of run-
off water seems to ignore now-a-days demand tp water quality in the recipients.
This practice mainly considers a hydraulic protection of the sewage treatment
plant for continuous optimal function and not a conscious protection of the
recipient. The aim is to have this practice replaced by instructive lines
for protection of recipients in connection with discharge of stormwater. The
performed investigation has to be seen as a step tending towards obtaining
the necessary knowledge of the impact on the recipients in order to reach
this aim.
INVESTIGATED AREA
The investigation was carried out in Skravad Brook located in the middle
of Jutland, Denmark (fig. 1).
The stream receives, about 5.5 km (3.4 miles) upstream the confluence of
246
-------
Town of Moldrup
Moldrup
sewage
treatment
' 0 0.5 1mile
Fig. 1: Map of the investigated area, Skravad
Brook.
247
-------
Skravad Brook and River Skals 8, biologically treated wastewater from Moldrup
sewage treatment plant. From the small town of Moldrup which has a combined
sewerage system and covers a reduced area of about 15 ha (37 acres) the over-
flow water is discharged through a ditch to Skravad Brook just downstream the
outlet from the sewage treatment plant (fig. 2).
overflow | overflow |
f
ditch
Skravad Brook
Farebro
Bro
Lund Brook
Bjerregrav Bro
—River Skals A —--
120m
20m
310m
1230m
2500m
1500m
Fig. 2: Schematic representation of the investigation area in Skravad
Brook. Stations for sampling and DO-concentration measurements
in inlet and recipient are stated.
Moldrup town has about 1,000 inhabitants and includes a few industries,
among them a dairy and a small slaughter-house.
248
-------
Upstream the outlet from the sewage treatment plant of Moldrup, Skravad
Brook is influenced by drain and surface run-off from mainly rural areas. The
relatively moderate influence on the stream upstream as well as downstream the
wastewater treatment plant was an important argument for the choice of Skravad
Brook as investigation area.
The investigation was started in autumn 1978. At this time the following
conditions were characteristic for the stream (station 5, 2.10.78):
Rate of stream reaeration K9(20) = 7 day~
21
Total respiration R(tot) = 2.5 gOg'm day
Production P = 0.3 -"-
These conditions cause little variation in DO-concentration throughout
day and night and a relatively high level of DO-concentration. In that case
the DO-concentration is 9-10 mg/£ throughout day and night with a DO-satura-
tion concentration of 11.0-12.5 mg/£. Even though the stream is influenced
by treated wastewater (BODg in the stream is 2-4 mgA during dry-weather con-
ditions) it is possible to maintain favorable oxygen conditions mainly because
of a slope of bottom of about 2% and a daily meanflow at station 5 of about
160 a/sec, (autumn 1978). Apart from periods with extreme growth of aquatic
plants in the summer period it is possible to maintain a DO-concentration
throughout day and night exceeding 6 mgA.
With respect to the oxygen conditions the investigated stream can thus be
characterized as favorable for maintenance of a comprehensive flora and fauna.
Salmon and trout are thus often observed in the stream. The impact of the
run-off water on the DO-concentration in the stream is thus expected to be re-
flected on the background of the recently discussed oxygen conditions during
dry-weather situations.
OXYGEN CONDITIONS DURING DISCHARGE OF RUN-OFF WATER
From a theoretical point of view the classical Streeter-Phelps model /?/
is, during steady-state conditions, able to describe the deoxygenation and
physical reaeration in streams. Furthermore it is possible to develop the
model by considering more and better description of these processes. During
dry-weather periods it is often possible to maintain approximately steady-
state conditions, and removal of organic matter from the water body will be
equivalent to the oxygen demand whether the degradation takes place in the
water body or at the bottom. Concerning the deoxygenation at the bottom, it
is naturally not the same organic matter, which is actually removed from the
water as at the same time interval is degradated at the bottom. As the
Streeter-Phelps model is a simple oxygen balance these conditions are left out
of consideration and are for the steady-state situations without interest.
During non-steady-state situations, for instance run-off events, concen-
tration of organic matter will be increased in the stream. Removal of organic
matter from the water body will be increased too and exceed the actual degra-
dation at the bottom. This fact may result in a later increased degradation
249
-------
at the bottom and a delayed effect on the DO-concentration which may affect
the stream after passage of the discharged stormwater volume. Non-steady-
state condition thus requires distinguishing between an immediate and a de-
layed effect on the DO-concentration.
For river systems, influenced by effluents characterized by a BOD-content,
the DO-concentration in the stream is of special interest. In connection with
combined sewer overflow, change in the DO-concentration in the stream is in-
fluenced by:
I. The resulting DO-concentration obtained immediately after mixing
of combined sewer overflow and water from the stream.
II. The combined sewer overflow impact on the DO-influencing processes
in the stream.
re I. The following phenomena will influence the DO-concentration after
mixing storm sewage and river water:
a) DO-concentration in mixed stormwater and raw wastewater before
discharge as a result of DO-concentrations in these two compo-
nents.
b) Flow of combined sewer overflow in relation to flow in the
stream. This relation is influenced by time of day and night
and time of year.
c) DO-concentration in the river upstream point of discharge.
d) The design of the sewerage system.
re II. The following processes are expected to be mainly responsible for
changes in DO-concentration of the stream:
a) Oxygen demand by degradation of organic matter.
A distinction must take place between:
1) An immediate oxygen demand which is supposed to be due to:
-Degradation in the water body.
-Absorption by animals (invertebrates) in the bottom sediment.
Both processes are due to degradation of mainly dissolved
organic matter.
2) A delayed oxygen demand caused by:
-Adsorption of mainly colloidal organic matter.
-Degradation of settled particulate matter.
Harremoes 1979 /8/.
b) Oxygen demand by nitrification.
c) Photosynthesis and respiration by the aquatic plants.
d) Reaeration from the atmosphere.
2.50
-------
METHODS
The purpose of the investigation was primarily to examine the stream re-
sponse to the combined sewer overflow. The processes mentioned in the previous
section point Ila were therefore of special interest.
It was thus possible to carry out the investigations by means of simula-
ted overflows. The day before the simulated overflow was carried out, river-
water from Skravad Brook and raw wastewater from Mpldrup in proportions 10:1
was pumped into a reservoir of 370 m3 at'the wastewater treatment plant of
Moldrup. On the day of the simulation this synthetic overflow water was let
out into the ditch and from there carried on the Skravad Brook (fig. 2).
The simulations were hydraulically as well as to pollutants carried out
at the locality like real overflow situations. The following data characterize
the simulated discharge on 9.11.1978: ..•••--
The duration of discharge':'. . .
Discharged quantity of water:
Dry-weather flow in ditch:
Maximum flow in the ditch during discharge:
Flow in Skravad Brook just upstream the outlet
from the ditch:
DO-concentration in the ditch during the discharge:
DO-concentration in Skravad Brook just upstream
the outlet from the ditch:
Quantity of BOD in the discharged water:
70 min.
2.15 m3
10 £/sec.
110 Vsec.
120 A/sec.
4.5-6.5 mg/£
10.5 mg/£
15 kg BOD5
It was essential to measure the expected immediate and delayed effects on
the DO-concentration in the stream caused by the overflow.
Throughout the whole period of investigation it was found that the DO-
coneentration curves through day and night in dry-weather situations at the
stations 2-6 had nearly uniform shapes. This proved very profitable for the
evaluation of DO-variations caused by overflow. All, the DO-concentration
curves will thus be identical, partly by displacement of DO-concentration
level, partly by change of time corresponding to the time of travel between
the stations concerned.
In order to measure the effect of discharge on. the DO-concentration in
the stream, the DO-concentration curve from station 2, which can be regarded
as unaffected by the discharge, therefore, by comparison with the DO-concen-
tration curves from stations 3-6, states the impact at these stations during
and after passage of discharged watervolume.
As an example fig. 3 shows the DO-concentration before, during and after
the discharge on 28.9.78. At this figure the solid line shows the measured
DO-concentrations while the broken line (the displaced curve from station 2)
251
-------
•DO-cone., station 5
—* DO-cone., station 2 displaced
1
10
8
6
4
2
k DO -cone.
(mg/t)
r *
t— = —
60min. and 0.7mg/l
u
ic*7
r
1
immediate e
1 — *-^,
ffect
delayed
"""^X— — 3
effect
X-
;••"•
hour
gOO 1QOO ^00 12°° 1300 i^OO ^gOO igOO lyOO igOO igOO Of
Fig. 3: DO-concentration at station 5 before, during and after dis-
charge on 28.9.78 compared to expected DO-concentration if
no discharge took place (displaced curve from station 2)
indicates which DO-concentration may be expected in station 5 if no discharge
took place. The difference between such two curves for each station will thus
at any time state the effect of the discharge on DO-concentration at the
station concerned.
IMMEDIATE EFFECT IN THE STREAM
As earlier mentioned the immediate effect on the DO-concentration in a
stream will take place when the discharged polluted watervolume moves down-
stream. Degradation in the water body as well as absorption at the bottom may
influence the DO-concentration.
Fig. 4 expresses the stream response on a simulated discharge carried out
in the period from 12:00 to 15:30 on 8.11.78. The discharge was hydraulically
moderate as maximum flow in the ditch was 50 &/sec. and only for half an hour
the flow was greater than 20 £/sec. The DO-concentrations at station 3 shows
the effect on the stream as the discharged overflow by outlet from the ditch
is mixed with the stream water. The flow in the stream before discharge was
120 A/sec.
The immediate oxygen consumption which occurs at the 2.5 miles distance
from station 3 to 6 can e.g. be recognized by following the observed two
minima on each of the DO-concentration curves. These minima appear at 13:00
and 14:50 at station 3 and are observed at 15:10 and 17:00 at station 6. Com-
252
-------
/
11
10
o
(mg/l) x— go -cone, station 2 displ
*"°-S
,
V^
sp^A-*
i
>w '
"—«
""""•X >
"**•*— —5
c c
^L><
S— 0"^
aced
-0.1mg/l
hour
^ *
day
11
10
Q
i u\j V.UIK,. u UU-IUMC., ilUUUII *t
(mg/l) x — DO-conc., station 2 displaced 15 min.
-x^-
_ x —
k
N>— c
A?-*
>^M
*--•*—
-•-,,,
-*"^»<
x — .x
^>o-<
x — x._
o '
x— »•-
0 t
X X
0— *
x — x
0 fj
r-—^r
hour
An. r
day
t
12
11
10
9
. -^ u DU-lUlll., blUllOII U
P0"Ji?nc' — "x D°- conc" station 2 displaced
(m9/l> 60min. and 0.6 mg/l
^
i
; — x — •;
s\
y*- \
^
a^-o
f ^^
1 '&--.,
-—o^-^
i~- x— .,
^jf*~
___^««_)
X >
-"•*—.>
i x )
— <^
i— X 3
— O 1
— 0— '
_houi
'of
day
12
11
10
M DO-conc.
(mg/l)
DO-conc., station 6
DO-conc, station 2 displaced
130min and 1.1 mg/l
12°° 14°° 16°° 18°° 20°° 22°° 0°° 2°° 4°° 6°° 8°° 10
hour
day
Fig. 4: DO-concentrations at the stations 3 to 6 before, during
and after a discharge compared to expected DO-concentra-
tion if no discharge took place (displ. curve from st. 2).
253
-------
pared to the expected DO-concentration expressed by displacement of the DO-
curve from station 2, the effect of the immediate oxygen consumption is obser-
ved. In the investigated stream with good atmospheric reaeration it cannot be
expected that a low value of DO-concentration caused by discharge of. combined
sewer overflows can be maintained or developed. The following day on 9.11.78
where a greater decrease in DO-concentration was generated by a simulated dis-
charge, it was observed that there could not be obtained greater difference
between actual and normal expected DO-concentration than about 2 mgA at sta-
tion 6 (fig. 5). These facts were connected with a 6005 mean value in the
stream of about 20 mg/H during passage of the polluted watervolume.
A Max. change in DO
during passage of
discharged volume
(mg/l)
3
0.5
1.5
discharge 9.11.78
discharge 8.11.78,
13°° station 3
discharge 8.11.78,
1 A50, station 3
2.5
station number
•distance from
point of
discharge (miles)
Fig. 5: Maximum difference in DO-concentration between the actual
value during passage of discharged watervolume and normal
expected value if no discharge took place.
Based on measurements of 9.11.78 with la stream temperature of 9°C the
following parameters were calculated during passage of the polluted water-
volume:
Rate of stream deoxygenation
Rate of total BOD-removal in the stream
^ = -0.15 day
Kr = -0.55 day
-1
-1
In smaller streams with values of K-| and Kr as mentioned above and at
the actual reaeration (K2(20) - 7 day~') it does not seem possible to maintain
a difference greater than about 2 mg/& between actual DO-concentration in the
stream during a discharge event as simulated and the normal value if no dis-
254
-------
charge took place. If greater change in DO-concentration in such situations
can be avoided immediately after mixing combined sewer overflow with stream
water this seems to be a very important factor for maintaining sufficient DO
in the stream.
DELAYED EFFECT IN THE STREAM
As mentioned before the delayed effect on the DO-concentration in the
stream must be caused by adsorption and sedimentation of organic matter on the
bottom due to the fact that removal of BOD from the watervolume exceeds actual
degradation during storm events. On fig. 4 this delayed effect can be obser-
ved. At station 3 thus after 4 p.m., at the stations 4, 5 and 6 correspond-
ingly displaced time of travel 15, 60, and 130 minutes. Measurements with
tracer (rhodamin B) indicate that the observed effects on DO-concentration
cannot be due to delay of polluted water in the dead-zone volume of the stream*
Dependent of the actual possibilities for adsorption, sedimentation, desorp-
tion and resuspension the observed delayed effect on the DO-concentration will,
with respect to magnitude and extension in time, be changed from one discharge
event to another. The hydraulic conditions in the stream may be important for
this effect. The delayed effect seen on fig. 4 shows, however, what was ob-
served several times. It is thus typically that the delayed effect had a dur-
ation of about 24 hours, and that the situation then was near the steady-
state situation. ;
The calculated total respiration (Rtot) of the stream immediately before
discharge on 8.11.78 will, due to decreased influence by the sewage treatment
plant, be reduced downstream (fig. 6). If values of Rtot are calculated
immediately after discharge on 8.11.78 and 9.11.78 an increased relative im-
portance is observed at the most distant stations (fig. 6).
A Rtotte02-m-2-day-1)
after discharge 9.11.78
after discharge 8.11.78
before discharge 8.11.78
^•station number
0.5
1
1.5
2.5
• distance from
point of
discharge (miles)
Fig. 6: Total respiration in Skravad Brook before as well as
after discharge on 8.11.78 - 9.11.78.
255
-------
In spite of the fact that the measured changes in the respiration of the
stream are relatively large (^ 100%) the absolute values are, however, of in-
significant influence on the DO-concentration in the described case. The oxy-
gen consumption by degradation of the adsorbed colloidal and settled particu-
late matter is thus spread over such a long period that the effect on DO-con-
centration is reduced.
As mentioned, the actual conditions for adsorption and sedimentation of
discharged organic matter will influence the delayed effect on the DO-concen-
tration. To throw light on this there were set up mass balances for the sec-
tion between stations 3 and 5. The calculations were made for SS and CODpar^,
the latter defined as the difference between the COD of a non-filtered ana a
filtered sample. Fig. 7 shows the processes influencing the SS mass balance.
SS Jnput
station 3
Water body
SS output
station 5
Erosion
Sedimentation
Fig. 7: Mass balance for SS for the section between station 3
and 5.
Based on measurements during the period of discharge on 9.11.78 the
results are as follows:
For SS:
For COD
Erod - sed = Input - Output = 12 kg SS/80 min.
(^35% of mass input at station 3).
part:
Erod - Sed = Input - Output =1.1 kg COD/80 min.
(^ 4% of mass input at station 3).
These results indicate that, due to the increase in water level owing to
the hydraulic conditions, erosion takes place in the stream. The substance
which is eroded has, however, in comparison with what is discharged, a much
lower value of COD (10-50 mg02/g dry matter contrary to about 1000 mg02/g
dry matter). Only unimportant quantities of particulate organic matter are
thus added to the bottom. The delayed effect on the DO-concentration is thus
likely to be due to adsorbed colloidal organic matter at the bottom.
256
-------
RESULTS FROM REAL DISCHARGES
As mentioned, the investigations were mainly concerned with the impact of
simulated discharges. However, current measurements of a few parameters took
place. It was thus demonstrated that the actual sewerage system normally does
not cause great changes in the DO-concentratioh in^the stream immediately
after mixing combined sewer overflows and water from the stream. Results ob-
tained at station 3 for 30 of a total of 66 discharges in the period 30.6.78 -
20.11.78 are shown on fig. 3.
Maximum difference between
actual and steady -state
DO -cone, station 3
11 days of preceding dry weather
by discharge on 20.7.78
50
100
150
200
250
300
Qmax
"Maximum
flow in ditch
Fig. 8: Maximum change in DO-concentration at station 3 compared to maxi-
mum flow in ditch during 30 discharges in the period 30.6.78 -
20.11.78.
257
-------
An arbitrary example of these real discharges that took place on 5.10.78
at 10:00-16:00 with maximum flow 170 a/sec, in the ditch at 13:20 is shown on
fig. 9. As found during simulated discharges an immediate effect as well as
a delayed effect on DO-concentration is observed. This fact does not indicate
any different impact between real and simulated combined sewer overflows on
the DO-concentration of the stream.
DO-conc., station 3
10
8
DO-conc. — x — DO-conc, station 1 displaced
(mg/l) 10min.and 0.6 mg/l
£^
<- .$1*
•>
x— *r£
r**£
)fet»Kx:««r
W-
,' . .•'!".' l|, '||i | .Jl
X f)c— <
"," ' : r L
ir—cx— - <
X—
10
8
i DO-conc.
(mg/l)
o— DO-conc., station 5
x — DO-conc., station 1 displaced
70min. and 0.4 mg/l
hour
of
day
x o-i
A X.
p*>X>~o*.
P^PQ.-
%
00
V*
* f*-'**—,^!
s^
f~— -x — ,
e — x
l-o^-0"
III
'-O
goo goo IQOO 12°° 14°° 16°° 18°° 20°° 22°° 24°° 2°° 4°°
hour
of
day
Fig. 9: DO-concentrations at the stations 3 and 5 before, during and
after discharge on 5.10.78 compared to expected DO-concentra-
tion if no discharge took place (displaced curve from station
1).
CONCLUSIONS
The investigations show that combined sewer overflows to a small stream
result in an immediate as well as a delayed impact on the DO-concentration in
the stream. Due to this fact the degradation of organic matter loading on to
the stream is spread over a longer period of time than just the time length
of discharge of polluted overflow water. This result is obtained from real
as well as simulated overflow situations.
The importance of this effect to water quality will largely depend on
actual possibilities for degradation of BOD in the waterbody and absorption
at the bottom compared to adsorption of colloidal and settling of organic
particulate matter. Actual DO-concentration in the stream after mixing com-
bined sewer overflows with stream water and actual reaeration rate will
affect the stream quality too.
258
-------
REFERENCES
Schaarup-Jensen, K., T. Hvitved-Jacobsen and E. Mortensen: Recipienters
•reaktion pa regnvandsafledning fra et faelleskloakeret byomrade,-
april 1979. • '
Harremoes, P.: Betydningen af forurem'ngen fra regnafstromning for valg
af urbane aflobssystemer. - En oversigt over nordisk litteratur og vur-
dering af status. • Jrettonde nordiska symposiet om vattenforskning,
Roros, 2.5-5.5 1977. Nordforsk publication 1977:2.
3. Johansen, L.: __ Afledning af regnvand, may 1974.
4. Homer, R.W., L.B. Wood and L.R. Wroe: London's stormwater problem,
.Water Pollution Control Federation, Journal, vol. 49, No. 1, 1977.
5. Pitt, R. and R. Field: Water-Quality effects from urban runoff, Ameri-
can Water Works Association, Journal, vol. 69, 1977.
6. Smith, R. and R.6. Eilers: Effect of Stormwater on Stream Dissolved
Oxygen, Journal of the Environmental Engineering Division, Proceedings
of the American Society of Civil Engineers, vol. 104, No. EE4, 1978.
7. Streeter, H.W. and E.B. Phelps: A Study of the Pollution and Natural
Purification of the Ohio River, US Public Health Service, 1925.
8. Harremoes, P.: Udledning af organisk stof i vandlob, Vand 4", 1979.
9. Hvitved-Oacobsen, T.: Et vandlobs iltforhold under pavirkning af
regnvandsafledning fra et fas lleskloakeret byomrade, Vand 4, 1979.
259
-------
Moderator:
Fifth Session
IMPACTS ON ESTUARIES
Raymond P. Canale
University of Michigan
Ann Arbor, Michigan
260
-------
IMPACTS OF INTERMITTENT URBAN DISCHARGES ON RECEIVING WATERS
John L. Mancini
Manhattan College
Bronx, New York
Abstract
The paper will present an illustration of the impacts of CSO discharges
on bathing beach areas of New York City. The illustration will also define
the treatment required and the costs for controlling water quality. A unique
statistical analysis used in the project will be discussed.
The proposed paper will be divided .into two segments.
Segment #1 -will discuss an analysis of coliform distribution in New
York Harbor. The paper will present calculated and observed time variable
coliform distributions over several storm events for the major water bodies
that make up New York Harbor. Data and calculations will be presented for
different size rainfalls and for different Hudson River flows. This segment
of the paper will conclude with economic data on the cost of controlling
coliform levels at selected beach locations in New York Harbor. Therefore,
loads, fate and effects of storm water associated coliform distributions will
be discussed.
Segment #2 - will address the fate and effect of intermittent discharges
of toxics. In particular, information from data in the Trinity River, will
be employed to calculate partition coefficients and equivalent removal rates
for Cd, Cu, and other metals. Comparable calculations for the equivalent
removal of Lindane, DDE, DDT and several other organic toxics will also be
provided. This will be a discussion of fate of toxics.
The effects part of this segment will employ dose response data for Cd
to illustrate calculations of mortality of an organism due to a storm load.
The effect part of the paper will present a procedure for calculations. In
situ data are not available to test the calculation procedure. The issues
of concern in rational evaluation of effects will be identified in a quantita-
tive manner.
Introduction
This paper deals with three interrelated considerations for defining the
impacts of intermittent discharges on receiving waters. Initially a classical
261
-------
approach employing non-steady state water quality modeling for coliform in
New York Harbor is discussed. The problem setting is a complex one due to
the number of sources, the stratified nature of the estuary and the general
scale of the analysis. It is, however, an illustration of the application of
well defined technologies for the analysis of the impacts of intermittent
discharges. The second part of the presentation addresses a technique which
can be considered for defining a wet weather equivalent to the 7 day - 10
year low flow criteria used in many water quality standards. This technique
may also be considered as one method of providing information on the probable
setting for local and regional water quality problems caused by intermittent
discharges.
The third section of the presentation illustrates one method of employ-
ing existing data on biological effects to begin addressing the questions
associated with numerical wet weather criteria which could ultimately be con-
sidered in the context of wet weather standards.
The second and third segments of the presentation deal with the very
initial results of an on-going research effort and are therefore preliminary
in nature. It is hoped that a technical dialogue on wet weather water qual-
ity criteria will be stimulated.
New York Harbor Analysis of Coliform
Figure 1 is a map of the New York metropolitan area. The location of
bathing beaches at which elevated coliform levels have been observed is
shown on the map. One objective of the recent New York 208 Study was to
define the treatment required to protect these beaches during the swimming
season^' . Figure 2 illustrates the segmentation of the water quality model
and the load generation model. The loadings to ,the system consisted of treat-
ment plant discharge, leakage from regulators and discharges from both com-
bined and separate sewered areas. The water quality model was a tidally
averaged non-steady state calculation which included vertical segmentation
and associated analyses to represent the circulation in this stratified
estuary^'. The water quality model was compared to observed chloride and
coliform data collected under seven conditions as illustrated in Table l(3).
The rainfall varied from 9.5 inches to .3 inches while the peak Hudson River
flow ranged from 115,000 cfs to 10,000 cfs.
Figure 3 illustrates the comparisons of observed and calculated chlor-
ide for a period when the flow ranged between 21,000 and 58,000 cfs. Figures
4 and 5 present comparisons of calculated and observed coliform levels for
two surveys considering the Hudson River and the East River. The comparisons
of calculated and observed data were considered adequate to provide a basis
for examining treatment requirements. In general the results of the analysis
indicated that discharges within 4 to 6 miles of the beach required control
and treatment; while those further away from the areas to be protected had
smaller effects on coliform levels at the beach and treatment was not requir-
ed to meet existing coliform standards. These findings provide an opportun-
ity to stage capital and operating expenditures over time obtaining benefits
in a sequential manner directly associated with expenditures. The cost of
262
-------
BEACHES
Fig. 1. Study area map.
263
-------
,
•£;.••"...•:••"'••..•••':•••../ ./ : -' ""-!
'•VV'"- SANDY! /*••. ••'" /.. ATLANTIC OCEAN
""\\":\:':.HOOK
-------
35
30
Q.
Q.
25
>-*
t 20
<
CO
15
10
0
10/17/75 HUDSON RIVER FLOW = 21,000 CFS
- NYC 208 PHA II DATA
50 40 30 20 10 0 -10 -20 -30
MILES
35
10/22/75 HUDSON RIVER FLOW = 58,000 CFS
SURFACE: * DATA
CALCULATED
L BOTTOM: • DATA
25 " -—CALCULATED .
t 20
CO
10
5
0
50 40 30 20 10 0 -10 -20 -30
MILES
Fig. 3. Illustration of salinity verifications, Hudson River
transect.
265
-------
HI
o
o
o
H
o
LU
C3
Q
CC
co
oJ
o
CN
O
CN
LO
0)
O
CN
o
T
o
O C/3
«- UJ
o S
CM
O
CO
13
O
CC
LU
m
LU
in
cc
\-
cc
LU
CC
2
00
I
s
£
c
o
V)
TD
I
o
**—
"o
o
1
•*-•
•s
o
CO CD «* CM
swuojnoo ivioi DOT
CO CO •<* - CN
SIAIdOdHOO 1V1O1 DOT
266
-------
EAST RIVER TRANSECT, UPPER BAY TO LONG ISLAND SOUND
o
co 7
oc R
0 6
LL
=J 5
O
J '
o~ 3
LU
U-
0 2
O
-1 1
•
8
i 7
tr
£ 6
5' 5
O
O
-i. 4
0 '. o
1 1 i O
LL •'
0 2
0
-1 1
^
5/4/77 A<
™
-
- .,'•> *"\ ,» — \
• - Siti.' m * '
_ i^|t * /I>K \
\V'~~\y' 's| Xv.
~ N* * \SvN-
* '""^ ^"~"\* S T^"^^^^
- xv.l " ^--^^ 7\.
K' f "\ i ^^
: : \
'• - • ^
—. * ,
1 1 1 1 1 1 1 1 1
? 3 8 '13 18. 23.. 2
MILES
8
i 7
DC R
0 6
II
^J 5
0
^.4
^
o 3
LU
II
^ 2
cu
O
-1 1
' n
\j .
3
8,
oo 7
1 6
LL
_J 5
O
o ,,
_j'-
~"J '
o 3
LU
LL „
2
O :
-> 1
1
8 -2
5/5/77 B.
--,
- ' ••*" — "'•X
T 1 1 * T H "•-•*»
. |r''":~"j: i^fxi ,\
*'x>"' v" — \ • "\..
— x^-"\ '""^^v^
• NYC 208 COM- \« ^^>
POSITEDATA 'x
DATA UNCERTAINTY
RANGE
— CALCULATED
i i i i i i i i i i i
2 "3 8 13 18 23 2
; MILES :
5/7/77 D-
;• :
- -' >, ' '*
*» .'' •*.'.:
fJf_x*"~T~~ ? \ . v >
' j* * * i***^^^-. ! ' ' *
"?**'•*! *i T^^vT
• *^"m\ * ** ' * • ^^»
- ~'' *' *>v* I^X,
• \% j r ^s.
- . v%
—
i i i i i i i i i i i
,3 8 13 18 23 2£
MILES
Fig. 5. Illustration of fecal Coliform verifications. East River transect.
267
-------
controls, even on this scale, are substantial and could represent 200 million
dollars. The economic feasibility of the individual components is still
under discussion.
In summary, a well defined technology base exists to analyze receiving
water impacts associated with intermittent discharges from combined sewer
overflows and urban runoff.
TABLE 1
VERIFICATION CONDITIONS FOR THE COLIFORM ANALYSIS
IN NEW YORK HARBOR
Condition
1
2
3
U
5
6
7
Rainfall*
(ins)
9.5
2.6
2.2
0.3
5.1
Peak Hudson River Flow
(cfs)
71,000
81,000
25,000
115,000
2^,000
10,000
11,000
Total average rainfall from all operating gauges.
Wet Weather Frequency of Occurrence Analysis
Treatment of all runoff from urban areas would be prohibitively expen-
sive and the water quality benefits have not been assessed. One method of
beginning to address this issue is to examine wet weather criteria which
could eventually be considered for use in establishing wet weather standards.
One of the primary issues relates to the degree of environmental protection
which is required or can be supported by available money and technology . The
degree of environmental protection is often characterized in terms of recur-
rence frequency and a concentration associated with some biological or health
effect. As an illustration, water quality standards are often written in
terms which state that the dissolved oxygen shall be above 5 mg/£ at flows
equal to or greater than that associated with the 7 day - 10 year low flow.
This low flow has a value which can be determined by analysis of the stream
flow records at a location. Even if records do not exist at a given location
there are techniques available to estimate this flow based on drainage area
and records for similar nearby streams. There is a need to define a compar-
able frequency related parameter for wet weather events .
One parameter which could be considered is related to the joint occur-
268
-------
ence of stream flow and rainfall. Historical records are available for
stream flow and for rainfall in most regions of the country. Methods for
extrapolation to specific sites are either available or could be developed.
As an illustration of the approach, daily stream flow records from Williams
Creek in Austin, Texas were paired with daily rainf.all data from this city
for the period October 1, 1976 to September 30, 1977. The number of occur -
ences of both rainfall and flow were determined and the results are presented
in Table 2. It may be seen from the table that most of the rainfall events
were associated with higher stream flows. However, a number of rainfalls
occurred during low flow periods. {One method of determining the relative
impact of runoff loads at various combinations of rainfall and stream flow is
to calculate the dilution available in the stream.} This can be done by
defining the runoff coefficient and the extent (size) of the urban area.
Figure 6 contains an idealized illustration of a site. The runoff volume is
defined by the product of the rainfall intensity, urban drainage area, and
runoff coefficient. This volume is divided by a storm duration to yield a
runoff flow. In the example, daily rainfall data was used and the duration
is assigned consistent with this time period. It should be noted that hourly
rainfall data is usually available at class A stations and could be used in
an analysis. Assuming that the runoff coefficient is 0.2 and that the rela-
tive drainage area of the stream was 100 times that of the urban area, the
calculated dilution factor in the stream is presented in Table 3 for urban
runoff from a separately sewered area. The dilution factor is defined by:
Dilution Factor =
Runoff Flow
Total Estimated Downstream Stream Flow
This factor may be multiplied by the concentration of a contaminant in the
runoff to yield an approximation pf the maximum instream concentration. The
total estimated downstream flow has been defined by the sum of the runoff
flow and the stream flow.
The event with the lowest dilution (i.e. highest dilution factor) can
be considered the most critical -for many parameters. Dissolved oxygen is
one parameter where this may not be true. The combinations of rainfall and
stream flow can be ranked 'based on the dilution factor. In the illustration,
the most critical condition occurs at a rainfall in the range of 0.1 to 0.2
inches which was associated with a stream flow of below 0.1 cfs. Under this
condition the maximum instream contaminant concentration would be approxi-
mately 83 percent of the concentration in the runoff. If desired, the rank-
ing can be weighted by the probability of occurrence related to the total
period analyzed (i.e. 365 days) or by the number of days of rain (in this
case, 88). These procedures could be employed to define frequency in terms
of sequential events over one or several seasons or years. An alternate
approach could be considered which uses an extreme value analysis technique
similar to that employed to define the 7 day - 10 year low flow used in
current stream standards. This procedure employs the most critical event for
each year of record and the frequency distribution defined by these events is
examined to determine the basis for environmental protection.
Table 4 has been developed considering the situation presented in Fig -
ure 6 and assuming that the intermittent discharge is from a combined sewer
269
-------
TABLE 2
NUMBER OF EVENTS OF JOINT STREAM FLOW AND RAINFALL
WILLIAMS CREEK, AUSTIN, TEXAS
October 1, 1976 to September 30, 1977
Daily Rainfall (ins)
0(1) .005 .05 .1 .2 .3 .4 .6
.01
.05
.1
.2
to
o
a •'
& 1.5
3
6
15
300
l.l 1.5
0
5
22
*
28
9
18
*5
84
32
10
2
2
3
3
14
6
3
1
1
4
1
f.
,
1
1
5
4
2
1
, 1
1
3
it
1
2
/
, 4
, 2
1
3
3
1
6
1
3
2
NOTE: l) included trace rainfalls and no rain
2) 88 Rain days
3) 365 days of record
270
-------
Stream gage
Rain gage
Proposed urban
area
Point of dilution calculation
Note: Stream flow translated on a drainage area basis.
Fig. 6. Site used in the example calculation.
271
-------
U]
-------
TABLE
CALCULATED INSTREAM DILUTION FOR CSO EVENTS
(D(2)(3)(5)
Daily Rainfall (ins)
0 .005 .05 .1 .2 .3 .4 .6 .8 1.1 1.5 5
n
.01
05
i
2
w
o
& .8
0)
-p
m 1.5
'•3
6
IS
300
-
_
-•
-
-
_
_
.
• _
-
_
_
- _
-
_
—
—
-••
-
_
_
-^
.52
.03
.14
.08
.03
.01
.01
.12
.01
.17
.01 '
-
.02
.05
NOTES: l) Stream Drainage "Area/Urban Area = 100:1
2) Runoff Vol/rain Vol = .h
3) Runoff Duration = 2k- Hrs.
4) No OTerflows
2
5) Interceptor Capacity = 6 cfs/mi
6) Number of overflows = 3^
273
-------
system with an excess interceptor capacity of 6 cfs/sq. mi., a runoff coef-
ficient of .4 and the ratio of stream to urban drainage area of 100:1. The
dilution factor in the stream tends to decrease due to the reduction in run-
off associated with the interceptor. In addition, the number of days in
which overflows occur are reduced by the interceptor from 88 to 34. The con-
centration of most constituents is substantially higher in combined overflows
as compared to urban runoff; therefore, the water quality impacts would be
larger.
In summary, one method for defining critical wet weather events has
been identified and can be considered for use in developing frequency infor-
mation which could provide a part of the basis for wet weather water quality
criteria. The approach illustrated is in the early stages of development and
employed a large number of simplifying assumptions. In its most generalized
form the approach could consider hourly rainfall and/or stream flow records,
variable runoff coefficients, and variable constituent concentrations as a
function of rainfall intensity or volume. An additional use which can be
made of the proposed approach is as a screening methodology for identifying
the relative size of river and urban areas in which water quality problems
could be anticipated. The critical relative size would vary regionally due
to climate and also be related to the intensity of urbanization. The approach
has not been tested and application should not be considered until adequate
development and site specific testing has been carried out.
Wet Weather Biological Impacts
As indicated previously the second part of wet weather criteria deals
with the biological or health effects of in -stream concentrations of contam-
inants. In the context of wet weather phenomena, it is necessary to consider
the time of exposure in addition to the concentration of the contaminant.
This is due to the transient nature of the intermittent loads from storm run-
off events. As an illustration of one possible approach to evaluation of
biological impacts the following example has been developed.
Figure 7 contains data on the percent survival of the snail (Hiys a
integra) as a function of time for several concentrations of cadmium. These
data are replotted on Figure 8 employing coordinates vhich can be employed
to develop equation (1) relating percent survival to exposure time and cad-
mium concentration.
waere:
K = 0.3 tc°'6
K = % kill
t = expos ure time (days)
c = cadmium concentration
(1)
Rewriting equation (1) in terns of population and differentiating with
respect to time yields :
dt
.3C-6 _
100 x o
(2)
274
-------
100
c
CD
£
Q.
75
ID 50
CO
25
0
Cadmium concentration (jug/1):
• <0.05 (control) n 27.5±2.7
o 3.0±0.3 A 85.5±7.6
• 8.3 ±0.3 A 238 ±16.4
4 7
14 21
TIME, days
From Spehar, R. L, Anderson, R. L, and Fiaudt, J. T.,
Environ. Pollut. 15, 1978, 195.
Fig. 7. Percent survival of Physa Integra (snail) after
28 days of exposure to cadmium
275
-------
OJ
=1
O
H 100
<
cc
h-
2
UJ
O
O
O
Q
<
O
10
AAvg. over concentration c
Kit = percent killed/time
0.1
= 0.3ft?0-6, where:
K= percent snails killed
t = exposure time, days
c = cadmium concen-
tration, g/l
I I I I I I III I I I I I MM
i i 11 ii
1.0 10
Kit
Fig. 8. Cadmium concentration vs. average snail mortality/time.
276
-------
where:
P = initial snail population
P = snail population at time "t"
The data employed was from a series of mortality experiments conducted
at different concentrations of cadmium. The cadmium concentrations were
maintained at a constant level in each individual mortality experiment. The
expression in equation (2) defines the rate of change of the snail population
as a function of the time invariant cadmium concentration.
A s torm water discharge into a completely mixed lake can be considered.
In order to minimize the mathematical complexities of this illustration zero
ourflow from the lake was assumed. The time history of cadmium concentration
in the lake is defined by_ equation (3) assuming cadmium is removed b/ sorption
and settling at a rate "k"^) .
\here:
C = Co e' (3)
C = cadmium concentration at time "t" (yg/&)
C = initial cadmium concentration (yg/Jl)
k = equivalent rate of removal (I/Day)
t = time (Days)
Assuming that a measure of the irate of change of the population in
experiments .run at constant but different concentrations is equal to the rate
of change of the population wien the concentration is varying with time ;: v?
allows substitution of equation (3) into equation (2) and integration yields :
for
t
Po @ t = o and C = Co @ t = o
,,0.6
= Po(l
kx200
,-k x .6 x t - N..
-(e - 1))
(4)
Equation (4) relates the population of snails to the cadmium concentra-
tion and time after a storm event in a simplified lake system. Calculations
are presented in Tables 5 and 6 using equation (4). In Table 5 it is assumed
that a small lake receives a large urban load generated by three different
rainfall events . The calculations in Table 6 are for the same rainfall
events . cons idering a lake xhich is five times as large. Information on popu-
lation reductions can be combined with estimates of' the frequency of rainfall
occurence to provide a basis for criteria to be employed in evaluation of wet
weather standards for different types and s izes of water bodies .
In any generalization of the above illustration the initial cadmium
concentration "Co" would reflect the size of the urban drainage area, land
use, size of the rainfall, antecedent conditions and other features of the
sys tern. In addition, the value of initial concentration "Co" would reflect
the initial dilution for various sizes and types of water bodies and could
277
-------
TABLE 5
INIT. POPULflTION= 500 NUMBERS
INIT. CONC. = 28 PPB
CHEM. RERCT10N RflTE= . 1 I/DRY
TIME/DflYS
8
5
10
15
20
25
38
35
48
45
58
POPULRTION
508
468. 901
431. 936
410. 478
394. 582
382. 806
374. 082
367." 619
362.831
359. 284
356.656
JiKILLED
0
7. 81975
13. 6128
17. 9043
21. 0836
23. 4389
25. 1837
26. 4763
27. 4333
28. 1432
28. 6688
INIT. f-OPULRTION= 500 NUMBERS
INIT. CONC. = 10 PPB
CHEM. REfiCTION
TIME/DflYS
8
5
10
15 :
28 ;:
25
38
35
48
45
58
RflTE= . 1 1/DflV
POPULflTION
500
474. 205
455. 095
440. 938
430. 45
422. 681
416. 925
412. 661
409. 502
407. 162
485. 428
SKILLED
0
5. 15911
8. 98107
11. 8124
13. 91
15. 4639
16. 615
17. 4678
18. 0996
18. 5676
18. 9143
INIT. POPULflTIOH= 500 NUMBERS
INIT. CONC. = 3.
CHEM. REflCTION
TIME/DflYS
8
5
18
15
20
25
38
35
48
45
58
33333 PPB
RflTE= . 1 1/DflY
POPULflTION
500
486. 656
476. 771
469. 448
464. 023
460. 004
457. 027
454. 821
453. 187
451. 977
451. 08
SKILLED
0
2. 66871
4. 64574
6. 11836'
7. 19539
7. 99918
8. 59465
9. 03579
9. 36259
9. 60469
9. 78405
CHEM. CONC.
20
12. 1306
7. 25759
4. 4626
2. 70671
1. 6417
. 995742
. 603943
. 366313
. 22218
. 134759
CHEM. CONC.
10
6. 06531 .
3. 67879
2. 2313
1. 35335
. 82085
. 497871
. 301974
. 183156
. 11109
. 0673795
CHEM. CONC.
3. 33333
2. 02177
1. 22626
. 743767
. 451118
. 273617
. 165957
.100658
. 0618521
. 03703
. 0224598
278
-------
TABLE 6
INIT. POPULflTION= 2560 NUMBERS
I NIT. CONC. = 4 PPB
CHEM. REACTION RfiTE= . 1 1/DflV
TIME/DflVS POPULflTION SKILLED
0 2500 8
5 2425. 57 2. 97722
10 2370. 43 5. 18279
15 2329. 58 6. 81673
20 2299. 32 8. 82717
25 2276. 9 8. 9239
30 2260. 3 9. 5882
35 2247. 99 10. 0803
40 2238. 88 , 10. 4449
45 2232. 13 10. 715
50 2227. 12. 10. 9151
CHEM. CONC.
4
2. 42612
1. 47152
. 892521
. 541341
. 32834
. J.99148
. 12079 :
. 0732626
. 044436
. 0269518
INIT. POPULflTION= 2598 NUMBERS
INIT. CONC. = 2
CHEM. REfiCTION
TIME/DflVS
0
5
10
15
20
25
30
35 ' . -
40
45
50
PPB
RflTE= . 1 1/DflV
POPULflTION
2500
2450. 89
2414. 52
2387. 57
2367. 6
2352. 81
2341. 85
2333. 74
232772
2323, 27
2319. 97
:
/•KILLED
0
1. 96423
3. 41937 ^
4. 49736
5. 29596
5. 88758
6. 32585
6. 65054
6. 89107
7. 06926
'7. 20127
CHEM. CONC.
2
1. 21306
. 735759
. 44626
. 270671 1
.16417 '-'-
. 0995742
. 0603948
. 0366313
.022218
.0134759
INIT. POPULflT I ON= 2500 NUMBERS. , .>.'•
INIT. CONC. = . 666667 PPB
CHEM. REflCTION RfiTE= . 1 1/DflV
TIME/DflVS POPULflTION SKILLED
0 2500 0
5 2474. 6 1. 01606
10 2455. 78 1. 76878
15 2441. 84 2. 32641
20 2431. 51 2. 7395
25 2423. 86 3. 04554
30 2418. 19 3. 27225
35 2414 3. 44021
40 2410. 88 3. 56463
45 2408. 58 3. 6568
50 2406. 37 3. 72509
CHEM. CONC.
. 666667
. 404354
.245253
. 148753
. 0902236
. 8547233
. 0331914
. 0201316
. 0122104
7. 406E-03
4. 49197E-03
279
-------
include specific considerations such as river flow and dispersive mixing as
appropriate. The type and size of the water body would also be included in
the analysis and influence the relationship between the time "t" and the spa-
tial location of calculated population reductions. In the current example
snails were the organism examined. It was assumed that the snails were sta-
tionary in the lake. Organisms with different characteristics relative to
the water column can be considered. As an example, plankton, which are trans-
ported with the storm pulse, could be analyzed as could fish which were con-
fined to regions of the receiving water as the storm pulse passed. In gen-
eral, equations (1) and (2) of the illustration would vary for different
organisms and contaminants and could extend from classical water quality
variables such as dissolved oxygen to toxics like cadmium and consider fish,
plankton or bottom organisms.
The removal rate "k" in the cadmium example would vary depending on the
contaminant being studied and the type of water and/or water body. For clas-
sical water quality variables such as dissolved oxygen, suspended solids and
coliform; conventional first order kinetics appear appropriate. For toxics
it is possible to evaluate an equivalent first order removal rate (6).
Once again the approach illustrated is in the early stages of development and
significant additional effort and testing is required before application to
site specific situations.
Conclus ions
The impacts of intermittent discharges on receiving waters may be exam-
ined using an array of methodologies ranging from reasonably well tested
approaches to those that are in the early stages of investigation. The
latter approaches require further development and testing before they can
be considered as practical tools. They do, however, hold some promise for
providing a portion of the basis required to develop wet weather criteria
which can be employed in the standards setting process.
References
1. Executive Summary -New York City 208, Hazen and Sawyer, 1978.
2. N.Y.C. 208 Task Report (PGP Task 314) "Seasonal Steady State
Modeling", Hydroscience, Inc., 1978.
3. N.Y.C. 208 Task Report (PCP Task 334) "Time Variable Modeling",
Hydroscience, Inc., 1978.
4. Spehar, R.L., Anderson, R.L. and Fiandt, J.T. Environ, Poll. 15,
1978, 195.
5. Manhattan College Summer Institute Notes - "Analysis of Natural
Systems" 1979.
6. "Environmental Report Tenn. Coloney Res.", U.S. Army Corps of
Engineers, Fort Worth Dist. Office, 1979.
280
-------
THE RESPONSE OF GREAT LAKES ESTUARIES TO STORMWATER RUNOFF
Authors: John R. Adams
Stephen M. Yaksich, Ph.D
Army Corps of Engineers, Buffalo District
1776 Niagara Street
Buffalo, NY 14207
ABSTRACT
Glacial rebound of the northern shore of Lake Erie has resulted in
drowned rivermouths of southern shore tributaries. Transport of material
through these estuaries depends not only on river stage but also lake stage.
This paper describes studies carried out in two Lake Erie triTutaries during
high and low flow conditions. It also examines the effect of localized urttan
runoff on one of the estuaries.
Three separate studies of water chemistry in the lo \er Maumee River,
an estuarine river section., were conducted by the Toledo Metropolitan Council
of Governments, the Buffalo District Corps of Engineers, and others, during
1974 and 1975.. Base flow river conditions and the resultant estuary chemical
variations \ere measured during the summer of 1974. Winter storm
runoff effects \ere measured during January and February, 1975. Additional
sampling and analysis was conducted during the summer of 1975, giving a
fairly complete record of water chemistry variability.
Winter storm runoff \as also measured in the estuary of the Cuyahoga
River, with simultaneous river and estuary sampling over the hydrograph for
selected pollutants.
This paper outlines the problems and complexities of chemical measure -
ments of rater quality in estuarine systems of the Great Lakes during varying
hydrolbgic conditions. The studies in the Maumee River point out the impor-
tance of sampling program design, and the difficulty in evaluating the trans-
port of pollutants through such estuarine systems. The summer studies were
conducted by TMACOG as part of an evaluation of in situ \ater quality, as
effected by a num ter of waste discharges from the Toledo area. The winter
storm runoff measurements, made simultaneously at both an upstream riverine
station and a do Tnstream estuary station xere designed to evaluate whether
the mass transport teing carried out into the basin was passing through the
281
-------
estuary and out into the lake. The question of an increase or decrease in
mass transport was also considered, to evaluate scouring and depositional
mechanisms in the estuary.
The results of the Maumee estuary measurements were compared vith the
Cuyahoga estuary in Cleveland, where similar sampling of river and estuary
transport ras carried out for the same storm periods.
Introduction
Many tributaries to the Great Lakes have river mouths which can scien-
tifically and legally be considered lacustrine estuaries. The fate of pollu-
tants in and through these estuaries presents a problem for water resource
planners. Transport of material through these estuaries depends not only on
river stage but also lake stage. This paper describes studies carried out in
two Lake Erie tributaries, the Maumee and Cuyahoga Rivers.
The studies considered include: the effects of water quality on the
estuaries and the effects of the estuaries on water quality. That is, how
the quality of runoff water from the agricultural watershed above the river
mouths and the local urbanized areas effect the estuaries, and how the pres-
ence of an estuarine reach effects water quality in the estuary and the
transport of water pollutants or nutrients to the water body-beyond, in this
case Lake Erie. . . : '
The effect of lake levels on the flow of river water through the estuary
has a dramatic effect on the quality of water in the "river." It has pre-
viously been the habit of waste load allocation planners to treat these
"river" segments in the same manner as any other free flowing segment. In
actuality, the behavior of the estuaries is much more lake like.
The present studies have spent a large effort on refining estimates of
loadings of pollutants and nutrients to Lake Erie. Part of this effort
included streamflow and water quality parameter measurements in the estuaries
and at upstream gaging stations during storm events. The purpose of these
measurements was to provide estimates of the transport of storm runoff
related pollution through the estuaries to Lake Erie.
Description of Sampling Programs
Three separate studies of water chemistry in the lower Maumee River were
conducted by the Toledo Metropolitan Area Council of Governments (TMACOG),
the Buffalo District Corps of Engineers, Lake Erie Wastewater Management
Study (LEWMS), and USEPA, during 1974 and 1975. Base flow river conditions
and the resultant estuary chemical variations were measured during the summer
of 1974. Winter storm runoff effects were measured during January and
February 1975. Additional sampling and analysis was conducted during the
summer of 1975, giving a fairly complete record of water chemistry varia-
bility.
282
-------
Figure 1 is a schematic diagram of the sampling locations utilizd in the sum-
mer water quality survey. Stations with numeric identification are TMACOG
stations while those with alphabetic identification were utilized by Jones &
Henry Engineers, Ltd. Samples collected at Stations 2, 5, 8, 11, 14, and 17
were analyzed for all the parameters listed in Table 1. The remainder of the
numeric station samples received laboratory analysis only for fecal coliform
organisms and chloride. All samples for laboratory analysis were taken at
mid-depth with a Kemmerrer or Van Doren Sampler, and were preserved imme-
diately after collection by placing on ice. Laboratory analysis for noncon-
servative parameters was begun within 6 to 8 hours of the collection of the
first sample. On a few occasions when immediate analysis was not possible,
chemical preservation plus refrigeration was employed.
Table 1 - Parameter List:
Maumee River Water Quality and
Mass Transport Studies
Station Location
Date
Time • r
Conductivity JL/
Total Phosphorus JL/
Ortho Phosphorus JL/
Nitrate & Nitrite JL/
Ammonia JL/
Total Kjeldahl Nitrogen
Total Solids I/
Suspended Solids —' ' ;
Total Organic Carbon 2J
Si09 JL/ - " '
Chloride'!/
PHl/
Extended Bod A/
Fecal Colif orms .!/
Calcium -2/
Magnesium -27
Potassium -=!/
Sodium -I/
• Iron 3/ ...
Copper .£/
Manganese .£/
Disolved Oxygen -27
L' Analyzed in all studies
2J Analyzed in Maumee Water Quality and Cuyahoga Mass Transport
.3/ Maumee Water Quality only
At each of the alphabetic stations, field instruments were utilized to
determine depth profiles of dissolved oxygen, temperature, and conductivity.
The instantaneous depth of the water at the sampling point was also recorded
to the nearest foot. Samples were collected once weekly at all stations
throughout the sampling period.
Sampling stations were chosen on a grid pattern beginning at river mile
(RM) 8.5, and extending to the mouth of the river at RM 0.0. The upstream
station was chosen as being the probable furthest extent -of the upstream
influence of pollution contributed to the river in the heavily urbanized sec-
tion of the river beginning around RM 6.5 and above all major point sources
in the estuarine' section of the river.
Independently from the water quality surveys conducted in the lower river
during the summer of 1975, both the U. S. Army Corps of Engineers and TMACOG
have engaged the Heidelberg College River Studies Laboratory in a study of
283.
-------
9.0
1 2
a
4
f
7
k
10
P
13
u
16
z
b
d
e
5
g
1
j
8
3
c
8.0
6 7.0
h
6.0
9
1 m
5.0
n
0
II
q
s
t
14
V
X
y
17
aa
cc
12 4.0
r
3.0
15
W
2.0
18 1.0
bb
0.0
en
Ul
_j
2
a:
UJ
>
a:
dd
MAUMEE RIVER ESTUARY SAMPLING POINTS
FIGURE I
284
-------
mass loading of pollution in the Maumee River Basin. In this study, a sample
of the Maumee River has been taken every day between December 1974 and July
1977. Samples were collected by an automatic sampler which was located
inside the pump house of the Bowling Green, Ohio water supply intake. The
sampler collects one 500 ml sample every 6 hours. During periods of stable
hydrograph, only one of the four daily samples was analyzed. Whenever a
storm event occurred and the hydrograph rose significantly above base flow,
all samples were analyzed to provide greater definition of the chemograph
over the storm. Flows for this station are available from the USGS gaging
station at Waterville.
Additional sampling of parameters not measured by LEWMS and TMACOG.
studies, namely biological oxygen demand and fecal coliform density, has been
done by chemists at the Lucas County Maumee River Waste Treatment Plant.
Samples which they took in the Maumee River at Waterville have been utilized
in this report.
Mass transport studies were conducted in both the Maumee and Cuyahoga
estuaries. The Maumee River was sampled at the Cherry Street Bridge in the
estuary (river mile 4.8) and above the estuary at the Waterville water treat-
ment plant, 2.0 miles upstream of the USGS water level gage (river mile
21.2). Flow measurements were made at Cherry Street by the U.S. Geological
Survey. Samples were collected at three locations on the cross section and
two depths. Parameters analyzed are shown in Table 1. The Cuyahoga River
was sampled in the estuary at Third Street Bridge (river mile 3.0) and above
the estuary at the USGS water level station at Independence (river mile
13.7). Samples were collected on the rising and falling state of the
hydrograph by the city of Cleveland Water Quality Laboratory. Estuary flow
measurements were made by the U. S. Geological Survey. Methods used for
chemical analysis for both studies are reported elsewhere (1).
Figure 2 is a map of the Maumee River study area. The locations of all
known wastewater discharges, municipal and industrial permitted and unper-
mitted, have been marked on the map. Also shown on the map are the locations
of 24 of the city of Toledo's combined sewer regulators which discharge to
the Maumee River and Swan Creek .within the study reach, and numerous
"unpermitted" dischargers, which are generally small package wastewater
treatment plants discharging less than 50,000 gallons per day of treated
sanitary sewage.
Analysis of Estuary Sampling Program
A series of samples were taken on the Maumee River, at Cherry Street
Bridge during the storm runoff event of 10 January 1,975.
285
-------
UAUMEE BAY
CHERRY STREET
SAMPLING STATION
e
LEGEND
••COMBINED. AND SANITARY
SEWER OVERFLOW
REGULATORS
» MAJOR PUBLIC AMD INDUSTRIAL
POINT SOURCE DISCHARGERS
&SMALL UMPERMITTED
DISCHARCERS
UPSTREAM EXTENT
MAUMEE RIVER ESTUARY WATER QUALITY SURVEY REACH
FIGURE 2
286
-------
Two samples were taken at each of three locations on the cross section
because of the width of the section and the location of bridge piers, as well
as the proximity of Swan Creek. After an analysis of velocity profiles indi-
cated that velocities were similar at each location, the samples were com-
posited. The results were three values of flux rate over the storm runoff.
These values were then compared with the corresponding hydrograph and mass
transport estimate for the Waterville station, about 19 miles up river
(Table 2). Interestingly, the mass flux rate is slightly greater in the
estuary for corresponding flows indicating no deposition is occurring. The
storage retention in the estuary between the two stations shows first a time
lag, then a slower recession rate, over the hydrograph (Figure 3).
Table 2 - Storm Transport Comparison - Maumee River
(9 thru 14 January 1975)
(D.A
Date
9 Jan 75
11 Jan 75
15 Jan 75
River Station
Waterville Gage
= 6,329 square miles)
: : Total .
: Flow :cfs/ ; Phosphorus
Time: (cfs) :sq.mi: kg/day
* * *
1000:17,600: 2.78:
* • •
1700:32,700: 5.16: 103.16 •
• • •
0900:13,000: 2.05: 19.08
! • • •
Estuary Station
Cherry Street Bridge
(D.A. = 6,500 square miles)
: : : : Total
: : Flow :cfs/ : Phosphorus
Date :Time: (cfs):sq.mi: mt/day
• » • •
• * • »
10 : Jan 75:1650:17,800: 2.74: 28.13
• • * •
• * • •
11 Jan 75:1430:33,400: 5.14: 105.97
• • • •
• * • •
14 Jan 75:1250:13,000: 2.00: 33,44
• • • •
On the Cuyahoga, the estuary was much easier to measure, with the Third
Street Bridge station providing a uniform, rectangular channel. The chemical
transport however, was far more complicated. During the storm of 21 May
through 28 May 1975, simultaneous chemical samples were taken and discharge
measurements were made at this station and at Independence, 10.7 miles
upriver. Both stations failed to demonstrate a distinct increase in chemical
concentration over the hydrograph, with the chemical variations large and
erratic for phosphorus as well as for other parameters. Eight discharge
measurements and 19 chemical samples were taken at Third Street over a 164
hour period. From these data, eight instantaneous flux estimates were made
for total phosphorus. These data were then compared with Independence.
While the concentrations of total phosphorus- varied considerably at both
stations, the hydrographs of the two stations were nearly simultaneous
(Figure 4) and the mass transport past the two stations was of about the same
magnitude (Figure 5). Both stations had a maximum discharge of about 12,000
cfs which occurred at 0900 on 22 May. Furthermore, the plots of instan-
taneous flux versus elapsed time showed both stations carrying a peak flux
rate of nearly 12 metric tons per day with approximately equal total mass
transport over the runoff period.
While neither of these analyses can be considered totally conclusive,
they do suggest strongly that the measurement of mass transport at riverine
287
-------
•-FLOW AT CHEMV STREET
•H
B-
3^-
5'
DATE
COMPARISON OF FLOW AT WATERVILLE WITH FLOW *T CHERRY STREET
s ••
1
K
I
ELAPSE TIME (HOUKS)
HYOHOOIMPHS OF CUYAHpa* RIVER AT INDEPENDENCE AND THUD
STREET BRIOOE . CLEVELAND; MAY gilt THROUOH MAY 28th
FIGURE 4
ELAPSE TIME (HOURSt
INSTANTANEOUS FLUX RATE OF TOTAL PHOSPHORUS
CUYAHOG* RIVER-MAY 2l«t THROUOH MAY 2Bth
FIOURE 5
288
-------
stations above the estuary sections will give representative estimates of
mass flux from the basin during high flow events. The localized effect of
urban runoff may be evident much sooner than any rise in stage in the
estuary, and so future station sampling should be initiated at the start of
the rainfall event. Also, it would appear that no significant storage or
discharge of accumulated pollutants occur in the estuary above either sta-
tion. Since, dredging is required in both channels on a large-scale basis,
this would appear to be an incorrect conclusion; however, it is likely that
the load measured in transport by this sampling is of a different character
than that material which fills the channel on the bottom of each river.
Summer Water Quality Surveys
Summer surveys in 1975 were designed to study the water quality of the
estuary in response to Its complex hydraulics and the inputs of pollution
from upstream and local runoff. There were several survey days when stagna-
tion was the apparent cause of extreme water quality degredation in the
estuary. 31 July 1975 was the worst, and is here discussed in detail.
Beginning on 19 and 20 July, a heavy rainfall was experienced in the
Maumee Basin and locally in Toledo. On the morning of 20 July, 0.97 inches
of rain were recorded between 1100 and 1200 at the Toledo Express Airport. A
total rainfall of 1.29 inches was recorded between 1000 on the 19th and 1500
on the 20th. This heavy local rainfall flushed a considerable amount of com-
bined sewage and stormwater into the river. At the same time the Maumee
River at Waterville was rising from 1,000 cfs on the 18th to 6,600 cfs on the
20fh and- bringing with it an estimated 324,000 pounds of BOD(5). This fairly
large summer storm did not produce enough water to exchange the entire
estuary volume and did not significantly alter water velocities in the
estuary segment. Velocities here were controlled by the rising and falling
of Lake Erie. During this time the Toledo STP continued to discharge at
least its long-term average load of BOD(5) of approximately 28,200 pounds per
day.
From 20 through 25 July, the Lake Erie stage gage at the mouth of the
river was at best jittery, rising and falling only a few tenths of a foot in
short period oscillations. Beginning 25 July, wind setups on the lake were
somewhat larger causing a 2.0-foot rise on the evening of 25 July.
Immediately the lake level fell and soon began the very regular oscillations
which characterize the Lake Erie seiche. This seiche was not sustained, but
damped until the day of sampling when the stage was virtually flat. By this
time the discharge of the Maumee River at Waterville had fallen to 500 cfs,
which was nearly as low as it would fall all summer.
The combination of factors during the previous 10 days was adding up to a
very bad time for the river. It had received a heavy load of oxygen
289
-------
demanding substances from both the local urban area and the rest of the Maume
Basin from storm runoff. River discharge was now low and the lake was
quiescent. It was very hot and river temperaures were as warm as they would
be all summer.
Figure 6 depicts dissolved oxygen, temperature, and conductivity data
gathered on 31 July 1976. All parameters measured showed marked stratifica-
tion. Dissolved oxygen is stratified by 6.2 mg/1, 7.2 mg/1 at the surface,
1.0 mg/1 near the bottom. There was no discernable sag point. Minimum DO
values were observed at RM 2.5 and 8.5. Temperature was vertically stra-
tified by 2 to 3°C at most points in the river.
Conductivity was higher throughout the river than on any other sampling
day of the summer, and was highest near the mouth of the river above the
Toledo STP outfalls. Maximum conductivity readings, 630 umhos/cm, are
apparent at RM 1.0 on the west bank of the river, just upstream of the Toledo
STP outfalls, and dropping off in all directions from that point. At RM 0.0,
conductivity is reduced to 560 umhos/cm as boundary waters of the Maumee
River and Bay mix. At RM 8.5, conductivity has fallen to 470 umhos/cm as
pure river water enters the survey reach. Figure 11 shows that conductivity
of the Maumee at Waterville had been as low as 380 umhos/cm on 18-20 July.
It is apparent that the high conductivity effluents of the Toledo STP and
secondarily the Gulf Oil Refinery are influencing the dissolved solids con-
tent of the entire estuarine volume below RM 8.5.
Such severe stratification occurred only on sampling dates that were pre-
ceded by a quiescent lake. Stratification, to some degree, was almost always
apparent, but all other sampling dates had been preceeded by at least the
regular mixing action of the seiche.
The Upstream Heritage and its Influence on Estuarine Water Quality
This section discusses water quality measurements upstream at the
Waterville gage and then relates it to water quality measured in the estuary.
This relationship varies at different times of the year, with different con-
comitant hydrologic regimes, and in different hydraulic segments. The flows
considered are high flows associated with storm events, intermediate flows
and low flows associated with the dry months of the year (but not necessarily
drought flows).
The bulk of the pollutants coming out of the Maumee Basin is transported
during a few large storm events which take place primarily during the winter
and early spring of the year. Approximately 80 percent the total annual
transport of sediment and phosphorus take place during 20 percent of the
time. The bulk of nutrients and oxygen demanding substances are also
transported at this time. During these high flow winter storms, water tem-
perature is low and reaeration is high so the pollutants do not have a strong
impact on the traditional parameters of water quality, but they do have their
impacts at other times and in distant locations. As shown previously
material is transported through the estuary during high flow events.
However, during runoff events associated with lower flows some of the
nutrient and oxygen demanding sediment settle as the storm waters of the
290
-------
_l
1 'S.
0
S
o
£ °
OT
UJ
UJ
DC
0
UJ
O
2
o
>*
en
o
i
z
o
a:
o
i
DISSO
HJAJ
9.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
29
28
27
26
25
24
600
500
400
LVE
* •• " f)|SSOI VFD OKYfiFN
x x * ,
X ' ' y .
X
°
e . o '
o a n
• o ,
a :
. . , . ' ' , • . ' .'\
11 II 1 1 1
0 , 1.0 2.5 4.0 5.5 7.0 8.5
.' -" "t . • • "
* TEMPERATURE
n n -V " -
0 ° : - D X V '
a ° n 5
u o
.•'. . '., .-.. O ;'..•'•
u n
' . . - - - ,, t
i - ,f • ~ ' " -
\ I I I I I 1
0 1.0 2.5 4.0 5.5 7.0 8.5
D n . . •
a ° x g, 5 CONDUCTIVITY
LEGEND
ox
TC x TOP (20%
DEPTH)
o « ° MID DEPTH
0 n BOTtOM
(80% DEPTH)
II ' ' ' | '
0 1.0 2.5 4.0 5.5 7.0 8.5
RIVER MILES
D OXYGEN, TEMPERATURE 8 CONDUCTIVITY FOR MAUMEE
RIVER SURVEY REACH 31 JULY 1975
FIGURE 6
291
-------
Mauraee flow into the estuary where the river deepens and widens and veloci-
ties decrease. Although this is a small amount of the total sediment
transported by the Maumee, it does affect water quality. As the water warms
later in the year, these sediments will demend oxygen and cause water quality
degradation.
There are other direct impacts. During storm events, the Maumee may
carry 2,000 mg/1 or more of suspended sediments. These sediments must be
removed for the water supplies of Defiance, Napolean, Bowling Green, and
Waterville. Also, nitrate exceeded the USEPA recommended maximum con-
centration of 10.0 mg/1 NOg-N on 14 days during calendar year 1975.
While the loadings are primarily of importance to the long-term quality
of the Great Lakes system, concentrations and their instantaneous time of
occurrence are most significant for water quality in the riverine and
estuarine segments. As discussed in th£ section on estuarine stagnation,
lake level variations and river discharge velocities are important variables
to instantaneous water quality.
Figure 7 is a plot of stream flow of the Maumee River at Waterville and
rainfall at the Toledo Express Airport near Toledo. The dates when sampling
was done in the estuarine reach are given by the dots along the date scale.
During the summer months, flows are reduced and so are pollutant con-
centrations. The river could always be considered enriched or eutrophic with
respect to the concentrations of nutrients for which standards have been set
for Lake Erie. Figures 8 thru 12 plot the concentrations of five of the
parameters which were measured daily (except for some downtime) in the Maumee
River at Waterville during the survey period of 1975. Nitrite-nitrate nitro-
gen (Figure 8) shows more of a seasonal dependence than a storm event depend-
ence. The abrupt increases correlate well with storm events, but the
greatest variance is with a longer time extending through the summer. It is
most interesting to note that large increases in nitrate concentration do not
take place during the storms which took place during September. Ground cover
is most complete at this time and most available soil nitrogen is probably
incorporated into growing plant material.
The suspended sediment plot shows a similar pattern, but with stronger
flow dependence. The sediment graph parallels the hydrograph very closely,
but a seasonal variation is apparent. The late summer storm hydrographs do
not produce sediment graphs of the same magnitude as the early summer storms.
The difference is small, but obvious. Again, ground cover is more complete.
It would appear that although ground cover does reduce sediment yield, the
protection is not complete.
The first of the high velocity late summer storms, 23 August, is smaller
than several of the later events, but produces the greatest peak suspended
sediment concentration. This concentration reflects the washout of sediments
from short-term storage in the river bottom during the summer. The
6 September event produces only a small increase over the elevated background
concentration and is perhaps a better indication of sediment runoff which can
292
-------
-iiaooo
UJ
10 20 I 10
1975
STREAMFLOW. MAUMEE RIVER AT WATERVILLE AND TOTAL
.DAILY PRECIPITATION. TOLEDO EXPRESS AIRPORT
FIGURE 7
9.0
6.0
^5.0
2,0
LJ
O
3.0
O
O 2.0
10
9 1.0
JUNE [JULy
|AUGUST|
I SEPT.
21
I
20
I
10
20
10
20
NITRATE+ NITRITE-N , MAUMEE RIVER AT WATERVILLE
FIGURE 8
293
-------
2500
10
20
SUSPENDED SOLIDS CONCENTRATION. MAUMEE RIVER AT
WATERVILLE
FIGURE 9
o
5
.SEPT.
21
10
20
10
20
I
10
20
TOTAL PHOSPHORUS CONCENTRATION. MAUMEE RIVER AT
WATERVILLE
FIGURE 10
294
-------
13
tr
o
co o
o 3;
a. o
Q o
UJ o
o
co
co
0.8
0.6
0.4
0.2
JUNE ,JULY |; .
|SEfT. ,
21 I II 21 I 10 20 I 10 20
DISSOLVED PHOSPHORUS CONCENTRATION . MAUME E
RIVER AT WATERVILLE
FIGURE II
600
500
400
300
200
^ 100
o
OS/CM
o
cr
o
r
'
JUNE JULY
,AUGUSJT
5EPT.
21
10
20
10
20
1
10
20
CONDUCTIVITY. MAUMEE RIVER AT WATERVILLE
FIGURE 12
295
-------
be expected from a late summer.storm. It is also, perhaps, a better indica-
tion of the effectiveness of ground cover protection in late summer.
The plot of total phosphorus demonstrates the same flow dependence as
suspended sediments and indicates that total phosphorus is strongly asso-
ciated with land wash. Dissolved phosphorus, on the other hand, does not
increase markedly with flow during summer storm events and in many instances
shows a slight dilutional tendency. Concentrations never decrease suf-
ficiently to decrease loads. Rather, the loads of dissolved phosphorus
always do increase with flow. This may be associated with land wash (both
urban and agricultural) and with combined and storm sewer overflows from
cities.
Dissolved phosphorus is the form of phosphorus which is most readily
available as a plant nutrient. Over the entire summer period, dissolved
phosphorus was present in the Maumee River at Waterville at a concentration
of 0.136 + 0.009 mg/1 (mean and standard error for 105 samples). This con-
centration indicates that there was almost always plenty of phosphorus in the
river for plant growth. The major source of this important plant nutrient
during the summer months was in the continuous input associated with munici-
pal sewage treatment plants and equilibrium dissolution from bottom sedi-
ments.
Conductivity, as shown in Figure 12, varied inversely with flow and did
not exhibit a seasonal dependence.
Figure 13 is a load graph of flux versus flow for BOD(5) (20°C, unseeded,
dark) in the Maumee River at Waterville over the period from September 1974
through April 1976. The data used for this graph was gathered as part of an
NPDES self monitoring program by staff of the Lucas County/Maumee River
Wastewater Management Plant. The samples are taken once weekly at Waterville
very near the location used for sampling in this study. The shape of the
curve indicates a very strong relationship between flow and the load of
BOD(5) being carried in the river. When flow goes up, the BOD(5) load
increases. When water runs off the land it carries oxygen demanding substan-
ces with it.
At mean flow, the Maumee carries from 25,000 to 200,000 Ibs/day of BOD(5)
compared to the long-term average BOD(5) loading of the Toledo STP of 28,700
Ibs/day. At higher flows above 10,000 cfs, the data does not include any
load measurements below 100,000 Ibs/day. Considering that the population of
the remainder of the entire Maume River Basin above Waterville is little more
than the population of the Toledo urban area, that it is scattered over 6,330
square miles and that BOD(5) is a nonconservative substance which decays with
time and distance in the river, it seems unlikely that the major source of
this pollutant is people. Agricultural land wash must be considered to be a
significant source.
Figure 14 shows the relationship of water quality in the Maumee River at
Waterville to the quality of the waters of the lacustrine estuary of the
Maumee at Toledo. The analysis was performed with water quality at
Waterville as the independent variable, and water quality in the estuary as
296
-------
o
o
,1 2 3456 7691 2 3456
9
6
E
Ixl043
2
J .
9
7
6
4
1
9
3
2
1
*
2
*
3
4
•
6
•
78
..
•
.
a
. •
••
,
9
•
•
-.
.
*
2
'
*
3
—
•
4
*
J
5
789
•
..
6
*
7
85
2 345 67891
• .
*
*
9
3
'
"•
2
'
3
•
4
5
6
78
9
BOD (5) LOAD (LBS./DAY)
LOAD GRAPH OF BOD (5) vs. FLOW,MAUMEE
RIVER AT WATERVILLE
FIGURE 13
297
-------
o
a
UJ
_J
O
o:
UJ
DC
UJ
UJ
V)
** 8.0
_i
2 6.0
•o 4.0
o
* 2.0
w
a
o
CO
iu CD
UJ
0.
CO
=>
co
a.
CO
a.
o
to
a:
o
o
200
150
100
50
0.2
0.15
O.I
0.05
80
60
40
20
(A)
r2 = .95
2.0 4.0 6.O 8.0
(C)
r2=.03
50 100 200
(E)
r2=.33
0.5 0.10 0.15 0.2
(G)
r2=.35
0.5
z 0.4
CO
< 0.3
0-2
O.I
a- 0.4
CO
•* 0.3
_i
o 0.2
s
a. O.I
I- ,
20
-J 15
—.
CD
s 10
CM
0 5
en
o
11
O o
5
500
250
(B)
2=.03
O.I 0.2 0.3 0.4 0.5
O.I 0.2 0.3 0.4 0.5
(F)
r2=.24
20 40 60 80
250
500
RELATIONSHIP BETWEEN THE WATER QUALITY OF THE MAUMEE
RIVER AT WATERVILLE (ABSCISSA) AND THE MAUMEE ESTUARY AT
TOLEDO (ORDINATE) JULY TO SEPTEMBER 1977;
(A) NITRI TE - NITRATE NITROGEN (B)AMMONIA NITROGEN
(C)SUSPENDED SEDIMENT (D)TOTAL PHOSPHORUS
(E)SOLUBLE ORTHO PHOSPHORUS (F) DISSOLVED SILICA
(G) CHLORIDE (H) CONDUCTIVI TY
ABSCISSA AND ORDINATE ARE THE SAME SCALE.
FIGURE 14
298
-------
the dependent variable. The concentrations representing water quality in the
estuary are an average of measurements made in the estuary for one day. Six
stations with samples taken at the middle of the river at mid-depth were
used. The stations were averaged to integrate all sources of input to the
estuary and the effect's of time of passage through the estuary. , The con-
centrations representing water quality at Waterville are the mean of each
parameter for 10 days prior to the day of sampling in the estuary.
The 10 day mean was used to allow for time of travel and mixing between
Waterville and Toledo. This travel time is not an unrealistically long time
due to the relatively low flow volumes and the large estuarine volume.
From Figure 14a it can be seen that nitrite-nitrate nitrogen con-
centrations in the estuary are closely related to concentrations at
Waterville. This close correlation (j™ = .95) indicates the major contributor
of nitrite-nitrate nitrogen^to the estuary is from, the upstream heritage.
Ammonia concentration at Toledo are independent of Waterville and are usually
higher (Figure 14b). This increase of concentrations indicates the existence
of local inputs of ammonia to the estuary. Suspended solids (Figure 14c) at
Toledo are independent of and usually less than those at Waterville. This
decrease indicates settling of solids during the low flow summer period.
However, since suspended solids in the estuary reaches a minimum of about 100
mg/1 independent of the Waterville concentration, it seems for this period
that the fine clays which remained in suspension remained relatively constant
at the flows observed.
Total phosphorus .concentrations on Figure 14d are well correlated
(r^ = .56). However, some settling of phosphorus attached to suspended sedi-
ment did occur. Soluble orthophosphorus concentrations (Figure 14e) are
greater at Waterville than in the estuary. This loss seems to indicate some
biological uptake occurring in the estuary. Soluble silica concentrations
(Figure 14f) are poorly correlated between Waterville and the estuary.
Chloride concentrations (Figure 14g) are greater in the estuary than at
Waterville and indicate sources of local input to the estuary. Conductivity
measurement (Figure I4h) 'do not show much variation between the estuary and
Waterville.
CONCLUSION
This analysis demonstrates that the estuarine reach of the Maumee River
is impacted by pollutants generated in the remainder of the basin; that the
water quality.problems in the Toledo urban area are not solely the results of
Toledo's own pollution of the river. Inputs of pollutants to the river
during even moderately small summer storms can be considerably larger than
Toledo's inputs. Impacts are increased by the fact that the entire volume of
such summer storms does little more than replace a single volume of the
estuarine volume. This'polluted water often becomes.stagnant in the estuary
for many days sloshing back and forth over the organically enriched bottom
sediments. In warm weather these events multiply to severely degrade the
quality of water in the Maumee River.
299
-------
During low flow periods and small events the estuaries act as traps for
suspended solids and total phosphorus. However, the bulk of this material is
transported during a few high flow events when it is carried directly through
the estuaries.
REFERENCES
1. U. S. Army Corps of Engineers, "Lake Erie Wastewater Management Study,
Preliminary Feasibility Report," 3 Vol, Buffalo, NY, December 1975.
300
-------
WATER QUALITY AND URBAN RUNOFF IN SELECTED CANAL
COMMUNITIES ALONG THE TEXAS COAST
by
Allen L. Messenger
Hays & Lindsey, Inc.
Austin, Texas
and
Tom D. Reynolds
Texas A&M University
College Station, Texas
ABSTRACT
Water and runoff samples from seven waterfront communities in the Galve-
ston Bay area were collected and analyzed in order to evaluate causes of canal
water quality problems. Until present, community design has been based on
optimum utilization of land area with little or no regard toward the effect of
development on water quality. The primary cause of water quality problems in
these communities appears to be urban runoff.
Samples of canal waters were collected over a five-month period and
analyzed for nutrients, oxygen damand, pesticides, and hydrological variables,
including Rhodamine dye concentrations. In general, canal waters exerted
BODc values of 2-10 mg/1 with no problems associated with toxic substances.
Evaluation of domestic wastewater data from centralized treatment facilities
indicates that these wastewater streams are not major sources of pollution
loading.
Runoff samples were collected from three rainfall-runoff events and
were found to contain significant amounts of carbonaceous material. Also,
the possibility of canal sediment resuspension by point discharge of runoff
was investigated using a canal model. The two-year frequency rainfall event
for the Galveston area was found to produce significant resuspension of high
BODs benthic sediments. These data were applied using the modified Streeter-
Phelps equations for estuarine dissolved oxygen analysis and were found to fit
actual conditions with reasonable accuracy.
301
-------
INTRODUCTION
During recent years population increases along the Texas coast have
resulted in a corresponding increase in housing demand. Most housing re-
quirements are being met by conventional urban developments. However, the
proximity of bay waters along the Texas coast allows urban planners the
unique opportunity to design and build waterfront communities. The basic
purpose in the design of these subdivisions is to provide each lot with
one side of waterfront footage. Access to the surrounding waters is through
a canal network which opens into the parent body, typically a bayou, bay,
or saltwater lake.
- '' *. b
At last count, there were almost 60 coastal waterfront communities
along the Texas coast. They represent a wide range in size, design, and
cost. For example, Bayou Vista located in La Marque has 1318 lots, approxi-
mately 40,000 feet of canals of dead-end design, and homes in the $60,000
range. In contrast, Jarbo Bayou has 80 lots, approximately 5,000 feet of
canal with continuous flow provided by Jarbo Bayou, and the homes sell for
over $100,000.
There are problems associated with canal communities which, although
independent of price, are dependent on the design and size of the community.
In several of the larger communities, utilizing canal networks of dead-end
design, fish-kills and dangerously high coliform and enteric virus concen-
trations have occurred (1). As a.result of the latter condition, the public
health department has banned the harvesting of oysters from the canals. In
addition, the safety of swimming in the canals has been questioned.
Seven communities in the Houston-Galveston area were chosen for study
in order to investigate the interrelationship between pollutional loading
and canal water quality. One basis on which these communities are differen-
tiated is domestic wastewater handling and treatment. One of the communities
utilizes septic tank systems while the others have wastewater collection sys-
tems and centralized domestic wastewater treatment. Urban runoff, another
major factor influencing canal water quality, is handled in one of two ways.
In some designs, runoff is allowed to run overland into the canal along the
entire canal length. In other designs, runoff is collected in open ditches
or storm sewers and then discharged at points along the canal.
302
-------
LITERATURE SURVEY
The most comprehensive report found on the natural waters surrounding
most of the study areas is "The Galveston Bay Study" by Gloyna and Malina(2).
Included in the report is information on all the bays in the Galveston Bay
system. West Bay is the outer bay in the Galveston Bay system, and as a re-
sult, has been impacted the least by the rapidly growing Houston-Galveston
area. Six of the communities studied are oh West Bay. According to this re-
port, the waters of West Bay are relatively unpolluted. It is reported that
samples taken from the bayous and along the sides of the bay indicate pollu-
tion loads from domestic sources were common enough to cause the water to
exert a BOD5 of 2 to 4 mg/1. , .
Trent and Lindall (3) present a summary of recommendations concerning
coastal canal development including: . .
I. Minimize disruption of the existing habitat as much as possible.
1) Restrict community development to non-wetland areas.
2) Access canals to communities' canal systems should be by the
shortest and least environmentally damaging route possible.
3) Turbidity and sediment dispersion should be controlled as much
as possible.
a) Complete all construction 'prior to connecting the system
to the surrounding natural waters.
b) Access canal construction should be conducted in a manner
which minimizes sediment dispersion.
c) Spoil disposal sites and easements should be reserved for
future canal maintenance.
II. The canal systems should be designed to maintain state and federal
water quality standards, especially for dissolved oxygen.
Another paper covering the effects of coastal development on water
quality, and plant and animal productivity has been published by Taylor, ejt
al. found that canal bottoms are soft and may exceed 90-percent silt and
clay whereas in undredged.areas, the bottom is up to 94-percent sand and
shell. They found the silt and clay bottoms in the canals were "unsuitable
for most bottom invertebrates found in other parts of the bay." They also
state of a 10-year old canal system: "In 10 years, recolonization of the
canal sediments has been negligible and it appears doubtful that soft sedi-
ments of bayfill canals will ever support a rich or diverse infauna " (4).
303
-------
Moore and Trent (5) monitored one of the canal systems studied for
this report. They were concerned mainly with the canal's effect on the
oyster, Crassostrea virginica. The study is unique because it was aimed
at utilizing canal communities for oyster production. However, low dis-
solved oxygen levels in the canals resulted in a high mortality among the
oysters making it unsuitable for oyster production.
In a report by Paulson (6), some of the most comprehensive work to
date is presented on basic water quality in canal communities. Coliforms
were found to increase while dissolved oxygen levels decreased with in-
creasing distance away from the canals' inlet. The 8005 levels were low
and no differences were found between natural and altered areas. Flush-
ing rates were found to be the same in both manmade and natural systems.
It should be noted that in Paulson's work, the canals under study were
shallower than the adjacent water bodies. This condition is the opposite
of the conditions existing in the Galveston area.
RESULTS AND DISCUSSION
The communities selected for study incorporate wide variations in
design. Design variations include: canal length, inlet design, domestic
wastewater handling and treatment, and runoff routing. Data collected
from seven canal communities are presented. Comparisons are made based
on general differences in deisgn and two communities are used to examine
the effect of canal age and wastewater handling on water quality. Hydro-
logical data include TOC, 8005, salinity, conductivity, temperature, coli
forms, Kjeldahl nitrogen, nitrite, nitrate, phosphates, dissolved oxygen,
and Rhodamine dye concentrations.
A total of seven communities were monitored and the resulting data
are presented in Table I. However, more detailed data will be presented
on only two communities. For further information and data analysis refer
to the thesis from which this paper was generated (7).
Effect of Canal Age
Figures 1 and 2 are graphic comparisons of hydrological data collected
in the three sections of various ages in Bayou Vista (Figure 3). These fig-
ures compare various hydrological variables and indicate only slight differ-
ences in concentrations with no trends apparent. These- data compare well
with previous work by Hall (8).
It should be noted from Figure 3 that the oldest section is located
where it has the best flushing and that the newer canals are located where
they have minimal flushing. Nutrient levels in the 24-year old canals aver-
aged lower (refer to Figure 2) than the newer sections. Similarly, TOC and
concentrations varied only slightly between sections (Figure 1) indi-
304
-------
IO
a
0)
J2
IS
C
S-
iS
in
•r-
C
a o
Q.
•o
c cc
IO
in
•t— •
1— O
1 "
cc
5
i*
CO O
U 3
3
?=;
0 +?
CJ i-
^
CO
1
0
co
«
in
i
en
en
i— i
i
^
d
CM
1
VO
en
f-H
1
O
in
1-1
n
.,_
C
(O
01
1
o
co
1
o
S
en
in
co
CSJ .
i
o
CM
1
0
in
CO
I
O
f^
*"*
1
en
o
o
<— 1
1
o z
f. a.
"o
o
1
o
VO
1
O
-
1
VO
CM
«— 1
r-1
1
in
CO
VO
i
O
CO
CM
1
VO
CM
*"*
CM
CO
•a-
in z
f— Q_
10 s:
u
C£
s_
0)
C
IO
305
-------
30
"S. 20
10
30
o on
o d(J
10
60
50
_ 40
I 30
20
10
8
l 1 I I I i I I
SALINITY
TEMPERATURE
TOC
24 YRS OLD -
13 YRS OLD --
10 YRS OLD -
•
BOD*
i
i
\
'%. \
i i i i
i
NOV DEC JAN FEB MAR APR MAY JUN
1975 TIME, months 1976
FIGURE I. —COMPARISONS OF HYDROGRAPHIC
VARIABLES IN CANALS OF VARIOUS
AGES IN BAYOU VISTA
306
-------
o>
3.0O
2.50
2.00
1.50
.80
.70
.60
.50
.40
.30
.20
.10
.70
,5O
.30
.10
10
I ; i
TOTAL KJELDAHL
NITROGEN
1.90
NITRATE
1,50 x-
EXPANDED SCALE"
1.30
PHOSPHATE
D.O.
24 YRS. OLD
13 YRS. OLD
10 YRS. OLD
I
1975
NOV DEC JAN FEB
TIME, months
MAR
1976
FIGURE 2. —COMPARISON OF NUTRIENTS AND D.O. IN
, CANALS OF VARIOUS AGES IN BAYOU
VISTA
307
-------
ID
TREATMENT PLANT-.
a OUTFALL-
• I
•^
•.
li
•'.
a
")
o
=
^
*" *
"t •[
i ;
•' J3
/ 03
': . °
• 1
1 '
^
f*
•
•
j
;•
s
.'•1
i
i
\
. *
ff
t
\
",
j
\.VA-
oo
FIGURE 3- - BAYOU VISTA CANAL SYSTEM SHOWING THREE
STAGES OF DEVELOPMENT AND THE STATION
LOCATIONS.
308
-------
eating that canal age is not a dominant factor in determining canal water
quality when flushing and circulation are also exerting an effect.
An additional variable differentiating between the three sections is
canal depth. Section I (10-years old) ranges 9.5 to 12 feet (2.9 to 3.7m)
in depth. Section II (13-years old) ranges from 5 to 11 feet (1.5 to 3.4m)
deep with the 5 foot (1.5 m) depth occurring at the wastewater treatment
plant outfall. Typically, the minimum depth in Section II is 8 feet (2.4m).
Section III (24-years old) consists of two smaller canals of about 8 to 10
feet (2.4 to 3.0 m) in depth. All depths are from mean sea level and are
approximations.
Bayou Vista
Bayou Vista, located in La Marque, contains 1318 lots and 540 homes.
The canals are of dead-end design and are approximately 3400 feet (1036 m)
in length. The lots have 50 feet (15.2 m) of waterfront footage and are
90-feet (27.4 m) deep. All homes are elevated on pilings.
Bayou Vista consists of three sections of varying age as shown in
Figure 3. The section with sample stations H-84 and H-85 is 10-years old.
The section with stations H-82 and H-83 is 13-years old and the section
with station H-81 is 23-years old. Station H-80 is located on Highland
Bayou outside the community.
The domestic sewage is collected and treated by a wastewater treat-
ment facility and is discharged at station H-86. Discharges averaged -
97,000 gallons/day (367 m3/day) over February and March 1976 with a BODs of
27 mg/1 and suspended solids of 85 mg/1.
Storm runoff is removed by two methods. Runoff from streets and front
yards is collected in ditches and discharged at points along the canals.
Runoff from the back yards (canal footage) is allowed to flow over the bulk-
heads ,
The canal on which station H-83 is located has a history of Gulf
Menhaden (Bravoortia patronus) fish kills. The kills occur after rainfall
during the critical summer months and, according to canal residents inter-
viewed, at night or early morning. This indicated a kill caused by dis-
solved oxygen depletion in the canal water due to an algal bloom and/or the
resuspension of canal sediments.
It should be noted that one of the most popular forms of recreation
in Bayou Vista is back-yard fishing, both at night and during the day. At
night, the fishermen shine powerful batteries of spotlights into the canal
waters and fish with good success. Catches of 15 to 20 game fish have been
recorded photographically. Late night monitoring revealed oxygen levels of
12.5 mg/1 under these lights although surrounding waters were generally less
than 4.0 mg/1. Attraction of fishes is by photo- or oxygen tropism.
309
-------
Canal waters averaged 5.3 mg/1 BOD5, 10.1 mg/1 TOC, 1.98 mg/1 total
nitrogen, and 0.586 mg/1 orthophosphates. Nutrient levels in the runoff
samples averaged slightly lower than the canal waters and BODg levels aver-
aged 58-percent higher in the runoff samples.
Dissolved Oxygen Profile
The largest single problem with the canals in any canal community is
oxygen depletion in the canal water. There are several contributing factors:
1) low or non-existent circulation,
2) poor tidal flushing,
3) loading into the canals from runoff, and
4) loading into the canals from septic tank and treatment plant
wastewaters.
The oxygen depletion which results from certain combinations of these
factors is the result of an increased algal population which produces appre-
ciable amounts of oxygen during the day but likewise consumes large amounts
of oxygen at night.
Figure 4 is a graphic summary of dissolved oxygen data collected in
Bayou Vista during July. It was collected from 11:00 p.m. to 1:00 a.m. The
top graph shows the dissolved oxygen concentration at surface and the lower
graph indicates the depth at which the oxygen level was less than 1.0 mg/1.
The data clearly show the relationship between canal length and dis-
solved oxygen concentration. The further down the canal from the inlet,
the lower is the concentration. In all three canals there was greater than
1 mg/1 of oxygen at the inlet at all depths monitored but there exists a
continuous decrease in the depth at which the dissolved oxygen goes to 1
mg/1 inland from the inlet.
Dye Study
In order to obtain data on flushing in dead-end canals, a dye study
was performed in Bayou Vista during July 1976. In the study, one gallon of
20-percent Rhodamine WT was added to the canal at station H-85 and mixed to
an approximate depth of 4 feet (1.2 m). Dye concentrations were measured
at 6, 12, 18, 24, 30, and 42 hours after dumping. Due to equipment limita-
tions, no tidal measurements were taken during the study although the diur-
nal tides are generally less than 1 foot (0.3 m) (9). Dye concentration was
monitored at seven stations along the canal with samples taken at top, 3-foot
(0.9 m) and 6-foot (1.8 m) and 9-foot (2.4 m) depths.
At 10 hours after the study began, a brief thunderstorm occurred.
Winds gusted to 35 miles per hour (56 m/hr) blowing parallel to the canal
310
-------
CO
to
00
Q
UJ
sf
ro
, i
OJ
«•'
^
_J
z
o
0
1
UJ
o
z
2^
en
o
UJ
_j
UJ
o:
i
i
z
UJ
— '
0
^
E
CD
2
^
X
UJ
CO
_J
^^
^J
o
ff
\-
co
' >
o
CD
UJ
UJ
tr
X
i-
u.
o
to
UJ
_J
u.
o
o:
a.
E
q
v/
CD
c5
O
CO
•z.
o
~
5
_J
UJ
cr
X
h-
0.
UJ
Q
Q
5
o
Ci
CL
O
h-
"Z.
CO
**
UJ
CD
U.
doi- (I/BUJ) N39AXO baAiossia
l/6ui O'T > .
HOIHM IV (»} HlcOQ
311
-------
WEST BAY
.330
CARANCAHUA
COVE
GULF OF MEXICO
FIGURE 6. - JAMAICA BEACH CANAL COMMUNITY SHOWING
COLLECTION SYSTEM AND SEPTIC TANK AREAS
AND STATION LOCATIONS.
314
-------
30
10
3O
o° 20
10
60
40
30
2O
10
8
^ 6
o>
1 II
SALINITY
TEMPERATURE
OF COLLECTION
AREA OF SEPTIC TANKS
TOC
BOO
I
|
I
I
1
NOV DEC JAN FEB MAR APR MAY JUN
1975 TIME, months 1976
FIGURE 7. - COMPARISONS OF HYDROGRAPHIC
VARIABLES BETWEEN AREAS OF A
COLLECTION SYSTEM vs. SEPTIC TANKS
IN JAMAICA BEACH
315
-------
.30
.20
JO
1.20
1.00
,8O
.60
.70
.50
.30
.10
10
NITRATE
TOTAL KJELDAHL
NITROGEN
PHOSPHATE
D.O.
COLLECTION SYSTEM
SEPTIC TANKS
I
NOV DEC JAN FEB
1975 TIME, months
MAR
1976
FIGURE 8. - COMPARISONS OF NUTRIENTS AND 0,0.
BETWEEN AREAS OF A COLLECTION SYSTEM
vs. SEPTIC TANKS IN JAMAICA BEACH
316
-------
Rainfall Runoff
Two methods of routing rainfall runoff are currently utilized in
canal communities. One method is to allow runoff to flow overland into
the canals without collection. This technique works well in canal com-
munities because of the relatively short distances involved. The second
approach is to collect runoff in open ditches or storm sewers and then
discharge the runoff at points along the canal. The practice of collect-
ing runoff intensifies water quality problems. Point discharge of runoff
concentrates runoff pollution where allowing runoff to flow overland dis-
tributes the loading along the canals's entire length. Point discharge
also increases the possibility that sufficient energy will be imparted to
the canal waters to resuspend bottom sediments. Resuspension of bottom
sediments in effect increases pollution loading due to urban runoff. .
P.O. Profile for Canal H-83 in Bayou Vista
As discussed previously, the Bayou Vista development in La Marque has
had problems with several fish kills occurring in Canal H-83. Residents
report that kills occur during the summer after rainstorms and at night.
These conditions indicate kills resulting from depletion of oxygen in the
canal waters. , ,
In order to model the dissolved oxygen depletion that occurs during
a rainfall runoff event, the equations and methods for estuarine analysis
as presented by O'Conner and Eckenfelder (10) are applied to Canal H-83.
Certain assumptions were made in applying this technique where data were
inadequate or where the basic equations were not applicable. All calcula-
tions were made using USC6 units, althougn SI units can be easily applied.
Runpff to be conisdered was from the area of land surrounding the
upper half of the 3400-foot (1036 m) canal and rainfall was assumed to be
2.25 in/hr (5.72 cm/hr) which is the 2-year frequency. Application of the
Rational Method for predicting runoff gives the runoff as Q = 5.26 cfs or
0.454 MGD (1718 nr/day) for a time of concentration of 30 minutes, and a
coefficient of runoff of 0.25.
At this flow rate the velocity, V, of the water in a canal 1050 ft2
(97.5 m2) cross-sectional area (100 ft wide and 10.5 ft deep) (30.5 m by
3.2 m) is equal to 18.0 ft/hr (5.49 m/hr). All four constants, that is
K£> E, Jp and J?> required to calculate the dissolved oxygen distribution
were calculated (see thesis for detail (7)). The oxygen deficit equation is:
D =
-J,X
-J0x
K2 ~ Kd
- e
D0e
(D
where LQ is the BOD of the, canal water, D is the initial oxygen deficit,
and x is the distance from the point of discharge. Insertion of data gives,
317
-------
D =
0.980
0.032 - 0.980
(2)
Application of Dissolved Oxygen Sag Equation
The sag equation was applied to two cases. The first case simulates
conditions in Bayou Vista where the sole source of 6005 is urban runoff.
This case is equivalent to point discharge of urban runoff where bottom sed-
iment is resuspended by the turbulence created by point discharge of the
runoff.
Case I. Initial calculations were conducted us in
collected during this study which averaged 14.3 mg/1 (L
runoff BOD values
and Dm (initial
D.O. concentration) which averaged 4.6 mg/1. Note that data use*d to deter-
mine L were not of first runoff and, therefore, the oxygen depletion result
ing from application of these data is less than might be expected under ac-
tual conditions. The D.O. sag equation was applied using these data and an
assumed critical dissolved oxygen level of 2 mg/1 was reached in. 30 hours.
Under actual conditions, residents report kills occurring 12 to 15 hours
after rainfall-runoff events; therefore, the model .used does not describe
the expected conditions accurately. Thus it became necessary to investi-
gate further into the problem to account for sediment resuspension.
Case II. Previous work by Hall (8) on the canal sediments indicates
a strong possibility that the point discharge of runoff into the canals is
sufficient to cause sediment resuspension. In order to investigate this
possibility, a hydraulic model of Canal H-83 was designed and constructed.
The width and depth of the model were 1/25 that of the actual canal.
The model was based on providing a plug-flow travel time of 12 hours. The
model measured 4 ft (1.2 m) by 6.5 ft (2.0 m), respectively.
In order to simulate the scouring action of runoff in the canal sedi-
ments, the Reynolds Number was used in the form Re=Vd/v, where V is the
velocity in ft/sec, d is the characteristic depth, and v is the viscosity.
The characteristics depth, d, used was sediment particle diameter. Since
sediment from Canal H-83 was used in the model, partical diameter, d, was
fixed. In addition, viscosity, v , was constant from model to canal for a
given Reynolds Number. Field measurements indicated that the gravity head
for runoff discharge into the canals was about 2.5 ft (0.8 m). The result-
ing velocity was 12.5 ft/sec (3.8 m/sec) which was used in the model.
The last scale-down required was the calculation of runoff flowrate.
This was done by determining the energy per unit volume imparted by runoff.
The energy is equal to mass flow per unit time multiplied by the elevation
318
-------
head. The runoff for one quarter of the drainage area which discharges run-
off into the canal was used to calculate mass flow per unit time. Since the
finished grade is about 2.5 ft (0.8 m) above the canal water surface, an
elevation head of 2.5 ft (0.8 m) was used.
The scale-down factor for energy can, therefore, be determined simply
by the ratio of model volume over canal volume. The flow rate is then cal-
culated by dividing the scaled-down power by elevation head and converting
mass flow rate to volumetric flow rate. The resulting volumetric flow rate
was 210 ml/min. The discharge outlet was then sized to provide 210 ml/min
at 12.5 ft/sec (3.8 m/sec).
The model was built to operate according to the criteria established
above. Concrete was used to form the shape of the canal bottom and the
canal sediment was distributed evenly over the bottom with a depth of about
1 inch (2.54 cm). The model was then filled with water and left for one
day to a-llow suspended solids to settle.
To run a test, the pump which provided the runoff flow was started
and samples were collected of the water close to the point of discharge.
The samples were then analyzed for suspended solids.
Work by Hall (8) on the - canal-sediments was then used to'convert the
suspended solids data to BODU. According'to Hall, the maximum oxygen de-
mand for Bayou Vista sediments was 17.94 mg 02/g sediment. Suspended solids
data collected from the model had an average value of 1.25 g/1 suspended
solids. Therefore, if 1.25 g/1 of bottom sediment are suspended by runoff,
the oxygen demand^ Lg, of the suspended is'the initial oxygen demand of the
canal water, 14.3 mg/1, plus, the oxygen demand of the sediment.
-. . L0 = (17.94 mg/g)(l.25 g/1) + 14.3 mg/1
= 36.7 mg/1
After the model testing, all variables and constants required to solve
the sag equation, Equation 2, were known. Figure 9 is.a summary of results
from application of the sag equation and indicates good comparability with
conditions reported by residents. The critical oxygen deficit of 2 mg/1
was reached in 11.4 hours versus the 12 to 15 hours reported under actual
conditions.
It should be noted that application of the D.O. sag equation requires
steady state conditions. Therefore, strict application of the equation re-
quired that runoff be assumed to be constant over the required application
period.
This is not, however, an unreasonable assumption. Although rainfall
intensity decreases with duration, the rate of runoff decreases more slowly
with time due to a corresponding decrease in the infiltration rate. Calcu-
lations for the Galveston area indicate, that at 11.4 hours the rate of run-
off for a 2-year frequency rainfall event will be 65 percent of the runoff
319
-------
ro
-------
rate at the minimum time of concentration. Thus, application of the modi-
fied Streeter-Phelps sag equation in the manner described approximates ac-
tual conditions encountered in the study area.
CONCLUSIONS
1. Septic tanks are not a suitable form of domestic waste treatment
in canal communities based on coliform levels in,the septic tank area of
Jamaica Beach (7,8).
2. Domestic wastewater should be collected .and treated at a community
wastewater treatment plant. The discharge from the plant should be located
so that it does not pass into the canal system.
3. Water quality in canal communities is strongly influenced by the
circulation of the surrounding natural waters through the system. To en-
hance 'circulation, the following criteria should be observed.
a) Canals should be oriented to maximize wind flushing, i.e.,
aligned with the prevailing winds.
b) Canals should not be dredged any deeper than is required to
allow safe boating. In most cases an 8-foot (2.4 m) depth
will be sufficient. If at all possible, the canal depth
should not be greater than the parent body of water.
c) Complex canal systems with restrictive inlets should be
avoided.
d) Designs which provide for circulation of the surrounding
waters through the canals should be encouraged.
e) The canals should be of uniform depth, however is possible,
the bottom should gradually slope from the furtherest point
inland to the parent body of water. This would help in pre-
venting stagrient water pockets.
4. If the canals are designed with sufficient circulation, the
runoff should not be collected. Instead, the roads and lots should be de-
signed with sufficient slope to remove runoff by overland flow into the .1
canals. Point discharge of runoff concentrates pollution and may cause T
sediment to be resuspended from benthic deposits. This design will not/
create the mosquito breeding problems of an open ditch collection system.
5. If runoff is collected for point discharge, drop boxes should be
used to dissipate excess energy prior to release into the canals in order
to prevent suspension of benthic sediments.
321
-------
6. Dead-end systems should be avoided if possible. When used, dead-
end canals should be connected by culverts to enhance circulation. Dead-
end canals of 3400 feet (1036.3 m) have had fish kills occur due to oxygen
depletion.
7. Storm-water runoff into the canals should be minimized as much
as possible and runoff from adjacent areas should not be discharged within
a canal led community.
8. Concrete bulkheads should be used instead of timber. Timber bulk-
heads are susceptible to collapsing which allows soil to enter the canal
and create stagnent water pockets.
REFERENCES
1. Goyal, S. M., ejb a]_., "Prevalence of Human Enteric Viruses in Coastal
Canal Communities." Journal Water Poll. Control Fed., 50, 10 (1978).
2. Gloyna, E. F., Malina, J. F., Or. "Galveston Bay Water Quality Study,
Historical and Recent Data." The Texas Water Pollution Control Board.
Austin, Texas (March 1964).
3. Linda!!, W. N., Jr. and Trent, L. "Housing Developments in the Coast-
al Zone of the Gulf of Mexico: Ecological Consequences, Regulations,
and Recommendations." MFR Paper 1163 (1976).
4. Taylor, John L., and Saloman, C. H. "Some Effects of Hydraulic Dred-
ging and Coastal Development in Boca Ciega Bay, Florida." Fishery
Bulletin 67:213-214 (1968).
5. Moore, D. and Trent, L. "Setting, Growth and Mortality of Crassostrea
yirginica in a Natural Marsh and a Marsh Altered by a Housing Develop-
ment." National Shellfisheries Assoc., Vol. 61 (June 1971).
6. Paulson, 0. L. and Pessoney, G. F. "Residential Canals along the Gulf
Coast." National Technical Information Service, PB-243-779 (1975).
7. Messenger, A. L., "Water Quality and Urban Runoff in Selected Canal
Communities along the Texas Coast." Thesis,Texas A&M Univ. (1979).
8. Hall, Ernest. "Water Quality of Some Coastal Canal Communities in the
Galveston County Area." Thesis, Texas A&M Univ. (1976).
9. United States Coast Pilot 5 Atlantic Gulf Coast of Mexico, Puerto Rico
and Virgin Islands."" 9th Ed., NOAA Nat. Ocean Survey.(July 1976).
10. Eckenfelder, W. W., Jr. and 0'Conner, D. J. Biological Waste Treatment.
Pergamon Press. New York, New York (1961).
322
-------
Sixth Session
ECOLOGICAL RESPONSE TO STORMWATER
Moderator: James S. Taylor
University of Central Florida
Orlando, Florida
323
-------
THE RESPONSE OF SUBTIDAL INFAUNAL COMMUNITIES
TO A CHANGE IN WASTEWATER DISCHARGE 1
Ronald M. Thorn and Kenneth K. Chew
College of Fisheries
University of Washington
Seattle, Washington 98195
ABSTRACT
Certain aspects of the structure of macroinfaunal communities in the sub-
tidal sediments adjacent to an intertidal combined sewage over (CSO) were
studied during 1978. The impact of this ephemeral source of raw wastewater
and street runoff on the communities was evaluated using samples taken after
periods of highest (April) and lowest (August) frequency of discharge. In
April, the community nearest the overflow was characterized by a high number
of individuals, a low number of taxa, a low species diversity, a high abun-
dance of the polychaete Capitella capitata, and a high relative number of
polychaetes. Subsurface deposit feeding species dominated the community.
There was an area of markedly low infaunal abundance bordering the region
of acute impact. Diversity and number of taxa were highest at the sites fur-
thest (i.e. > 1000 m) from the overflow. Bivalve molluscs were in relatively
high abundance, and carnivores, surface deposit feeding and subsurface de-
posit feeding species were approximately equally abundant at these sites.
The samples in August showed similar trends relative to the overflow in
number of individuals, number of taxa, species diversity and community comp-
osition by species, phyla, and feeding type. However, differences among all
sites in these parameters were less pronounced during this period. A notable
between season difference was the switch in numerical dominance from C. cap-
itata to the leptostracan crustacean Nebalia pugettensis at the site nearest
the overflow. This switch may be related to "decreased disturbance by effluent
flows near this site prior to the August Sampling.
A cluster analysis indicated that depth was of primary importance, and
that the CSO was of secondary importance in producing among site differences.
The communities at the shallowest sites and northward from the overflow were
the most altered by the effluent.
We concluded that the decreased frequency of overflows in August may be
responsible for the relative decrease in differences among the communities
at all distances from the overflow. Impact studies should be conducted over
1
Contribution No. from the College of Fisheries, University of Washington,
Seattle, Washington 98195
324
-------
several seasons in areas where substantial variations in wastewater and/or
natural environmental parameters occur.
INTRODUCTION
Most of the sewers in the city of Seattle, Washington, carry both sewage
and street runoff. During rainy periods, sewage treatment plants become
overloaded and some of the material in the sewers is released through combin-
ed sewer overflows (CSO). The CSO discharging the greatest volume of sewage
into marine waters in Seattle is located at Denny Way in Elliott Bay, Puget
Sound (Fig. 1). Overflows through this CSO occur approximately 38 times a
year with an average volume per overflow of 2.8 x 10° gal (10.6 x 106 1)
(Municipality of Metropolitan Seattle - METRO, unpublished data). The over-
flows are most frequent, of greater volume, and have the highest flow rates
during winter and early spring.
Effluent is discharged through a conduit located within a small cove in
the littoral zone. The conduit is surrounded by a steeply sloping boulder
sea wall, which extends approximately 200 m south and 1400 m north from this
point. The path of the effluent is marked by a scoured zone in the soft
sediment which forms the substrata in the cove directly seaward from the base
of the conduit.
Armstrong et^aJL (1978, 1980) studied the impact of this source of pollu-
tion on the adjacent intertidal and subtidal epibiotic and infaunal commun-
ities during spring (ie. April-May). They documented significant changes in
the structure of the communities, and these were measured as changes in
species richness and diversity, community composition, and trophic require-
ments. Armstrong et al_. (1980) were unable to locate in the literature sim-
ilar studies near CSO's, but did conclude that the modifications they saw
were in line with those reported for areas subjected to continuously flowing
'effluent from intertidal and subtidal outfalls (e.g. Reish 1959, Littler and
Murray 1975, Stevenson et. al_. 1975, Smith and Green 1976). Arfcstronget al.(1978)
recommended that, due to the variable nature of the CSO discharge, s^uoTes
be conducted during several seasons to document any seasonal aspects of the
CSO impact on the biological communities.
The purpose of the present paper is to compare the structure of subtidal
infaunal communities near the Denny Way SCO in an extreme case; that is,
shortly after periods of maximum and minimum number of overflows. Therefore,
we sampled 14 subtidal sites during April and August, 1978, and followed the
methods of Armstrong, ejt aL (1978) for sample processing and data analysis.
MATERIALS AND METHODS
Sampling sites were located at 9 and 13 m depth contours along seven tran-
sects which ran approximately perpendicular to the shoreline. Transects 1,
2, 3, 4, 6 and 7 were positioned 1270, 320, 260, and 180 m north and 150
and 440 m south of the CSO, respectively. Transect 5 was located immediately
seaward from the CSO.
Two Van Veen (0.1 m2) grab samples were taken at each site using the R/V
HYDOR. The texture, odor and color of the sediment was described, and a
small (3-5 g) subsample was extracted for volatile organics analysis. The
325
-------
•45'
STUDY VS: .•'.••-
AREA
22'
Fig. 1. The location of the study area.
326
-------
samples were live screened with gentle shaking in water through a 1 mm mesh
sieve on board. The material retained on the sieve was placed into glass jars
and preserved and stained in 10% formalin and Rose Bengal stain. In the lab-
oratory, the samples were rescreened (1 mm mesh) and rinsed with tap water
to remove the formalin. All organisms retained on the screen were identified
(usually to species) and counted. '••'"" ' *
Volatile organic content of the sediment was determined by drying the 3-5g
subsample for 30 hours at 100° C, recording the weight, burning the material
for 3 hours at 550° C, and reweighing. The loss of weight, expressed as a
percentage of the dry weight, gives a rough estimate of the organic carbon
content of the sediment (Morgans 1956).
Species diversity at each site each season was estimated using the com-
bined data from the replicate samples. The measure of Hurlbert (1971) was
used for this estimate, and it is calculated by the formula
,
"
(n)
= expected, number of species in a sample of n
individuals selected randomly from a collect-
ion of N individuals, S species, and N. in-
dividuals in the ith species.
i ' *... t -
Curves for the two most ab.undant taxocenes (ie. polychaetes, molluscs) re-
lating E(S ) to several values of n were constructed for each site each
season. > . .;
Species composition at the sites was compared using classification analy-
sis (Boesch 1977). The species abundances at each site were square root
transformed. Intersite .dissimilarities were determined using the Bray-Curtis
coefficient of dissimilarity (Clifford and Stevenson 1975). The sites were
then clustered using the group average sorting strategy (Sokal and Michener
1958). The results of the classification analysis are displayed as a dendro-
gram.
RESULTS
During the three months pipior'to the April sampling, 25 overflows occur-
red (total volume =275 x 10b 1), with an average flow rate of 45 x 103 1/min.
Three months prior to the August sampling, five overflows took place (total
Volume = 45 x 106 1) at an average rate ;of 34 xMO3 1/min.
The sediment collected in the grab samples graded from light brown sand at
transect 1 to fine black silt at transect 5 (Table 1). Sulphide odor and
volatile organic content were generally highest in sediment at the sites
nearest the CSO (Table 1). Sulphide levels generally decreased, and sediment
color lightened, at the sites south of the CSO. In April, black sediment
with a moderate sulphide odor was encountered at the 9 m site on transect 2,
which is located relatively far from the CSO. This site is in front of a
small cove, and this cove, may contribute to the development of eddy currents
and the deposition of suspended material.
Infaunal densities differed appreciably among the samples from 9 m in
327
-------
Ol
o
LOOOOCVl«*COf^«3-OOOOVOCOCOCM i— CO i— CO
i—O>
re
3S
>>
c
g
0
cu
.c:
•«->
S-
0)
4->
to
0)
o>
QJ OJ C7)
C C = T- :
O O •—
c c: to
•o
cu
o
to
.a
c
o
O = O :
(O
I— -)->
^a r—
-p
o en
S_= -r- S
+J r—
in
O
(B
o
o
to
.a
I
a> =
a)
S.
en
o =
rd
o
s-
CO
to
•M
c
•a
0)
to
a>
-»->
x
o
TCI
re(
o
o
a>
ia
i— C>J
co
i—
-------
April (Fig. 2a). The pollution indicating polychaete Capitella capitata ac-
counted for the majority (76%) of the individuals found at the site nearest
the CSO. Total density generally declined at sites immediately north and
south of the CSO, and increased at the sites furthest from the CSO. This
trend was also seen in August, although abundances near the CSO were not as
high as in April. C. capitata was in low abundance at all sites in August,
and the crustacean Nebalia pugettensis comprised the majority (52%) of the
individuals on transect 5. Density values were notably low at the 9 m site
on transect 2 during both seasons.
The highest densities along the 13 m contour in April occurred at sites
on the transects immediately north (4) and two transects south (7) of the
CSO (Fig. 2b). The bivalve mollusc Axinopsida serricata was numerically dom-
inant (44 and 59% of the total number of individuals on transects 4 and 7
respectively) at these sites. Very few (total=15) individuals of C. capitata
were found in the samples from 13m.
Density values for the samples taken at 13 m in August were substantially
different from those in April (Fig. 2b). Relatively low densities occurred
at the sites on transects 4 and 7, where values decreased by approximately
72 and 58% respectively. The highest densities occurred far from the CSO,
and these values were approximately double those noted in April at these
sites. The site on transect 5 showed the least change in this parameter.
The number of taxa found at both the 9 m (Fig. 2c) and 13 m (Fig. 2d)
differed among sites, and exhibited a general decline at decreasing distance
foom the CSO. An exception to this was the samples from the site nearest
the CSO. The higher number of taxa at this site may be related to the higher
number of individuals that were collected. These data differed between
seasons, although the trends were similar. The flux in number of taxa between
samplings was most pronounced on transects 5 and 6 at 9 m, and on transect
4 at 13m. The greatest number of taxa were consistently found in the
samples from transect 1.
Mollusc species diversity (as measured by curve height) was lowest at the
sites nearest, and immediately south of, the CSO at 9 m in April (Fig. 3a).
Mollusc diversity was correlated positively with distance from the CSO at
the 13 m contour during this season (Fig. 3b). In August, mollusc diversity
at 9 m was similar among all sites (Fig. 3c). Lowest diversities were en-
countered at sites nearest the CSO at 13 m during August.
Polychaete species diversity decreased with decreasing distance from the
CSO at 9 m in April (Fig. 4a), and showed a similar but less pronounced trend
at 13 m. Several sites at 9 m grouped with the site closest to the CSO in
August, which indicated similarities in polychaete diversity (Fig. 3c). The
site furthest from the CSO was again highest in this parameter. A gradient
relative to the CSO may be indicated at 13 m by the relatively low diversity
at the site on transect 4 and the relatively high diversity at the site on
transect 1 (Fig. 4d).
The classification of site samples taken in April revealed the primary
importance of depth in controlling the species composition, and the secondary
but substantial effect the CSO effluent had on the communities at some sites
(Fig. 5a). Depth appeared to be not as important in August, however the in-
fluence of the CSO was still evident during this month (Fig. 5b). Five sub-
groups were designated by letters, and these letters were plotted on maps
of the sampling locations (Figs. 6a, b). If transect 1 can be regarded as a
329
-------
500
; 400
100
APRIL'
111 INDIVIDUALS AT 9m DEPTH
CSO
SOD
400
/\ A-
APRIL!"
(b) INDIVIDUALSat 13m DEPTH
CSO
AUGUST
(d) TAXA AT 13m DEPTH
CSO
1 234567
TRANSECT
Fig. 2. Density of individuals and taxa found in the
subtidal samples from near the Denny Way Regu-
lator outfall during April and August, 1978'.
330
-------
(a) APRIL, 9m DEPTH
O 0
Q
UJ 10-,
250 325
(c) AUGUST. 9m DEPTH
10-T—
100 150
NO. OF INDIVIDUALS
200 250 275 368
Fig. 3. Species-richness curves for molluscs
collected near the Denny Way Regulator
outfall during April and August, 1978.
The numbers indicate sampling transects.
331
-------
30
K
a
g
-4
(a) APRIL, 9m DEPTH
60 100 150 200 250 300 1553
SO 100 ISO 200 250 300
30-
10
(c) AUGUST, 9m DEPTH
50 100 150 200 250 300
20
10-
(d) AUGUST. 13m DEPTH
50
100 150 200
NO. OF INDIVIDUALS
250 300
Fig. 4. Species-richness curves for polychaetes
collected near the Denny Way Regulator
outfall during April and August, 1978.
The numbers indicate sampling transects.
332
-------
SOT
(a) APRIL
A B
20
10-
13/4 13/7 13/3 13/3 13/2 13/1 9/7 9/6 9/3 13/5 9/2 9/1 9/4 9/5
501
Ib) AUGUST
30-
10-
13/1 13/3 9/1 13/6 13/2 9/1 13/5 13/7 9/3 13/1 9/2 9/7 9/6 9/5
SITES
Fig. 5. Dendrogram of subtidal infauna samples
from near the Denny Way Regulator
outfall, 1978. The site designations
are given as depth (m)/transect.
333
-------
(a) APRIL
Fig. 6. Positions of subgroups of sites from cluster
analysis of samples of subtidal infauna
collected near the Denny Way Regulator outfall,
1978.
334
-------
control, then modifications caused by the CSO effluent and evident in April
are seen at sites 9/2, 9/3, 9/5, 13/5, and 9/6. The gradient of influence at
depth (13/5) and distance (9/2, 9/3) is apparent. Interestingly, the commun-
ity along transect 4 appears unmodified.
August samples revealed that the CSO affected infaunal community compo-
sition at some sites different than those affected in April (Fig. 6b). Sites
9/5, 9/2 and 9/6 were still modified, and sites 13/4 and 9/7 appreared to
also be influenced. Lower flow rates prior to the August sampling coupled
with natural northward flowing currents may have resulted in the effluent.
being carried to the 13 m site on transect 4 rather than affecting the 13 m
site on transect 5.
A graphic display (after Snee 1974) of the proportions of individuals
among three major phyla found at each site provides a partial indication of
the influence of the CSO on the benthos (Fig. 7a). Infauna composition at
9 m on transect 5 was significantly different (P<0.05) from that at the
other sites during both April and August. The April and August samplings at
this site were also significantly different from one another. Most of the
individuals in April were polychaetes. Conversely, a large proportion of
Crustacea comprised the samples from this site in August. The samples taken
at 9 m on transect 1 were significantly different from one another, but they
did exhibit an infaunal composition that was relatively higher in mollusc
individuals and lower in numbers of polychaete individuals. The sites
closest to the CSO had increased ratios of polychaetes to molluscs.
A change between seasons in these data from 9 m was evident. Samples
from the same site were significantly different between samplings except at
transect 2. The greatest difference (i.e. as measured by distance between
points on the graph) is between samples at sites on transects 4, 5, and 6.
Considering transect 1 a control area, where seasonal changes can be regarded
as natural, it can be concluded that the effluent from the CSO tends to in-
crease between-season differences in the infauna at 9 m.
The graphical analysis of site from 13 m revealed that there were no sig-
nificant differences among sites or between seasons at a site (Fig. 7b).
However, the August samples from transect 4 were disproportionately high in
polychaetes which may suggest a possible influence of the CSO at this depth.
Polychaete feeding strategies are a reflection of the characteristics of
the sediment and overlaying water (Jumars and Fauchald 1976). Burrowing
(subsurface) deposit-feeding polychaetes were proportionately most abundant
near the CSO (Fig. 8a). At the site at 9 m on transect 1, the three major
feeding types were in approximately equal abundance (Fig. 8a). These results
were true for both the April and August samples; however, the differences in
proportion of feeding types among sites were generally greatest during April.
Significant between season differences were seen at sites on transects 4, 5,
6, and 7. Samples at sites on transect 6 and 7 in August contained large
numbers of the surface deposit-feeding polychaete Prionospio steenstrupi
which accounted for the position of these samples on Fig. 8a. At 13 m, there
were no significant differences among samples (Fig. 8b). The samples from
April tended to have proportionally more burrowing deposit feeders, however.
DISCUSSION
Effluent from the Denny Way CSO significantly affects the adjacent shallow
subtidal infaunal communities, and some aspects of this effect changed with
season. Frequent overflows of high volume and rate resulted in an increase
335
-------
(a) 9m DEPTH
ARTHROPODS
(b) 13m DEPTH
ARTHROPODS
Fig. 7. Proportions of the total numbers of annelids,
arthropods and molluscs sampled near the
Denny Way Regulator outfall, April and August,
1978. The numbers designate transects; those
in parentheses are for August. The circles
indicate 95% confidence intervals.
336
-------
100H
ounnowmo DEPOSIT FEEDERS
100K
CARNIVORES
100*
8URROWINS DEPOSIT FEEDERS
(•) 9m DEPTH
SURFACE DEPOSIT FEEDERS
Ib) 13m DEPTH
1QOX
CARNIVORES
SURFACE DEPOSIT FEEDERS
Fig. 8. Proportions of total numbers of polychaetes
within each of three feeding-type categories, for
samples collected near the Denny Way Regulator out-
fall, April and August, 1978. The numbers designate
transects; those in parentheses are for August.
The circles indicate 95% confidence intervals.
337
-------
in infaunal density, an increase in the abundance of the pollution indicating
polychaete Capitella capitata, reduced mollusc and polychaete diversities,
raised relative proportions of annelids as compared to other phyla, and a
predominance of burrowing deposit feeding annelids at the site closest to the
CSO. The effect was reduced, but evident, at sites immediately beyond this
area and at some sites to the north. Data on sediment quality corresponded
with these results, which suggest that the CSO effluent is responsible for
the modified conditions near the discharge.
After a period of infrequent overflows, the differences in community par-
ameters among sites were less pronounced. The greatest change in the commun-
ity between seasons occurred at the site closest to the CSO. Mollusc and
polychaete species diversities remained low; however these values resembled
those for sites immediately north and south of the CSO. Burrowing deposit
feeding annelids were proportionally most abundant at 9 m sites (on transect
3 and 5) near the CSO, and the proportions of annelid feeding types changed
significantly between seasons at all sites except those on transects 1, 2,
and 3.
In August, the leptostracan crustacean Nebalia pugettensis replaced C.
capitata as the numerically dominant organism at the site closest to the
CSO.This switch in dominance may at least be partially explained by a
change in habitat characteristics and food quality in the area. During high
flows, heavy deposition of organic debris favors subsurface deposit feeding
organisms such as C. capitata. When flows are less frequent, deposition
and disturbance are correspondingly reduced. N. pugettensis is a surface
dwelling organism that will swarm to areas containing large amounts or organ-
ic debris (Schmitt 1965). Because disturbance and deposition are reduced
in August, this species is not buried or washed away from the area. The
decline in the^C. capitata population may be due to natural die off, compr
etitive inferiority to N. pugettensis, or changed environmental conditions
near the CSO.
Classification analysis results showed that depth produced the primary dif-
ferences in community composition among sites and that effluent flows were of
secondary importance. If the transect furthest from the CSO can be regarded
as a control, stratification by depth in species composition by natural phen-
omena exists in spring but not in summer. Evidence from the classification
and the proportional analyses suggests that the effluent has its primary
effect at the 9 m sites on transects 2 and 5. Sites located between these
latter two are somewhat less affected. Sediment characteristics indicate
that effluent material is being deposited in greater quantities on transect
2. It has been shown that a lens of freshwater exists near the CSO during
overflows and that the water in the lens is turbid with effluent particulate
(METRO, unpublished data). Northward flowing currents may transport this
material along the shore until mixing (possibly influenced by the small
cove near transects 2 and 3) of the waters causes the material to be de-
posited on the bottom.
The patterns in the parameters relative to CSO disturbance follow the
models summarized by Stirn et al_. (1975) for subtidal infaunal communities.
Recent work has indicated that these patterns are largely explained by the
relative ability of the members of the local species pool to feed successful-
ly in the disturbed area (Word 1978). Burrowing deposit feeding species
species can withstand the constant rain of organic matter, whereas surface
338
-------
deposit and filter feeding organisms are not well suited for this type of
habitat. Word (1978) has developed an index that utilizes the relative
abundances of selected infaunal taxa of various' feeding types to illustrate
the degree of organic enrichment of benthic areas near sewage outfalls. The
index is highly sensitive to changes in organic content of the sediments and
shows that burrowing deposit feeding species predominate in areas of sewage
deposition. The patterns in feeding types demonstrated by Word in southern
California agree well with those seen by us near the Denny Way CSO.
We conclude that shallow subtidal infaunal communities subject to substan-
tial variations in wastewater flows will respond to a measurable degree in
concordance with these variations. The areas closest to the wastewater source
will exhibit the greatest fluctuations, and the magnitude of the temporal
changes will decrease with decreasing degree of influence (eg. distance).
The resiliency of the benthic community we studied was evidenced by a slight
recovery of the severely impacted areas during the period of infrequent dis-
charges. We suggest that studies evaluating the impact of human disturbance
on the marine environment, in regions where significant seasonal variations
in physical-chemical parameters occur, should be conducted during several
seasons. Results will vary by season in these areas and conclusions drawn
on one season of sampling may be misleading regarding the overall impact of
the pollution source on the environment.
ACKNOWLEDGEMENTS
We gratefully acknowledge G. Farris and R. Tomlinson of METRO for their
assistance and technical support. We thank J. Packer and K. Li for their
help with the field and lab work. C. Staude kindly verified the amphipod
identifications. The typing and editorial assistance of J. Norton is greatly
appreciated. This work was supported by a grant from the U.S.: Environmental
Protection Agency through the Municipality of Metropolitan Seattle (METRO).
LITERATURE CITED
Armstrong, J.W., R.M. Thoma K.K. Chew, B. Arpke, R. Bonn, J. Glock, R.
Hieronymus, E. Hurlburt, K. Johnson, R. Mayer, B. Stevens, S. Tettlebach,
and P. Waterstrat. 1978. The impact of the Denny Way combined sewer
overflow on the adjacent flora ,and fauna in Elliott Bay, Puget Sound,
Washington. A report prepared in cooperation with the Municipality of
Metropolitan Seattle, 102p. (unpublished).
Armstrong, J.W., R.M. Th-om, and K.K. Chew. 1980. Impact of a combined sewer
overflow on the abundance, distribution and community structure of sub-
tidal benthos. Marine Environmental Res. (in press).
Boesch, D.F. 1977. Application of numerical classification in ecological
investigations of water pollution. U.S. Environmental Protection
Agency, EPA-600/3-77-003, 115p.
Clifford, H.T. and W. Stevenson. 1975. An introduction to numerical classi-
fication. Academic Press, New York, 229p.
339
-------
Hurlburt, S.H. 1971. The non-concept of species diversity: a critique and
alternative parameters. Ecol. 54: 577-586.
Jumars, P.A. and K. Fauchald. 1976. Between community contrasts in success-
ful polychaete feeding strategies. Belle W. Baruch Libr. Mar. Sci.
6: 1-20.
Littler, M.M. and S.N. Murray. 1975. Impact of sewage on the distribution,
abundance and community structure of rocky intertidal macro-organisms.
Mar. Biol. 30: 277-291.
Morgans, J.F.C. 1956. Notes on the analysis of shallow-water soft substrate.
•k An. Ecol. 25: 367-387.
Reish, D.J. 1959. An ecological study of pollution in Los Angeles-Long Beach
horbors, California. Occ. Pap. Allan Hancock Foundation 22: 1-119.
Schmitt, W.L. 1965. Crustaceans. The University of Michigan Press, Ann
Arbor, 204p.
Smith, R.W. and C.S. Green. 1976. Biological communities near submarine
outfalls. J. Wat. Pollut. Control Fed. 48: 1894-1912.
Snee, R.D. 1974. Graphical display of two-way contingency tables.
Statistician 28: 9-12.
Amer.
Sokal, R.R. and C.D. Michener.
systematic relationships.
1958. A statistical method for evaluating
Univ. Kans. Sci. Bull. 38:1409-1438.
Stevenson, W., R.W. Smith, C.S. Green, T.S. Sarason, and D.A. Hotchkiss.
1975. Soft bottom benthos from an area of heavy waste discharge:
hierarchial classification of data. Southern California Coastal Water
Research Project, El Segundo, California TM 226, 38p.
Stirn, 0., A. Avcin, I. Kerzan, B.M. Marcotte, N. Meith-Avcin, B. Vriser,
and S. Vukovic. 1975. Selected biological methods for the assessment
of marine pollution. Pages 307-327 In E.A. Pearson and E.D. Frangipane,
eds., Progress in water technology, marine pollution and marine waste
disposal (Suppl.) Proceedings of the 2nd Inernational Congress, San
Remo. Permagon Press, New York, 487p.
Word, J.Q. 1978. The infaunal trophic index. Pages 19-39 In. W. Bascom, ed.,
Coastal Water Research Project Annual Report 1978. Southern California
Coastal Water Research Project, El Segundo, California, 253p.
340
-------
PRODUCTIVITY RESPONSES OF LAKE EOLA TO URBAN RUNOFF
by
Harvey H. Harper, III, Yousef A. Yousef, and Martin P. Wanielista
Civil Engineering and Environmental Sciences
University of Central Florida
Orlando, Florida 32816
ABSTRACT
Lake Eola is a land-locked lake located in downtown Orlando, Florida. It
is a focal point for the city and is visited by many tourists for its aes/the*;
tic appeal. No swimming, motor boating, or fishing is allowed. The lake was
drained down exposing 40% of the bottom and approximately 4 feet of mud was
removed in 1972-1973. However, input nutrients, organics and other compounds
from street stormwater runoff have not been effectively managed. Five years
later, excess fish and algae have to be removed periodically and the water
quality of Lake Eola is again questioned.
«(.
A research project is sponsored by the U.S. Environmental Protection
Agency, the Florida State Department of Environment Regulation and the City of
Orlando to determine the lake impacts of pollutants in stormwater runoff
optimal combinations of stormwater management methods.
Periodical water samples are being collected from the lake for various
stormwater events. Changes in water quality parameters with time through
each storm event are documented. Parameters analyzed include alkalinity,
hardness, solids, BOD§, TOC, nitrogen, phosphorus, and heavy metals for
particulate and dissolved fractions. Loading rates from nutrients and heavy
metals release to Lake Eola due to stormwater runoff are analyzed, and lake
impacts are evaluated.
Algal bioassay studies are performed to investigate stormwater impacts
on algal productivity. Periodical water samples are being collected from
various locations in the lake, mixed and filtered for limiting nutrient
studies using various concentrations of N, P, and Fe. Unialgal species of
Selenastrum, Choi ore!1 a and indigenous species are used and changes in chloro-
phyll "a" and biomass are measured. Initial results indicate that phosphorus
or nitrogen can be limiting at sometimes of the year. Maximum standing crop
seem to occur at N:P ratio of 15-20:1.
Also, similar bioassays are performed on a mixture of stormwater, coagu-
341
-------
lated stormwater and lake water at different ratios. Higher concentrations
of stormwater would inhibit algal productivity and smaller ratios would signi-
ficantly increase productivity. Additionally, no increase in productivity is
observed when coagulated runoff water is used. These experiments can be used
as useful tools to facilitate stormwater management decisions.
INTRODUCTION
Lake Eola is a small land-locked lake located in downtown Orlando,
Florida. The lake receives direct stormwater runoff from a watershed of
approximately 78.2 acres of commercial and 57.8 acres of residential areas
surrounding the lake. There are currently eleven active street drains which
drain stormwater directly into the lake (Figure 1). The level of the lake is
maintained between 87.0 and 88.5 -feet above sea level- by two drainage wells
which drain excess water into the underlying artesian aquifer. The natural
shoreline of the lake has been replaced with a stone'wall to prevent flooding
of the adjacent parkland, and numerous small patches of rooted emergent
macrophytes exist along this wall. Howrver, no rooted submergent plants of
any kind have been noted in the lake. Physical characteristics of Lake Eola
are listed in Table 1.
Floating masses of dead algae and fish and their accompanying odor are
a common occurrence in Lake Eola. With the exception of areas near the shore-
line, the bottom of the lake is covered with an accumulation of loose floccu-
lant partially decomposed organic matter which is easily disturbed. The loose
nature-of this material makes it difficult for rooted submergent plants to
exist and with the exception of a very small area near the shoreline, no
rooted plants of any kind were seen in Lake Eola. In areas near the center
of the lake this organic matter, subjected to long periods of anoxis and
reducing conditions, has formed into sapropen complete with the characteristic
hydrogen sulfide smell.
A restoration project of Lake Eola was undertaken in 1972. At that time
the lake was partial-ly drained and approximately 40% of the bottom was cleaned
of muck and silt and covered with relatively clean builders' sand. Existing
stormwater drains were also extended into the lake approximately at the 8-10
foot water level. The lake was then refilled with water from the drainage
wells. However, no efforts were undertaken at that time to manage or treat
the stormwater entering the lake, and now the water quality of Lake Eola is
again questioned. Buildup of flocculant sediment matter is increasing rapidly.
Large masses of algae can be seen floating along the shoreline, and fish and
duck kills have been reported periodically during spring and summer months
following heavy rain events. Also, Salimonella, Shigella and Clostridium
botulinum have been isolated from the w&ter and shoreline sediments in the
lake.
The lake is a focal point and tourist attraction in the heart of downtown
Orlando, Orange County, FLorida. Therefore, a research project was initiated
during 1978 and supported by the US EPA, Florida Department of Environmental
Regulation, City of Orlando, and the Engineering and Industrial Experiment
342
-------
343
-------
UJ
CO
*"
eC
O
1—4
0£
O
U.
ft
•eC
i '
O
UJ
UJ
o:
o
u.
t— i
to
h-4
Ul
y
£y*
•a:
g
1
^^
o
H-4
to
JC
Q.
..^
4_)
•p™
•«->
c
3
O"
S-
co
•M
CO
E
s- a>
CM tO S- LOU3
E •»-» O O 0
O CO
r>. j= o xx
CM O
«>CM • O CO
cr> o^ r^* co •
O • CM • t>.
i— o — * co co
CO
5-
^*
CO
O CO
^- 5
&H I""*
3 O
to • >•
CO CO
1 % 1 ^
ia id
E E
•r- *r~
X X
0 0
S- S-
Q. O-
Q. Q.
*^™^. ^^^s **^*
•4-^ *4— ^ *4^
EM- 1- E t-
E r—
CM CM CO I"-. O CM
CD CT* OO • »"• • tf> •
• • • CM ^t" VO r~"
CO CT> VO CM r— «5f
^^^ s—^ *^**
+J
C
CO CO
•r- O.
r— O
CO r—
S- CO
J= O >
•4-> JS CO
CL C/> O
x: co
4J ca ««- co
o. o c
CO E 'r-
O 3 J= r—
E •<-> CO
C •!- O> S-
(d x c o
s: s — i to
E ^""^
CM •!->
r-. oo <»-
• •
r— U3 00
CM OO
*•—••
i—
CO
>
CO
~J /
S
to
CO
4J 0
C -Q
CO ^
fS
a. 4->
O J=
r— • CH
CO «r-
> CO
CO 3=
a
CO
co en
E
> «c
344
-------
Station at the University of Central Florida to develop:
1. The nature and extent of pollutional loads from stormwater
runoff to Lake Eola
2. Impact of these loadings on the lake ecology
3. Management techniques to minimize stormwater effects
4. A proposed plan for the restoration of Lake Eola
EXPERIMENTAL METHODS .
Water samples were collected from stormwater drains during stormwater
events at selected time intervals to determine pollutional loads released to
the lake. Also, Lake Eola was divided into six representative locations for
field measurements and sampling for water quality analysis and bioassay exper-
imentation. Water quality analyses were performed at each station on a month-
ly basis for a period of one year beginning in July, 1978. All samples were
collected from a depth of one meter using a 2.0 liter Kemmerer water sampler,
stored in one gallon polyethylene containers which were completely filled to
eliminate gas exchange and placed on ice in the dark for return to the labor-
atory. Dissolved oxygen and temperature profiles were recorded monthly at
0.5 m intervals and at the bottom using a YSI Model 54A Dissolved Oxygen
Meter equipped with a remote sensing probe. Seechi disk depth was also deter-
mined at each sampling station.
Water samples were collected and returned to the Environmental Engineer-
ing Laboratory at the University of Central Florida for water quality analysis
and for use in bioassay experiments. The following determinations were per-
formed on each sample collected: pH, turbidity, organic carbon, inorgainc
carbon, nitrate nitrogen, ortftophosphorus and chlorophyll "a" along with an
analysis of heavy metals which included: zinc, lead, chromium, nickel,
copper, aluminum, iron, cadmium and arsenic. Total Kjeldahl nitrogen and
BOD5 were also determined on selected samples. Measurements of pH, turbidity,
and chlorophyll "a" were performed within 4 hours of collection. Other anal-
yses were conducted within the time specified by EPA in Methods for Chemical
.Analysis of Water and Wastes (1976). Chlorophyll "a" concentrations were
determined from a calibration curve using a Turner Model 111 Filter Fluoro-
meter. The calibration curve was prepared by calculateing chlorophyll "a"
concentrations in water samples from Lake Eola using the trichromatic spectro-
metric acetone extraction method as described in Standard Methods for the
Examination fo Water and Wastewater, 14th Edition and comparing these values
to relative fluorescence values.
Heavy metal analyses were performed on concentrated samples using a
Spectrometrics Incorporated Spectrospan III Plasma Emission Spectrophotometer.
Samples were concentrated by adding 2.0 ml of concentrated HNOo to 100 ml of
sample in a 250 ml Erylenmeyer flask and heating at 95°C until a volume of
approximately 10 ml was achieved. The sample was then brought up to 20.0 ml
with glass distilled water and stored in a covered polypropelene container for
measurement. All glassware used in metal determinations was acid-washed be-
fore each use with a 1:1 solution of hot hydrochloric acid followed by 5
rinses in glass distilled water.
345
-------
mg/1
7 8 9 10
4
mg/1
DISSOLVED 123456789
OXYGEN —' • ' ' ' ' ' '
9 10 II 12 l3TpMp5ft 26 27 28 29 30
o- -o D.O.
TEMP
FEB 10,1979
JULY 10,1979 ,
FIGURE 2. VARIATION IN TEMPERATURE AND DISSOLVED
OXYGEN WITH DEPTH IN LAKE EOLA IN THE WINTER
AND THE SUMMER.
346
-------
CO
I2O
11.0-
10.0-
9.0
8.0'
40.0
30.0
20.0
10.0
0
10.0;
9.5
9.0
8.5-
8.0
8.0
SO
4.6
2.0-
O
DISSOLVED OXYGEN
(TOP 0.5m)
•CHLOROPHYLL ?a"
TURBIDITY
S O
—1978-
N D
M
-1979-
T
M
FIGURE 3. AVERAGE PHYSICAL CHEMICAL PARAMETERS OF
iVATER SAMPLES COLLECTED FROM THE TOP 0.5m IN LAKE EOLA.
347
-------
FIELD ANALYSIS
Temperature profiles in Lake Eola indicate little variation between the
stations recorded on a specific date and variations in temperature between sta-
tions were generally limited to 1°C or less. The highest temperature recorded
during the test period was 30.9°C, occurring during August. Lowest lake water
temperature was measured during February with a yalue of 11.1°C. Seechi disk
measurements at stations fluctuated slightly during the test period with re-
corded values between 0.8-1.3 meters, the lowest values occurring during the
period of September to December.
Dissolved oxygen concentrations in the top 0.5 meter of the water column
fluctuated between 7.5 and 12.6 mg/1 over the test period. Concentrations of
dissolved oxygen, although usually at or above saturation near the surface,
drop periodically during the spring and summer months to 1 mg/1 or less at
depths of 4.0 meters or greater as shown from Figure 2. Although measurements
of dissolved oxygen have indicated that the deeper areas in the lake can occa-
sionally become oxygenated by wind action, it seems reasonable to assume that
areas in Lake Eola below 4-5 meters deep remain anoxic during much of the
spring and summer. Besides releasing large quantities of I-^S and C02 from
bottom sediments, these conditions can also cause release of considerable
quantities of phosphorus. This release from sediment material was measured
in Lake Eola to be as high as 250 mg-P of orthophosphorus per meter square
after approximately two months of anoxic conditions (Marshall, 1979). If this
release is conservatively considered to be limited to areas greater than 4.5
meters deep, the area under this depth would be approximately 30,000 square
meters and the expected release may be as high as 7,500 grams of orthophospho-
rus. A release of this magnitude, if mixed into the water column, would in-
crease average in situ orthophosphorus concentration by as much as 0.023
mg/1-P.
LABORATORY ANALYSES
Lake Eola water was found to be somewhat alkaline with pH values ranging
from 8.3 to 9.5 (Figure 3), presumably due to the high rate of algal produc-
tion. At the average pH value in Lake Eola of 8.86, approximately 96%
of the inorganic carbon present would exist in the bicarbonate state. The
remaining 3% would be in carbonate form with only a minute percentage of free
C02 in Lake Eola. Thus, it appears that algal production is depleting this
source almost as fast as it enters the water. After the free C02 is utilized,
those photosynthetic organisms which are capable of utilizing bicarbonate ions
also begin to do so. It is not unreasonable to assume that the composition of
algal communities in Lake Eola is determined and regulated not only be season-
al variations but also by the type of inorganic compounds present. PH values
were typically lower in fall and winter months when algal production would be
expected to decrease and higher in the spring and summer when production is at
a maximum. Turbidity increases during spring and summer (Figure 3)
are also indicative of this increased production.
348
-------
Nutrient data for Lake Eola, in many cases, does not follow typical pre-
dicted cyclic patterns, which is to be expected considering the irregularity
in both quantity and quality of the pollution source (Figure 4). Concentra-
tions of orthophosphorus fluctuated between 0.01 and 0.04 mg/1 during the test
period. Although concentrations would normally be expected to reach minimum
values during the highly productive spring and summer months, peaks in phos-
phorus were found in August and May, presumably due to stormwater additions
and/or phosphorus release from bottom sediments. Organic carbon was lowest in
the winter months as expected due to decreases in algal production and storm-
water inputs with a corresponding increase during the spring and summer. How-
ever, a very large increase in organic carbon was measured during October and
November and was presumed to be due to mixing of sediment material caused by
circulation increases typical of the fall season. Inorganic carbon appeared
to be lower during the spring and summer as algae utilized this substrate as
a food source with slight increases during the nonproductive winter months.
A two-fold increase in concentration was recorded in April 1979 presumably
due to stormwater inputs. Although the data on nitrate nitrogen is limited,
it appears that nitrate concentrations experienced decreases during spring
and summer as algae utilization increased. Concentrations of chlorophyll "a"
were extremely varied over the test period. Values measured at individual
stations ranged between 6.2 ug/1 during October to 41.4 yg/1 in July. Al-
though the data is very limited, concentrations of BODs seemed to be very low.
Average BODs at individual stations ranged from 3.0 to 4.5 mg/1.
Total metal concentrations measured in composite water samples collected
in Lake Eola are listed in Table 2. No apparent trend was observed from the
data presented and the fluctuations seem to remain within a narrow range and
oscillate around an average value.
BIOASSAY EXPERIMENTATION
Bioassay experiments were conducted to evaluate algal responses
to nutrient changes in Lake Eola water and to study the impact of untreated
and coagulated stormwater runoff on algal cell production. In this research,
four different types of sample preparations were used: raw unfiltered lake
water, water filtered through a 0.45 micron Millipore filter (F), water which
was filtered then autoclaved (F/A), and water which was autoclaved then fil-
tered (A/F). In addition to indigenous algal species from Lake Eola, two
different laboratory cultured test species were also utilized, Chi ore! la py-
renodoaa and Selenastrum capricornutum. Both of these species are solitary,
non-motile green algae which posses a wide tolerance towards environmental
conditions and occur in waters of diversified composition (EPA 1978). A sum-
mary of experimental conditions for bioassay studies conducted during this re-
search is listed in Table 3.
The validity of results obtained from a bioassay experiment hinges upon
the ability of the experiment to determine the maximum growth response and
growth limiting nutrients of a test water. Since these responses are an inte-
gration of the combined effects of ion solubility and ion availability to the
test organism, any factor which alters these conditions, such as sample treat-
349
-------
O)
INORGANIC CARBON
FIGURE 4. AVERAGE PHYSICAL CHEMICAL PARAMETERS OF
WATER SAMPLES COLLECTED FROM THE TOP 0.5m IN LAKE EOLA
350
-------
CM
LU
CD
I—
t/5 -=C
C£ O
LU LU
OO •
O LO
o •
CD
•z.
l—i D-
O
00 M
CD LU
1— O
•zz cc
LU U_
o LU
01—
o
_J LU
< -1
I— _J
LU O
•Si O
r~
"•x,^
O)
£
O
.(_> *
to
S-
CD
O
c
o
o
ea
CD
'
CD
to
to
3
r-
en
eC
-o
O
c
S-
• S
*•— *
ja
O-
0)
i —
1
1
1
1
CO
CO
o
0
CO
o
o
o
en
CM
o
CD
CM
o
o
•
o
cr>
CO
<3
0
LO
CM
.0
0
CO
o
CD
CO
CO
0
co
1 —
o
en
r-.
o
CM
i
i
1
o
10
0
o
en
o
o
o
LO'
0
o
CM
0
o
•
0
•vl-
LO
o
o
Cn
CM
CD
o
£
o
CD
r—
*-^
0
^
CO
o
en
r*^
^*^
co
r—
co
i
i
1
co
LO
0
o
£
o
CD'
O
O
<&
CO
CD
^
CO
CO
o
en
^^^
{Q
O
^
1
1
'
^
O
O
LO
o
o
o
0
o
•
o
o
o
o
•
o
LO
^~
.0 •
o
o
CM
CD
CD
en
10
0
o
,_
CM
r—
CD
LO
CO
CM
CD
en
^^
^^^
0
CO
*
co
•
r—
CM
<£>
O
O
O
0
CD
O
O
CD
O
O
O
O
•
O
0
CM
O
O
O
CO
O
o
10
co
o
0
LO
^*
o
o
<£>
CO
CM
CD
en
f-^
^*^
o
CM
<£>
•
LO
CM
r-.
LO
o
CD
LO
o
o
0
^
o
o
•
o
LO
LO.
o
o
CM
CO
o
o
CO
CO
0
0
1--.
CD
r-
o
«D
CM
CM
o
en
r*^
**^
o
1 —
f^-
CM
•
co
CM
00
LO
O
O
CO
0 ,
o
0
CO '
LO
o
•
o
r—
o
o
•
o
• -
LO
CO
o .
o
en
CM
o
, CD
en
10
CD
o
o
CO
"—
o
L
CO
CM
CD
cu
Dl
to
s-
CD
•$
351
-------
CO
LU
_1
ca
co
UJ
co
co
co
eC
O
1—4
CO
D;
CD
co
o
o
o
UJ
UJ
O-
X
UJ
o
£
CO
-a
r—
1— T3 --v.
a -a CT
Luce E
CU
u_
•o
cu
-a
•a a.
to -^
E E
as
5-
3 2:
q-
q-
0
E &«
o;
to
V) E
CO fO
S5
CO
E
S- O)
CO CU E
fO fO fO
_,3=£
d)
fB
a
O O CD O O O r
LO LO U
OOOO O O C
0 0 C
O LO LO U
r-^ O O C
1 CD CD O 1 1
. — LO LO L
0 0 C
o
0 0 C
LO LO LO L
IOOO 1 1
r— i — i — r
CU
O O 4->
LO CD O CO
1^ r— i— CU
CD 1 1 1 S CD O C
o o o
•a
E E
3 3
(O CO (O CO S- S-
r— i — r— r— (/) to
CU (LI d) CU (C ro
S- S- S- S- E E
OOOO CU O)
S IE JE IE "0] "QJ
o o o o co co
_
(/) CO (/) C/) C/) t/1
OOOO O O
a. a. a. a. o. o.
E E E E E E
OOOO O O
cj> o c_> o cj o c
oo ex) co co en en
CD «D CO -^- «* i —
r— i— O r— r— CM
00 O r— CM r— CM
-OOOO O O
n
D O O O O O O
3
r»
3 LO LO
1 • • 0 0 00
O O CD
D
•
0
o o
O r— r—
III O O O
"O ' — -
CU "O
- — - +* CU
-0 (0 H->
O) r— (O
O H-> O 3 O CD r—
LO rO LO CD LO LO 3
D o CD i cu i ca i i en
O S- O O CD CD (O
+J 0 0
E 0
• 3 CU
i— U_
(0
3 tO I/) tO tO
i- 33 33
•i-* o o o o
to E E E E
ca cu cu cu cu
E CD cr> en CD
CU *r- *r- *r- *r-
r— T3 T3 T3 T3
CU E E E E
CO 1— 1 1 — 1 1— 1 1 — I
a: u. «=c
IT -=c u,
d) CU CU CU CU CU CU
y> to to to to to to
o o o o o o o
a. o. o. o. o. a. a.
= E E E E E E
O O O . O O O O
J> O O O O O O
en en en
CO O t*-.
r— CO r—
co ^t- r~-
352
-------
ment, may bias bioassay results. Chemical analyses conducted during this
research indicate that autoclaving, filtration followed by autoclaving, and
autoclaving followed by filtration all result in significant decreases in
concentrations of organic carbon, inorganic carbon, and orthophosphorus. How-
ever, nitrate nitrogen concentrations show substantial increases due to these
treatments (Table 4). Similar results have been reported by Filip and Middle-
brooks (1973) in tests on mesotrophic pond water. They suggested that oxida-
tion attributable to the physical disturbance of the sample as it passes
through the filter probably accounts for higher nitrate levels in treated
samples. It seems reasonable that increases in temperature and pressure
during autoclaving may also be responsible for chemical changes of various
parameters in water samples. Oxidation of organic carbon, nitrogen and phos-
phorus may occur altering the composition and quantity of these nutrients
while at the same time, slightly increasing the pH. Autoclaving of the sample
either before or after filtration, besides substantially reducing concentra-
tions of organic carbon, inorganic carbon, and orhtophosphorus can also cause
reductions in concentrations of calcium, phosphorus, iron, and lead by preci-
pitation during autoclaving (Filip and Middlebrooks, 1975). These precipi-
tates may be resistant to resolubilization under bioassay conditions (Envi-
ronmental Protection Agency, 1978). Oxidation of organic wastes may also
occur rendering these compounds less toxic. It appears that autoclaving of
a sample tends to produce conditions which may enhance algal production by
increasing the availability of some nutrients and at the same time decreasing
concentrations of organic compounds and certain metals. The maximum yields
obtained in this research using test waters which were autoclaved then fil-
tered were much larger than any other type of treatment.
From the results obtained during this research, it appears that auto-
claving of a sample may not be a suitable treatment technique for use in
studies where the effects of complex organic compounds are to be determined.
In these cases, filtration should be considered. If, on the other hand, a
relatively clean' oligotrophic water is to be assayed and it is desired to de-
termine the amount of algal biomass that can be grown from all nutrients in
the water, including thos contained in filterable organisms and other parti-
culate matter, then autoclaving of the sample seems to be the proper technique.
As a general rule, samples which are autoclaved will produce a much larger
standing crop than filtered samples.
ALGAL RESPONSES TO NUTRIENT ADDITION
A summary of nutrient addition bioassay experiments using Lake Eola
water is listed in Table 6. As seen in this data, both nitrogen and phospho-
rus were able to stimulate algal production in Lake Eola on certain dates.
Although the data is limited, it appears that stimulation of a particular
water by nitrogen or phosphorus may be related to the background phosphorus
concentration at the beginning of the bioassay. Whenever the orthophosphorus.
concentration before nutrient additions was approximately 0.02 mg/1 or less,
algal production was stimulated by a phosphorus addition. Above this value,
nitrogen was shown to stimulate growth. Typical response curves of Selena-
strum, to various nutrients added to Lake Eola water are shown in Figure 5.
353
-------
UJ
CO
CO Z3
O-
!±! c?
UJ co
s: co
a: o
CD
O
s_
O)
Q_
O-
*
o
^y
i
CO
o
o
E E
IO O
O tO
E ^-^
l — i
O
•r- E
E O
05 ~£
S- to
O 0
T"
Q.
tfl
OJ
o "cL
E
=8= to
CO
4-^
0) E
E *^
tO tO
co cu
h-
^ — ^ . — * CO
<£> CO 1
"""Jc? "^i . « ""^co
COLO i— « CTI"* COCO
LO i ojt~~ cocn i — i
I* | OJ | 1 |
— "
en LO ^— -
i— 1 ID IQ
LO-^ — " o+ o+ o^"
.LO . •> • •> .+
r~»i— i C~.LO COLO LO "
** r~~* — ' <^t*OJ OJ"^~
+ o ++ + + + +
> — - S — -
*-~^*
' — - OJ * — -
LO OJ OJ
LO- — - O+ O 1 CO 1
.CO . " . •> . •>
i — i ^Of""* COO LO^
" r— CO COLO COLO
I CD I 1 | I i 1
^^^ ^^* •<• — ' ^^
x-s o vo . — -
LOLO O* — ' O * LO^-^
.1 .1 . « .1
OJ " CO " OC3 IO "
i — O LO'-O •sJ'CD COCO
1 OJ I CO | r— 1 | CD
1 1 1 1
x-^ X — ^ x~ ^ *~~ ^
LO' — ' 0^5 COOJ O1^"
• 1 .+ •+ .+
i — " i — " CO" i — "
1 1 + 1
^ ^ V ^ V ' N ^
OJ OJ CO -^f-
T3
cu res
S- i— «=C U_
».
+-> O LL. <
t— +->
III ej;
.
* — •
E
. *r—
E
0
ID
, —
IO
>
q-
o
CD
E
tO
C£
*
354
-------
LO
oo oo
^. ^.
oo o
•=C 1-1
Q
Q
CD LU
O I-H
CtL
Q
LU O
CQ i— i
O 3:
00 C£.
oo
LU
>- o
eC LU
OO
00 LU
- i—i
oo
CJ
o
o
•r-
+->
to
rv*
r—
V) -»-> 3
0
n
•5
^
i—
t—
a
1 E •*->
. E oo
" i- E
- CU i — i
0 X
LU
X
to
M- ^.
0
O
•p
4
f
Q
' f \
V— '
E
O
CJ
.jj
oo
E
I-H
1
CU
CJ
E
i
•r—
X
rd
_>
r*
2
4
= -a
frt p-"-
(O P™~
= CU
r ^ i
i — to i —
l — tO ^
QJ fO O J
cj s: E
_
_trj^-~.
_l CD
>- p.
"T~ •* ^
CJ
E
J *r-
to E
CU 55
H
t
- 01
S-
cr>
3 co
CU n3
£
•p
r
4
rEZ
5 ^ — ^
-> i — i —
-Q O ^-.
Or rn
^> w j
4J E
T
p-
3 E — '
- O
CU CJ
•r^
>- E
cu E
1—
£
£
f
— eu
ix. E
" -U)
5 to
OO CU
r-
ci
(
E
O
CU O. O
4J E CU
f
O to i —
Q OO i —
0
o
p—
,
en
LO
CM
CO
CO
o
•=3-
cn
p.—
(O
. p_
r-«
S
o
.E
CJ
LO
0
Lu
CO
CD
^^
CO
oo
,.(.
O^
^.D
en
o
oo
oo
CM
(O
, —
.£
o
c—
CJ
cn
oo
oo
Lu
00
CO
CD
•^^
p_
p—
0
CM
oo
^J-
•~
00
CM
0
r^
CO
to
p~"
£
o
3=
CJ
r^
•*
o
ecC
•^^_
LU
CO
CO
•^^
o
p—
CO
0
r~-
i-^.
oo
i —
,_
CO
oo
oo
E
3
S-
4J
CO
(d
E
CU
'cu
00
0
en
o
to
to
c.
CU
"aj
00
o
1 — '
CM
•=c
Lu
en
co
*v^
oo
LO
o
CO
CM
CO
cn
o
CM
oo
CM
E
3
S-
4J
to
to
E
CU
'cu
oo
0
o
p—
Lu
•^^
^c
cn
^~
CM
•^^
CM
<* LO
, ,
p— r^
p— en
CM P—
cn i —
LO CO
0 0
LO cn
h~. r-~
r~~ p-^
CO tO
0 O
E E
CU CU
cn CD
•i— 'r-
•a -a
E E
i— i i— i
OO p—
co cn
0 0
cu cu
E E
0 0
E E
cn en
CD 1^
OO i—
^^t "-^
«^~ ^^
355
-------
A close examination of the data indicates that the N:P ratio may, under
certain conditions, be more influential in regulating algal production in
Lake Eola than the actual concentration of nitrogen or phosphorus alone. The
fact that N:P ratios can play an important part in determining the potential
bioinass in a system is clearly demonstrated in Figure 6. In these experi-
ments an optimum weight ratio of total soluble inorganic nitrogen (TSIN) to
orthophosphorus was found to be between 7.4 and 21.4. Chiaudani and Vighi
(1974) report optimum weight ratios between 4.5 and 9 in studies on Italian
lakes. It seems reasonable to assume that when a nutrient is present in a
minimal quantity that a point will be reached in terms of the concentration
of a particular nutrient at which the availability of the nutrient will be-
come so low that further growth will be limited by ion availability rather
than by its relationship to other ions. It appears then that a threshold
level of phosphorus exists below which the growth of an organism is regulated
solely by the concentration of phosphorus present, assuming all other essen-
tial nutrients are also present. Above this threshold value, phosphorus is
in relative excess and the organism is no longer limited by ion availability
alone. As concentrations of ions increase, organisms are virtually surround-
ed by an abundance of nutrients. Since some molecules have greater affinities
for binding sites on the cell membrane than others, these binding sites may
become blocked by certain compounds, limiting further nutrient uptake. In an
environment with an excess of nutrients, cell growth will be optimum at a
nutrient ratio where nutrients needed for growth can be readily taken up by
the cell in approximate quantities necessary for growth. As seen in Table 6,
an orthophosphorus concentration of 0.060 mg/1 was not sufficient to produce
excess ion availability. However, when the concentration is increased to
0.110 mg/1 as was the case in the experiment using lake water collected on
3/18/79, optimum growth occurred at a weight ratio of 6.4:1, which is within
the range previously reported. It appears then that the threshold concentra-
tion within Lake Eola above which sufficient phosphorus ions are available
so that growth may be limited by the relative abundance of these ions rather
than their actual concentrations is between 0.060 and 0.110 mg/1 P04-P. In
other words, when the concentration of orthophosphorus in Lake Eola is approx-
imately 0.10 mg/1 or less, algal production is regulated by the concentration
of orthophosphorus alone. Above this concentration, it is assumed that an
excess of phosphorus is available and algal growth is regulated by the rela-
tive availability of nitrogen or phosphorus. However, since concentrations
of orthophosphorus recorded in Lake Eola during this research were 0.04 mg/1
or less, it can be concluded that, except during periods of heavy P loadings
by stormwater inputs or by long periods of anoxia, algal production in Lake
Eola is regulated by the concentrations of added phosphorus alone.
STORMWATER IMPACT
Average water quality characteristics for runoff water from Lake Eola
drainage basin for eight storm events are presented in Table 7. Water quality
parameters seem to fluctuate over a range of more than ten fold in some para-
meters such as TOC and N03~. Also, more than half of the nutrients released
to the lake in stormwater runoff appear to be in solution. Calculations of
the mass loadings from these storm events are presented in Table 8. Mass
356
-------
CQ
oo o;
oo LU
CO
<
z _l
o o
Q
Q
I—CD
z: z
LU 1—I
2*
££
U- LU
- rv
E C
E 0-r
OJ -r- C
•r- -P I
S- -r- T
4-> -0 C
3 -O S.
z: eC a
-a
a> E
4-> 3
o E
-a x
O> ra
i- s:
CL.
03
3 OJ
E E
X i.
ca a>
s: Q.
X
LU
in
o
•i—
4-J
•r—
O
ra
ra
^"*
o
O>
Q.
(O
oo
E
to
r—
ra
S?
o
s'-o"
E " —
X 'I-
ra >~
s: —
7)
- EE ••""^**
- 3 -a
J E r—
3 •(- O)
3 X •«-
3 (O *>•
- 5C **— •*
*^— s
•Q 1— •
r— "s^.»
-»— '
-O r—
aj o;>
•r~ E
>- — ••
z^
cooi
0 E^
CL. 1 —
1 ^.
«^- D^
0 E
CL. '
^
c
0)
E
ro
a)
*~
-a
0) -0
i- -a
i — uT rf" cc «c uT
•r- 1—
U_ Q
[ | |
OO O"i CT> O^ O^
r*^ r**** r*"11* r*** r*-"
****** "'Xi** *x"'«* ****** ******.
O *«d- r- i— CO
i— r—
-------
CONTROL
- O.OSmg/IA—^X.
O.lmg/IA—-&\
24 6 8 10 12 14 16 18 20
INCUBATION PERIOD (DAYS) l
FIGURE 5. RESPONSES OF SELENASTRUM TO VARIOUS NUTRIENTS
ADDED TO LAKE EOLA WATER SAMPLES COLLECTED 1-14-79
(LAKE WATER FILTERED THEN AUTOCLAVED).
358
-------
RATIOS FOR MA/.GROWTH
YIELD
S
-: 5;l 10 :| 15=1 20:1 25 :| 30=1 35H
NITFIOGEN TO PHOSPHORUS RATIO
FIGURE 6 .CHANGES IN MAXIMUM GROWTH YIELD IN A
SELENASTRUM CULTURE INCUBATED IN LAKE EOLA
WATER DUE TO CHANGES IN NITROGEN TO PHOS-
PHORUS RATIOS. (LAKE WATER COLLECTED 3-18-79).
i
01
o
I
i .:
1 ,
359
-------
ui
_j
oo
D:
UJ
o
CO
co
CS
U-cC
o z:
\—\
CO «=C
o o;
i-< Q
CO <
I— I _J
C£. O
LU UJ
CJ> UJ
_1 Q
CC UJ
=3 h-
CXO
UJ
Di _l
UJ _J
UJ
CD
UJ
in
1 S- Q-
O O 1
.c .G i—
+J Et-^
S- t/l O>
0 0 £
Q.
CO 1
COr—
o --^
i- 05
•!-> E
•r-»
z:
O r—
o --»
«-g
U> —
o en
03 E
co en
JD
S- ID
£5
•
.ii
r—
, <
£
• • (
O T3 *•
0) G C
CL O ^
GO O !
in
r—
"
to
5
o
•z.
^
1—
•
in
in
0
"fO
•(->
O
•*..
3
•=d-
«*
O
CO
•> co
r-
LO
CO
•
<^J-
to
to
CM
«*
•
co
CO
CD
0
CM
•*
CT>
•
«D
CM
CO
•
•*
co
i
i
i
cr>
vo
CO
o
00
r~-
r-
un
CM
o
LO
in
o
•=1-
,•
CM
•
O
oo
CM
OO
CO
o
CM
CM
r-~.
I-,
en
•
00
<£>
•
CT>
un
en
o
C\J
oo
r~-
r-
CD
OO
t-«.
«*
r^
0
o
CM
«d-
0
LO
IO
0
CM
•
O
co
CM
r~^
LO
en
CM
r—
CO
«*
CO
10
vo
10
l-~.
o
CM
CO
oo
r*.
LO
CM
t~-.
LO
1—
0
CM
ID
0
:
r—
*
CO
r—
l£>
0
CO
O
en
oo
oo
•=j-
CM
LO
CM
r—
1 —
vo
r-.
•*
0
o
CO
r--
co
CD
CO
CO
O
en
CM
0
LO
vo
•
0
•
o
en
•
CO
r—
IQ
•=d-
•*
r--.
•sf
I—
LO
CO
CO
co
r~»
CO
•*
0
co
r-.
0
r—
oo
r^
<£>
o
o
CM
CO
«d-
O
•
r^—
P^.
OO
g
co
r— •
«=d-
CO
r— •
«*
<=!-
en
to
co
en
r—
LO
l~-
co
o
00
r~~
LO
co
CO
CM
o
en
CO
0
o
CO
•
co
oo
«3-
CM
•^1-
oo
CM
•^-
LO
LO
r-«
CM
LO
to
LO
IS3
IO
LO
CM
•*
r^
0
0
10/12/78
CO
LO
O
O
en
oo
en
CM
co
10
0
CM
CO
CM
co
oo
f~-
CM
CO
CM
LO
en
r--
co
CO
co
CO
<3-
o
0
oo
r--
r-.
o
360
-------
TABLE 8
MASS LOADINGS FROM LAKE EOLA BASIN
Date
7/11/78
7/14/78
7/25/78
8/03/78
8/10/78
9/15/78
10/12/78
11/07/78
Average
Calculated
Loadings*
(kg/yr)
Mass Loading, kg/ha-cm
SS
18.5
4.4
8.9
9.3
4.5
5.9
13.0
1.8
8.3
58,432
BOD5
2.6
0.6
0.3
0.9
0.3
2.0
1.0
1.1
1.1
7,744
TOC
4.1
6.0
1.0
3.6
1.6
20,0
18.0
13.4
8.5
59,840
Soluble
Ortho-Phosphorus
0.06
0.02
0.02
0.02
0.03
0.02
0.02
0.03 .
0.03
211
*Assumes a drainage basin area of 55 ha and average rainfall of
128 cm/yr.
361
-------
loading average 8.3, 1.1, 8.5, and 0.03 kg/ha/cm of rainfall for suspended
solids, BOD5, TOC, and orthophosphorus, respectively. Assuming the drainage
basin area is 55 hectares and the average rainfall is 128 cm, the total solu-
ble orthophosphorus released to the lake is 211 kilograms/yr. Similarly,
loadings from suspended solids, BOD5 and TOC will equal to 58, 432, 7,744,
and 59,840 kilograms per year, respectively. If all the suspended solids are
transported to the area of the lake deeper than 6.5 meters, a buildup of
solids approximately 0.15 meter deep will accumulate per year in the deep
areas of the lake. This material, in addition to rapidly decreasing available
water depth in the lake, also creates a significant oxygen demand during its
decomposition and it may be responsible for anoxic conditions during spring
and summer months in the deep areas. Of the dates on which stormwater was
collected the lowest concentrations of virtually every stormwater parameter
seemed to occur during the months of July and August. These months are
typically characterized by frequent and intense storm events. Since the dry
period between storms is very short, the accumulation of transportable mater-
ial in contributing areas of the watershed is minimized, and nutrient input
into the lake per storm is reduced. However, since storm events occur almost
daily during this period, the total input of nutrients over this rainy season
will be very large although it will be stretched out over a period of several
months. The frequent and intense nature of these storms serves to reduce
concentrations of toxic heavy metals and organics while at the same time
supplying the lake with a constant source of phosphorus, nitrogen, and carbon.
This continuous supply of nutrients with reduced concentrations of toxic
elements combined with warmer water temperatures during these summer months
provides enriched conditions for algal growth, and one would expect the most
rapid rate of algal growth to occur during this time. However, because dilu-
tion of algal populations by stormwater and the constant removal of the upper
layers on the lake via the drainage well are constantly lowering these popu-
lations, the highly enriched growth which is occurring during this period is
npt reflected by in situ chlorophyll "a" concentrations.
In contrast to the enhanced algal growth conditions experienced during
the summer rainy months due to stormwater events, runoff entering the lake
after prolonged periods of drought may produce severe toxic effects on aqua-
tic life in Lake Eola. Contaminants have been allowed to accumulate within
the watershed over this period, and when a storm event occurs, the mass
loading to the lake is many times larger than that experienced during a rainy
period. This large influx of toxic and oxygen demanding wastes can be lethal
to many forms of aquatic life. Evidence of such a phenomenon was recorded in
March, 1979 when a rain event occurred after a 6 weeks dry spell. Concentra-
tions of organic carbon as high as 400 mg/1 were measured in stormwater run-
off entering the lake during this event. Two days after this event, dissolved
oxygen concentrations had been reduced from saturation near the surface to
4 mg/1 at a depth of 1 meter and to virtually zero below 2 meters. Numerous
large-mouth bass averaging 2-3 pounds were also found floating in the water,
and large masses of dead filamantous algae had accumulated in thick mats over
much of the lake's surface. The color of the water itself was changed from
its characteristic blue-green tint to a grey-green appearance. Seechi disk
depth was reduced to less than 0.5 meter. After approximately 5 days, condi-
tions began to improve, and after 10 days physical conditions in the lake, as
determined by dissolved oxygen profiles and Seechi disk measurements, had
362
-------
returned to near normal values for this time of year. However, it should not
be inferred that the lake had returned to the. same ecological condition as
before the storm event. The damage caused to the lake system by increases in
sediment buildup and loss of animal species.are difficult to document.
A summary of .stormwater addition bioas.say experiments using Lake Epla
water is listed in Table 9. As seen in this data, maximum yields, were pro-
duced in virtually every case with a mixture.,of 25% stormwater .runoff, result-
ing in increases of 87-731% over control flasks.. If Figure 7 is considered.
6 typical algal response to stormwater additions it can be seen that the
addition of stormwater in any concentration, even 100% .stormwater runoff,
resulted in an increase in algal production, over the control. However, 't when
stormwater was added in concentrations of 25% or greater, the growth curve
was typified by an initial die-off of algal populations, the magnitude of
which being a function of runoff concentration. This di.e-off is due largely
to the presence of toxic elements in the stormwater which is discussed, .
along with its implications, in a, later section. If a typical h inch storm ;
fell on the Lake Eola watershed the corresponding runoff, if completely m :
mixed throughout the water column, would represent an addition of only 2% to
the total lake volume. At this concentration of stormwater runoff the
effect on algal production would be greatly minimized. However, the time
necessary for the .stormwater-to mix throughout the entire lake, depending on
weather conditions, is probably on the order of several days. It seems
reasonable to assume then that stormwater, concentrations of 25% or greater
may exist near stormwater outfalls for at least 1 day and that the, die-off
predicted at this concentration would occur in these areas.; ,. , .
EFFECT OF COAGULATED RUNOFF ON ALGAL GROWTH ; . . -.
One of the techniques under consideration for the restoration of Lake
Eola includes chemical coagulation of a portion of the runoff to remove
phorus and heavy metals. Alum, A1^(S04)-ISH^O, was selected as a,coagulant
for test purposes. An alum dosage ;of 24Q m.g/1 at a, final pH of 5.5 was de-
termined through jar tests to provide optimum phosphorus removal. .Coagulation
under these conditions resulted in a reduction of 96% of orthophpsphorus, and
87% of nitrate nitrogen (fable 9). With the exception of aluminum and arsenic
coagulation also reduced concentration of every heavy metal.tested. As seen
in Table 9 and Figure 7, coagulation of the stormwater resulted in substan-
tially lower growth responses, as compared with uricpaguTated test flasks, at
every concentration of stormwater tested. The maximum percent increase in
algal production over the control flasks was reduced frpm 731% using
untreated stormwater to 8% in a coagulated sample. ,
An interesting result of these coagulation experiments is that the un-
treated sample required 11 days to reach its maximum yield while the coagula-
ted sample obtained a maximum in only 6 days. Greene (1976) suggests that a
growth lag in algal response is often experienced in waters containing toxid
compounds. Certain algal forms are able 'to produce extracellular substances
which can form chemical complexes with the growth inhibiting substance. The
absence of this characteristic growth lag in the coagulated samples suggests
that sufficient toxic elements had been removed by this process so that algal
363
-------
CO
I—
2:
UJ
S
UJ
CL.
X
UJ O£
UJ
Z I—
0«=C
I-H 3
1— 1 41
en Q _i
Q CD
UJ eC Ul
_J
CQ o: uj
•^g
jjg CD
O I-H
1— CO
co ra
u_
o
J>«
oi
CO
CU
+•> to r—
E re s- o
Q) CU CD S-
O S- > •)->
S- O O E
CU E O
Q. HH O
(ft
>>-E E
re o 3 -o
a re E i—
CU -r- CU
t»- Di X-r-
o re >-
0 S
E
to
4-> -t-
tfl C
CD re
^— en
^.
O
s-
• E S
re o •!- 3-0
S E 0 E •—
S O 3 -r- -
•i-J s- 2:
CO Q-
f—
re-a
4J i—
E E OJ — -
E E ^" ^***
•r- -I- CD
X S_i— E
s o.'aj
X 0
UJ
z:
i
CO
o
z:
D-
1
(/) ^-
O O
•i- a_
•1-5
to
'si
CD +J
t^t f^
O
re re
-E O)
C_5 S-
h-
O)
"o. -a
E a)
re -i-3
co a) o
•)-> O)
re i—
o
o
^^ ^^
ID f—
CT> CO
LO
co o
re re
i— i—
r~~ r—
CU O)
s- s-
o o
^^ r~"
(T* C~
CJ C_>
^^ ^^
LT5 LO
OJ OJ
r*^ co
OJ CO
• •
co 10
i i
i i
0 0
co un
• •
0 0
B^
^*"*«. L'i.
[_i_
00 CO
r**1* p*"*
•"^x. "*^»^
^o oo
r— 0
O i—
r— r—
S^ ^^
i— CO
CO
1 —
r— 10
r"-*
in (/)
3 3
0 0
E in E to
CD CD CD CD
OVi- COT-
•t- O -I- O
T3 O) T3 O)
E a. E a.
>-< co i— i co
^^ ^^
LO LO
OJ
OJ -=d-
f*»^ ^D
* •
CO i—
O CD
en en
vo a
r— CO
err i—
OJ O
• ***^
O 0
•o
CD
4J
cu E re
E 3 r—
O i — 3
•ZL •=£. en
re
0
o
en CD
r^» r^«
**^^ ^^^
co co
OJ OJ
^T ^T
364
-------
o>
E
X
52
UJ
on
o
LU
O
0.5
0.2
CONTROL
5%
10%
25%
50%
UNTREATED RUNOFF
COAGULATED RUNOFF (i«m,n
ALUM)
i
i
i
l
0 2 4 6 8 10 12 14 16 18
INCUBATION PERIOD (DAYS)
FIGURE 7 RESPONSES OF INDIGENOUS AGAL SPECIES
IN LAKE EOLA TO VARIOUS CONCENTRATIONS OF STORM-
WATER RUNOFF AND COAGULATED RUNOFF (WATER
SAMPLES COLLECTED 4-30-79, STORMWATER COLLECT-
ED 4-28-79).
365
-------
forms were no longer inhibited. Thus, it seems that coagulation of storm-
water not only removes nutrients and limits algal production but it also pro-
duces a product which is less toxic to aquatic organisms.
TOXIC EFFECTS FROM STORMWATER RUNOFF
Although the algal production measured in these bioassays would certainly
correspond to eutrophic conditions, it appears that production in Lake Eola
may actually be partially inhibited by toxic elements in the stormwater runoff.
The growth response of Selenastrum in a synthetic algal medium was measured
(Harper, 1979). Since this medium is by design a phosphorus limited medium
(EPA, 1978), then algal growth should continue until all available phosphorus
has been utilized by organisms. A total of 120 mg of dry cell weight per
liter of Selenastrum were produced from the 0.186 mg/1 of phosphorus contained
on the synthetic algal medium. From this relationship, it can be calculated
that 0.01 mg/1 of phosphorus will produce, assuming all other essential
nutrients are present in sufficient quantities needed for growth, 6.45 mg/1
of Selenastrum cell mass. Accordingly, the 0.017 mg/1 of phosphorus present
in the lake water sample collected on 8/10/78 should have produced 10.97 mg/1
of dry cell mass assuming no toxic effects existed and all other nutrients
necessary for growth were present. As seen from Table 6, none of the bioassay
experiments with the exception of the one in which EDTA was aldded, were able
to produce the maximum predicted yield.
The fact that the addition of EDTA was responsible for increasing the
experimental yield strongly suggests the presence of heavy metal stress in
Lake Eola. The additions of EDTA to a lake water sample collected on 2/21/79
resulted in a two-fold increase in cell yield (Table 6). An inhibition of
47% was calculated for this experiment between control flasks incubated with
and without EDTA. EDTA may act to stabilize ferrous iron, increasing the
availability of this ion for aquatic growth while at the same time suppressing
heavy metal toxicity by chelation (Miller, et. al., 1976). However since the
addition of iron to this experiment in control flasks without EDTA did not
stimulate algal production, it was concluded that the increase in production
caused by EDTA was due largely to suppression of heavy metal toxicity. Toxic
levels of selected heavy metals to Selenastrum along with average Lake Eola
water and stormwater levels are listed in Table 10. It can be seen that nor-
mal background concentrations of copper and zinc in Lake Eola itself are
sufficient to cause continuous incipient inhibition of algal production in
certain species. Average heavy metal concentrations in stormwater collected
near Lake Eola indicate an algicidal level of copper with a zinc concentration
sufficient to cause complete inhibition. Although no information is listed
for toxic levels of lead, the average concentration of 1500 yg/1 of lead
found in stormwater runoff at Lake Eola can be assumed to exhibit almost
certain algicidal effects. Experiments reported by Shiroyama (1976) and
Greene (1975) on the toxicity of zinc to Anabaena and Selenastrum indicate
that a zinc concentration corresponding to the average Lake Eola concentration
of 0.048 mg/1 was sufficient to reduce the yield of Anabaena and Selenastrum
to only 20-40% of the maximum yield. Similiar concentrations of lead and
arsenic are also present in Lake Eola water and may exhibit similar inhibi-
tory effects. Synergistic effects due to the simultaneous presence of various
heavy metals may increase this toxic effect ajid further limit algal growth.
366
-------
UJ
CIS
o;
UJ
^
1—
I — r
3
CIS
;^
o
ej;
2
—4
H^J
^
oo
^^
UJ
1
LU
OO
0
i— 0
f—
UJ
— 1 00
CO _l
^i c^
L_ [_
LU
g^
>~
^>
III
LU
DC
U-
o
x— -x
,1 —
oo
o
1 — 1
I
r^
^v
H-
UJ
0
p^
0
o
f>^
LU
l__
^£
3
s:
D;
o
I—
oo
o
UJ
1
1 — '
o
t — i
X
o
f—
i
i.
cu
(G
^
E
O
-p
oo
to
o
UJ
cu
(O
—1
r-«
res
•a
•r—
O
cn
r—
t~
0
a> •<-
QJ *r~
i— JQ
Q.T-
0 E
O i — i
E
•»-> O
E •!-
0) -P
.|_* >r_.
Q.^3
•p~ "r—
O -E
E E
i— i i— i
•P
CU
CU
^~
UJ
co co LO o vo ^f- o
r— CM (Ni oo •=!- cr> co
LO «*• LO
,-T
m LO r— en cr> oo co
Jj;. CO CD CM LO
0 O 0 1 II 1
O O LO • 1 1 1 1
co t-~- vo i i i i
o o o i ii i
CTl CM CO I.I 1 1
r— 1 .1 1 1
O O O 1 1 1 1
LO CO LO 1 II 1
3 E T3 J2 S- 10 -i-
o MO
E •*
•I- f-.
M CT>
n
S- JE
S-
(i) res
JQ O)
res to
ct. 01
o;
U-' S-
•a -P
E res
res 3
•P =
•p n
GJ E
+3 -P
S- 3
res E
CO 5-
• O
_i a
UJ
o
rv*
rs
0
oo
-367
-------
LAKE MANAGEMENT PLAN
A plan for restoration of Lake Eola had been developed and supported by
the US EPA, Florida DER, City of Orlando, and College of Engineering, UCF.
The plan was based on findings from the previous studies and the essential
elements are summarized as follows:
1. Reduction of nutrient input, especially phosphorus, by diversion
of stormwater runoff, storage and/or biological uptake.
2. Inactivation of phosphorus release from bottom sediments in areas
of the lake deeper than 4.0 meters, due to anoxic conditions.
Studies are currently underway to evaluate the possibility of
using water treatment sludges for inactivation.
3. Coagulation of stormwater before release to a limited area of the
lake on a pilot scale for experimentation.
This plan is currently considered for implementation and a continuous
monitoring system was set to measure the nature and extent of lake productiv-
ity changes. Detailed analysis and discussion of the proposed management plan
is beyond the scope of this presentation.
SUMMARY AND CONCLUSIONS
From all indications, both visually and analytically, Lake Eola appears
to be a lake in severe ecological distress. Persistent algal blooms exist
virtually year round. Populations of the macroscopic algae Chara and the
filamentous green algae, Spirogyra, became so dense during the summer months
along the shoreline that in many instances it became very difficult to launch
the small Jon Boat used in this research.
Concentrations of dissolved oxygen in Lake Eola, although usually at or
above saturation near the surface, drop periodically during the spring and
summer months to less than 1 mg/1 in deep areas of 4.0 meters or more water \
column. Phosphorus from the bottom sediments was released up to a l;evel of
250 mg/nr after two months of anoxic conditions.
Bioassay experiments were utilized to study the impact of stormwater run-
off, untreated and coagulated on productivity of Lake Eola water. Addition-
ally, possible toxic effects from heavy metals released to the lake have been
discussed. It appears that bioassay techniques could become a useful tool
provided proper experimental conditions and test organisms are selected. The
following conclusions are reached:
1. Management of stormwater runoff is essential if restoration of
Lake Eola is to be successful.
2. Considerable amounts of nutrients are released to Lake Eola by
stormwater runoff. Most nutrients are released periodically for a
few hours at a time over a period of six months during spring and
summer. More than half of nutrients released are in soluble form.
368
-------
3. Phosphorus is probably the most important nutrient in regulating
algal production in Lake Eola. Addition of phosphorus to Lake
Eola waters will result in an increase in algal cell production.
4. From bioassay experimentation, stormwater up to 25% mixture with
lake water will enhance algal production. A mixture of lake water
and dry weather flow from storm drains did not produce changes in
algal production.
5. If stormwater runoff is coagulated with alum before its release to
. the lake, no significant increase in algal production will be pro-
duced.
6. Stormwater contains heavy metals and/or organic compounds which
may be in concentrations toxic to algal species in lake water.
REFERENCES
APHA-AWWA-WPCF-.. Standard Methods for the Examination of Water and Wastewater,
14th Edition, American Public Health Association, Washington D.C., 1975.
Bartlett, Larry, Rabe, Fred W.., and,Funk, William H. "Effects of Copper, Zinc,
and Cadmium on Selenastrum capricornutum," Water Research, Volume 8,
pp. 179-185, 1974.
Chiaudani, G. and Vighi, M. "The N:P Ratio and Tests with Selenastrum to
Predict Eutrophication in Lakes," Water Research, Volume 8, pp. 1063-
1069, 1974.
Filip, Daniel S., and Middlebrooks, Joe E. "Evaluation of Sample Preparation
Techniques for Algal Bioassays," Water Research, Volume 9, pp. 581-585,
1975. .
Greene, J.C., Soltero, R.A., Miller, W.E., Gasperino, A.F., and Shiroyama T.
"The Relationship of Laobratory Algal Assays to Measurements of Indige-
nous Phytoplankton in Long Lake, Washington," Biostimulation and Nutri-
ent Assessment, Ann Arbor Science, Ann Arbor, pp. 93-125, 1976.
Greene, Joseph C., Miller., William E., Shiroyama, Tamotsu, and Merwin, Ellen.
"Toxicity of Zinc to the Green Alga Selenastrum capricornutum as a
Function of Phosphorus or Ionic Strength," Proceedings: Biostimulation
and Nutrient Assessment Workshop, U.S. Environmental Protection Agency,
pp. 28-42, October 16-18, 1975.
Harper, Harvey H. Ecological Responses To Urban Runoff in Lake Eola, M.S.
Thesis, University of Central Florida, Orlando, FL, Dec., 1979.
Marshall, Frank. Phosphorus Interactions With Lake Eola Bottom Sediments,
M.S. Thesis, University of Central Florida, Orlando, FL, March, 1980.
369
-------
Miller, W.E., Greene, J.C., and Shiroyama, T. "Use of Algal Assays to Define
Trace-Element Limitation and Heavy Metal Toxicity," Proceedings of the
Symposium on Terrestrial and Aquatic Ecological Studies of the North-
west, Eastern Washington State College, Cheney, WA, pp. 317-325, 1976.
Schlindler, D.W., and Fee, E.J. "Experimental Lakes Area: Whole Lake Experi-
ments in Eutrophication," Journal Fisheries Research Board of Canada,
Volume 31, No. 5, pp. 937-953, 1974.
Shiroyama, T., Miller, W.E., Greene, J.C., and Shigihara, C. "Growth Response
of Anabaena flos-aquae (Lyngb.) De Brebisson in Waters Collected from
Long Lake Reservoir, Washington," Proceedings of the Symposium on Ter-
restrial and Aquatic Ecological Studies of the Northwest, Eastern
Washington State College, Cheney, WA, pp. 267-275, March 26-27, 1976.
U.S. Environmental Protection Agency. The Selenastrum capricprnutum Printz
Algal Assay Bottle Test, National Eutrophication Research Program,
Corvalis, OR, p. 82, 1978.
U.S. Environmental Protection Agency. Methods for) Chemical Analysis, of Water
and Wastes, Environmental Monitoring and Support Laboratory, Cincinattl,
OH, 1976.
370
-------
WATER QUALITY AND BIOLOGICAL DEGRADATION
IN AN URBAN CREEK
by
Robert Pitt* and Martin Bozeman**
Woodward Clyde Consultants
San Francisco, California
ABSTRACT
This preliminary report presents the initial results and conclusions
from the EPA-sponsored demonstration study of the water quality and bio-
logical effects of urban runoff on Coyote Creek, near San Jose, California.
This first phase included investigating various field procedures that would
be most sensitive in evaluating water, sediment and biological changes in the
creek as it passed through the urban area. The procedures identified.as most
promising are currently being used in additional Coyote Creek studies.
The report describes the characteristics of urban runoff affecting the
creek, sources of urban runoff pollutants, effects of urban runoff and
potential controls for urban runoff. Local urban runoff characterization
information is summarized, based on a previous EPA sponsored demonstration
project in the area (Demonstration of Non-Point Pollution Abatement Through
Improved Street Cleaning Practices-EPA grant No. S-804432, Pitt 1979)"and
from the local "208" study (Metcalf and Eddy 1978). Sources of urban runoff
pollutants in the study area are being investigated as an important part of
the field activities of the project and include sampling runoff from many
source areas (such as street surfaces, parking lots, landscaped areas,
rooftops and rain).
Various short- and long-term biological sampling techniques were used
to evaluate the fish, benthic macroinvertebrate and attache algae conditions
at many stations in the creek, above and within the urban area. Creekwater
and sediment samples were also obtained and analyzed for a broad list- of
parameters. In most cases, very pronounced gradients of these creek^ quality
indicators were observed, with the urbanized portion of the creek being
significantly degraded. Current additional monitoring is being conducted
to identify the urban runoff control goals necessary to improve creek quality
to adequate levels.
* Private Consultant, Route 1, Blue Mounds, WI 53517 (608) 437-3456.,
** Aquatic Biologist, Woodward-Clyde Consultants, 3 Embarcadero Center,
Suite 700, San Francisco, CA 94111 (415) 956-7070.
371
-------
INTRODUCTION
The purpose of this paper is to discuss the information that lias been
assembled from a variety of urban runoff studies. Some of these data and
findings have been, published recently; others are not generally available and
should be considered preliminary findings (i.e., subject to further review).
This information is presented here to stimulate interest and discussion by
those concerned with urban runoff problems.
This paper is separated into three major topics: (1) characterization
of urban runoff, (2) sources of ur.ban runoff pollutants, and (3) effects of
urban runoff. Much of this information is based on a recently completed EPA
project (Storm and Combined Sewer Section), Demonstration of Hon-point Pollu-
tion Abatement Through Improved Street Cleaning Practices (Pitt 1979) and the
initial phases of an on-going project (also .for the Storm and Combined Sewer
Sewer Section), Water Quality and Biological Effects of Urban Runoff on
Coyote Creek (Pitt and Bozeman 1979). Information has also been obtained
from other runoff projects that have also characterized the sources and prob-
lems of urban runoff and have developed potential control measures. Local
planning agencies that have participated in "208" Studies (pursuant to Sec-
tion 208 of the Clean Water Act [PL 92-500]) are very interested in urban
runoff. Approximately 30 demonstration projects have been started throughout
the country under the National Urban Runoff Program (NURP) of the EPA's Water
Planning Division. These projects will examine many aspects of urban runoff
pollution in various areas of the country.
Projects for the Storm and Combined Sewer Section have been conducted
by Woodward-Clyde Consultants (WCC) in study areas within the Coyote Creek
watershed in San Jose, California. The location of this watershed in rela-
tion to major cities in the San Francisco Bay Area is presented in Figure 1.
The Coyote Creek watershed is located about 50 miles southwest of San Fran-
cisco and typically receives 10 to 20 inches of rain per year. The water-
shed, which covers about 400 square miles, is mostly undeveloped, open
rangeland, semi-wilderness state park, and low-intensity agricultural land.
About 25 square miles at the northern end of the watershed is a heavily
urbanized portion of San Jose.
372
-------
The main channel of Coyote Creek is about 80 miles long; the southern
half lies in rugged coastal hills and the northern half in gently sloping
valleys. Coyote Creek empties into the southern end of San Francisco Bay.
Typical average daily flows in the northern part of the creek are less than
50 cubic feet per second. Major storm flows, however, can approach 1000
cubic feet per second; these flows are controlled by two dams.
The study area is located between the farthest downstream dam (Lake
Anderson) and the first major confluence (Silver Creek) which is well within
the City of San Jose. Of this 20-mile study section, approximately 5 miles
are urban and 15 miles are non-urban areas. The non-urban section is char-
acterized by relatively low-intensity agricultural or open space. Sampling
stations were located in both the urban and non-urban sections of the stream
(to allow for comparison).
CHARACTERIZATION OF URBAN RUNOFF
For this study, runoff water quality data obtained from the street
cleaning demonstration study (Pitt 1979) were examined and analyzed by an
equilibrixim water chemistry computer program to estimate both the specific
chemical compounds that would remain soluble and those that would settle out
into receiving water sediments. This information (see Table 1) shows that
most of the urban runoff pollutants are soluble and would be carried in the
water columns of receiving waters. Almost all (95 percent) of the lead com-
pounds present in urban runoff are expected to be insoluble particulates
and, depending on their size, may settle out in the sewerage or the receiv-
ing water sediments.* Other parameters that were monitored will mostly be
present in soluble forms. :
The information presented in Table 1 and data obtained from the Santa
Clara County area— wide wastewater management plan (Metcalf and Eddy 1978)
were used to estimate the urban area unit pollutant yields for the study
area. These estimated yields, on a pounds per acre per year basis, are pre-
sented in Table 2. The estimated yields of urban runoff affecting the moni-
toring stations in Coyote Creek are also shown. Non-urban stations are
affected by runoff from undeveloped and agricultural areas, and urban sta-
tions are affected by runoff from those areas as well as the runoff from
urban areas. The pollutant yields affecting the urban stations are all sub-
stantially greater than the quantities affecting the non-urban stations. For '
example, the total solids discharged into the creek in urban areas were more
than one hundred times those of non-urban areas. The amounts of lead dis-
charged at urban area stations were several thousand times those of npn— urban
stations.
SOURCES OF URBAN RUNOFF POLLUTANTS
A major area of uncertainty concerning urban runoff is how the contri-
butions from pollutant sources in the watershed contribute to the outfall
* This was substantiated in field studies by the large concentrations of
lead that were found in the urban creek sediments.
373
-------
' ' f %1V ' f ^X y'J "" £ ' >•
11 * ' > ^ < c
PACIFIC OCEAN <• - ^
'vX ""' "•',-. < 'N.,
^JiVl^'-j , j. " ,
COYOTE
CREEK
WA TE RSH E D
10 15
miles
Figure -1. San Francisco Bay Area showing the general
location of the Coyote Creek watershed.
374
-------
Table 1. FLOW-WEIGHTED URBAN RUNOFF CONCENTRATIONS OBSERVED AND
CALCULATED (mg/1, unless otherwise noted)
Range*
Parameter
Total Ca^^
Ca-"
CaS04
CaHCO^
Total Me
MgHC03+
Total K
K"1"
Total Na
Ha*
Total Cu
Cu+
CuC03
CuHP04
CuOH+
Total Cd
CdS04
CdCl*
Total Zn
Zn""
ZnP04
ZnS04
Total Pb(2,3)
PbC03
Hin.
2.8
2.8
^0»1
^0.1
1.4
1.4
0.1
1.5
1.5
2.1
2.1
0.013
0.01
<0.001
<0.001
<0.001
<0.001
<0.002
<0.002
<0. 001
<0.001
0.10
0.10
<0.001
<0.001
0.20
<0.001
<0.001
Max.
19
18.5
1.7
0.4
6.2
5.9
1.1
0.2
3.5
3.5
27
27
0.05
0.05
0.01
0.01
0.01
0.001
0.004
0.004
0.001
<0.001
0.32
0.31
0.003
0.02
0.76
0.01
0.02
Average
13
12.2
1.1
0.1
4.0
3.9
0.6
2.7
2.6
15
15
0.03
0.02
0. 004
0.005
<0.001
<0.001
0.002
0.002
<0.001
<0.001
0.18
0.18
0.001
0.011
0.42
0.003
0.01
Range*
Parameter
Total Cr
Cr(OH)2
CrOn*
Total C03(2)
Total HC03
HC03~
Total S04
so4~
Total CL
CL~
Total Ortho_P04(3>
MgHP04
Total N03
N03~
Total Hg
Kjeldahl N
pH, pH Units
ORP, mV
e
Temp. , C
Sec. Cond, umhos/cm
Turbidity, NTU
Total Solids
Total Dissolved Solids
Suspended Solids
Volatile Sus. Solids
Dissolved Oxygen
BOD5
COD
TOC
Hin.
0.009
<0.001
CO.OOl
<0. 001
•C0.001
<0.001
6.3
6.2
3.9
3.9
0.2
0.2
<0,001
0.3
0.3
<0.0001
3.1
6.3
78
15
33
37
110
34
41
5
6.5
17
77
19
Max.
0.03
0.05
0.03
0.055
150
0.01
27
25
18
18
6.0
4.4
0.008
1.5
1.5
0.0002
15
7.0
140
15
130
86
680
160
570
40
9.9
30
350
290
Average
0.04
0.02 ,
0.006
0.019
53
0.007
18
17
12
12
2.4
1.7
0.002
0.7
0.7
<0. 0001
6.7
6.8
120
15
93
49
300
91
210
23
8
24
200
110
NOTE: The following coapounds ate expected to occur in urban runoff in solid form:
(1) CaHP04
(2) PhC03
(3) Pt3(P04)2
*The Binimum and maximum range values represent individual atom flow-weighted average values. There
is much larger variation when individual values during the atoms are considered.
Source: Pitt and Bozenan 1979
375
-------
/•"N
H
fTl
W
0)
rS
CO
o
4J
>—x
2:
0
I-H
H
<$
H
co
pa
9
a
o
5S
M
EH
O
w
fe<
%
***
o
tJ
w
M
>•<
P*4
b
^^
§
Pi
J2J
*2
CQ
g
rJ
S3
s
•<
•
OJ
w
CP
«<3
EH
m
O
,ti
a
a
M
(1)
a
cd
•u
CO
•H
O
t3
C
cd
c
0
fH
cd
4J
CO
*
x>>
CO
Q)
i-H
Jjjjj
^
^
0)
QJ
Jj
^J
,
U
cd
0)
l (_(
O
•u cd
-r4-^-
C rQ
S3 rH
Vl
0}
1
^
cd
&
0 O O
r-- CM rH
o o o
VO CM rH
o o ^o
in i— i
O/— . /— >
^-M? ^^
v v y
V V V
v v v
O<— ^ r~i
\_J VwX
m oo Q Q
O O O
H U M
\O
^
VO
VO
•*
in
r-H
O
V
rH
Q
\/
rH
O
V
in
CM
13
3
•0
VH
V
$4
O O
-* CO
m r--
CO CM
OO CM
CM CM
rH
rH O
0 O
V V
rH rH
0 O
O O
w
rH ' — 1
0 0
0 0
V V
CO O
rH rH.
-3-
2
i
o
rC3 TJ
4J Cd
S-l ,*,
£•4
cfl
*^3
c
o
rQ
«
CD
cd
CJ
"S
3
CD
fH
•K
a\
r-.
a\
r^
(3
N
O
m
"O
c
cd
4-1
4-1
S
• «
0)
o
3
O
CO
376
-------
Table 3. URBAN RUNOFF POLLUTANT CONCENTRATIONS FROM MAJOR AREAS
(mg/1, unless otherwise noted)
Parameter
pH, pH Units
Specific Conductance,
(umhos/cm)
Turbidity, NTU
Total Solids
BOD 5
COD
0-P04
Total P04
Kjeldahl N
NH3
N03
S
so4
As
Zn
Pb
Cr
Cu
Total Coliform Bacteria
(MPN/lOOml)
Fecal Coliform Bacteria
(MPN/lOOml)
Fecal Strep. Bacteria
(MPN/lOOml)
Fecal Coliform/Fecal
Strep. Ratio
Outfall
7.8
185
29
162
8
97
0.23
0.34
1.52
0.25
0.74
4
13
<0.01
0.06
0.08
0.009
0.013
>2400
>2400
>2400
—
Gutter Flow
7.5
130
100
235
13
172
0.12
0.31
2.41
0.42
0.42
2
7
<0.01
0.14
0.67
0.049
0.029
>2400
920
>2400
<0.4
Parking
Lot
7.0
45
26
340
22
176
0.47
0.49
1.47
0.35
0.13
<1
<1
0.02
0.23
1.09
0.071
0.046
540
350
>2400
<0.2
Park
Puddles
7.3
2400
21
2140
3
69
0.32
0.42
1.32
1.23
285
15
38
0,10
0.01
0.035
0.010
0.031
49
49
920
0.5
Commercial
Tar and
Gravel Roof
7.5
155
1
186
7
131
0.02
0.07
4.37
1.06
0.22
5
21
<0.01
0.08
0.019
<0.005
0.11
170
9
17
0.5
Residential
Composition
Shingle Roof
6.5
11.2
a
18
3
19
0.08
0.10
0.71
0.50
0.09
a
<1
0.01
0.18
0.017
<0.005
<0.005
<2
<2
920
<0.002
Rain
6.4
10.4
<1
30
3
12
0.03
0.03
0.64
0.36
0.09
<1
<1
<0.01
0.04
<0.01
<0.005
0.010
8
2
<2
^
Source: Pitt and Bozeman 1979
377
-------
yield. Sources that are farther from the storm drainage system and require
overland flow have a very low yield when compared with parking lots or street
surfaces that are impervious and located adjacent to drainage systems. Pre-
liminary results from the Coyote Creek project (Pitt and Bozeman 1979) which
examined potential sources of urban runoff pollutants and involved the col-
lection of runoff samples during rainstorms from small areas within San Jose
(e.g., building roofs, parking lots, gutter flows) are presented in Table 3.
In that study, rainfall and outfall samples were collected for chemical
analyses. As expected, in almost cases rain had the lowest pollutant concen-
trations while parking lots and gutter flows had the greatest. Puddles that
were monitored in a city park had much greater concentrations of total sol-
ids, nitrates, and higher values of specific conductance exhibited than any
of the other samples.
Table 4 is a generalization of common urban runoff pollutant sources.
Ho particular source is expected to contribute significant quantities of all
the pollutants, but some of the sources are expected to be important contri-
butors. For example, street surfaces contribute significant amounts of
heavy metals; oxygen—demanding materials and nutrients are thought to come
mostly from landscaped areas and vacant land.
Table 5 is a generalization which shows the pollutant contributions to
major urban surfaces and the approximate delivery yields to the outfall.
Vacant land and landscaped areas are among the most pervious surfaces in
urban areas and are located farthest from urban drainage systems. Therefore,
they contribute little flow and only a small fraction of the pollutants in
urban drainage systems. For example, landscaped areas and vacant land com-
prise almost half of the land surface in the San Jose study area, but only
about 5 percent of the rain that falls on these surfaces contributes to the
outfall flow. Similarly, very little of the potential pollutant yield from
these surfaces is expected to affect the outfall.
Rooftops, which comprise 15 to 20 percent of surface areas in the San
Jose study area, are also located relatively long distances away from the
storra sewerage system. Rooftops are impervious; however, most of the roof
drainage systems in San Jose are not directly connected to the storm sewer-
age system and require considerable overland flow. Therefore, the outfall
runoff yield from rooftops is expected to be only about 30 percent.
Sidewalks, which comprise about 5 percent of the San Jose study area,
are located closer to the storm drainage system, but some of their drainage
flow is directed towards adjacent landscaped or other pervious areas. There-
fore, only about half of the runoff yield from sidewalks enters the receiving
378
-------
f-l
M »H 6) «M
Vi -U O 4-> U-t
0) a co co o
.e 3 « c
M QJ
1
•u
y tQ
3 01
'CO CO
C
o
4J
a c
O Qj
$
•o
a)
ex
at
U 03
a> «
•o a)
00
C
WH co
£
• co
a> o
01 Sl
CO 9
CO
CO
o.
s
*o
o
ei
X
X
X
diment
01
CO
X
3
•a
II
& 33
X
X
CO
trient
:?
a
i
X
•2
u
3
X
• X
m
a
4J
w
X
CO
01
•o
•H
U
I
CO
0)
sticid
£
X
- •
X
X
X
a)
CO
a
"S
0
CO
i-l
3
X
w
S
4J
1
4J
s
1-1
If,
X
X
X
CO
iH
a
•rl
01
4J
3
u
.&
1
4J
0
379
-------
Table 5* POLLUTANTS CONTRIBUTED BY VARIOUS URBAN SOURCES
Pollutant Sources
Dust fall
Precipitation
Tire Wear
Auto Exhaust Particulates
Other Auto Use
(Fluid Drips, Wear Prod.)
Vegetation Litter
Construction Erosion
Other Litter
Bird Feces
Dog Feces
C*t Feces
Fertilizer Use
Pesticide Use
De-Icing Use
Proportion of Urban
Surface Area
Delivery Yield
to Outfall*
Lawn and
Landscaped Areas
X
X
X
(Adjacent)
X
(Adjacent)
X
X
X
X
X
X
402
5%
Vacant Lots
X
X
X
X
X
X
X
X
102
5%
Rooftops
X
X
X
15-202
30%
Sidewalks
X
X
X
X
X
X
X
X
X
5%
45%
Parking Lots
X
X
X
X
X
X
X
X
7X
50%
Street Surface
X
X
X
X
X
X
X
X
x
15-20;:
75%
* This is the portion of the total pollutant load
45JE of the pollutants which are washed off side
for.
that leaves the surface and eventually arrives at the outfall (e.g.,
walks get to the outfall; 55% is "lost" enroute and cannot be accounted
380
-------
xvaters. Parking lots, which comprise 7 percent of the San Jose study area,
are mostly paved and impervious. Again, some of the runoff from the parking
lots (especially at hones and apartments) is directed towards adjacent imper-
vious areas, and only about half of the parking lot runoff is expected to
reach the receiving waters.
Street surfaces, however, are located very close to the storm drainage
system and are almost impervious. They comprise about 15 to 20 percent of
the study area, and most of the runoff from the street surfaces is expected
to reach the outfall. However, some of the street surface flow does not
reach the outfall, possibily because of evaporation and infiltration through
streets in poor condition,,
Dust fall and precipitation affect all of these major components. Al-
though not a major pollutant source, dust fall is a mechanism for transport
of pollutants. Most of the dust monitored in urban areas is resuspended
particulate matter from street surfaces or wind erosion products from vacant
areas. Some point source air pollutant emissions also contribute to dust-
fall pollution, but the bulk is contributed by other sources already consi-
dered (e.g., street dirt and other fugitive dusts).
Automobile tire wear is a substantial source of zinc in urban runoff
and is mostly deposited on street surfaces and adjacent areas. About half of
the settleable particulates resulting from tire wear settles on the street
and the remaining material settles within about 20 feet of the roadway. Auto
exhaust particulates also contribute significant amounts of heavy metals (es-
pecially lead) and mostly affect street surfaces and adjacent areas. Other
automobile pollutant contributions are associated with fluid losses from
drips and spills, and by mechanical wear products. Heavy metals and asbestos
are also associated with those losses. Most of these pollutants directly
affect parking lots and street surfaces and some pollutant material lands on
adjacent areas through wind and automobile-induced air turbulence.
Vegetation litter can be a significant pollutant component in most
source areas. For example, leaf fall in San Jose is an important street sur-
face pollutant in the autumn. Although animal feces can contribute signifi-
cant quantities of nutrient and bacteria to urban areas, this pollutant
source mostly affects vacant land and landscaped areas. Fertilizer, pesti-
cide, and herbicide use is generally associated with landscaped areas, but
large amounts are routinely used to manage plant growth on and adjacent to
impervious surfaces. Fertilizers may be used in large quantities in road
landscaping projects. In cold climates,- the use of deicing materials also
contributes salt and particulate materials to roadways,. parking lots, and
adjacent areas.
Table 6 is based on preliminary results from the Coyote Creek project
and estimates the percentage contribution of the various pollutants from the
different source areas studied. In most cases, rooftops contribute the least
amount of pollutants while the pervious areas contribute the majority of
381
-------
3
£
H
r*
Et,
O
PS
rJ
1
g
0
H
V)
§
CO
,0
0
H
1
•&
8
K
z
0
M
£a
PQ a
H 0)
Cd 0
H IH
S5 00 CO
o u c w
iu -a c ^ 01 B
o B co v l-i co
PS rH Pk CO
(0
rH (X-H
O Ol >-i
O 1-. 01
01 4-1 tJ
b C/3 O
a
ctj O K
o »w 01
0) -H U
to rH O
O CO
O m
O -H 4-1
H rH CJ
3
O,
-------
solids, oxygen demanding materials, and some nutrients. Parking lots, street
surfaces, and sidewalks are expected to contribute some nutrients as well as
a proportion of heavy metals and bacteria, to the total outfall runoff yield.
Local erosion products and some contaminants from motor vehicles account
for most of the street surface dust and dirt material (by weight). Where
streets are in good repair, only minor contributions are made by wear of the
street surfaces. The specific makeup of street surface contaminants is a
function of many site conditions and varies widely. Many pollutant sources
are specific to particular areas and on—going activities. For example, iron
oxides are associated with welding operations, and strontium (used, in the
production of fireworks and flares) would probably be found on streets in
greater quantities around holiday times or at the scenes of traffic acci-
dents.
Relative deposition values for the different pollutant sources are sum-
marized in Table 7. These values represent the percentage of the total pol-~
lutants deposited in urban areas; however, the amounts deposited exceed the
pollutant yields to the outfall. In comparison, Table 8, shows the relative
yields from these pollutant sources to the total outfall runoff yield. The
deposition rates of some pollutants are relatively high for some impervious
areas, but these source yields are reduced substantially when infiltration
is considered. Automobile activity is responsible for most of the heavy
metal yield in the runoff and about half of the total solids yield. Vegeta-
tion sources contribute most of the oxygen-demanding materials, while animal
feces and fertilizers are thought to contribute most of the nitrogen in
urban runoff.
A summary of predicted San Jose urban runoff characteristics is pre-
sented in Table 9. Conditions vary substantially for other areas of the
country but these data do point out some important aspects of urban runoff.
Of all the total solids deposited in urban areas, only about one-third will
reach the outfall. Only about 10 percent of the nutrients and oxygen-demand-
ing materials deposited would affect the receiving water quality, but most of
the heavy metals deposited in the area would affect the receiving waters.
The remaining pollutants that are washed off the source areas and do not
reach the outfall would be accumulated in other areas of the urban environ-
ment. The most significant: pollutant "sinks" in urban areas are soils,
groundwater, and plants. For example, many studies have shown significant
concentrations of heavy metals in roadside soils and vegetation (Farmer and
Lyon 1977; McMullen and Faoro 1977; Olson and Skogerboe 1975; Pitt and Amy
1973). Deicing salts have also been shown to affect shallow groundwater near
urban areas. Much of.this material (about 15 percent of the total deposition
of total solids) can be associated with dust fall. Most of this dust fall,
however, consists of resuspended particulates from streets and vacant areas
and, except for point source air pollution emissions that may settle out, is
not an actual source of urban runoff pollutants.
383
-------
Table 7. RELATIVE ANNUAL DEPOSITION VALUES FOR POLLUTANTS FROM VARIOUS
SOURCES (percent)
Source
Precipitation
Tire Wear
Auto and
Street Use
Vegetation
Construction
Erosion
Bird Feces
Dog Feces
Cat Feces
Fertilizer Use
Other
Total
Solids
10%
5
20
40
20
2
3
BOD5
3%
<1
1
95
<1
1
TKN
5%
1
15
<1
<1
40
<1
40
Pb
<1%
10
90
<1
<1
Zn
15%
80
1
2
<1
Cr
<1%
10
70
10
10
Cu
10%
5
50
2
30
Table 8. RELATIVE POLLUTANT YIELDS OF TOTAL OUTFALL RUNOFF (percent)
Source
Precipitation
Tire Wear
Auto and
Street Use
Vegetation
Construction
Erosion
Bird Feces
Dog Feces
Cat Feces
Fertilizer Use
Other
Total
Solids
10%
5
50
30
5
5
3
BOD5
10%
1
10
70
<1
5
TKN
20%
10
10
<1
1
30
<1
30
Pb
<1%
1
100
<1
<1
Zn
20%
80
<1
<1
Cr
<1%
3
90
1
5
Cu
5%
2
80
<1
15
384
-------
CO
o
H
H
CO
H
3
H
O
O
Pn
Pn
O
S3
§
1
P
W
w
o
>D
5
CO
Q
H
H
C_>
W
ft
1
Ov
0)
rH
-s
H
u
a
5
.n
04
R
f
3-3
4J -rl
^5
H
d
rH
d
in
rH
d
o
rH
10
en
0
o
«H
1
It
*3 o
*Jj^
3 >o
i
TJ
V T3
4J rH
OJ -H
1
3
O
*
o
f-
o
o
lu
*J C M to
«
a)
O T3 0) CD
a 8 £ O
CU H
s^
a)
«
«
(0 N
U >H
O 13 rH 0)
« a i-t a
00 CU
(5
g
•H
3
u
1
s-§s s
.4'-SB
W)
^
4J
?*
83
O >»
CO MH
«) m
o o
3 3
o pi
to
o
1,"
a
385
-------
EFFECTS OF URBAN RUNOFF
This section summarizes the field and laboratory activities conducted
to examine the water quality and biological effects of urban runoff on
Coyote Creek. Preliminary work involved sampling Coyote Creek during the
fall of 1977 and spring and early summer months of 1978. The objectives
were to characterize (on a macro-scale) sediment quality and biological con-
ditions as the creek passed through urban areas. The types and magnitude of
changes in the sediment quality and biological community in Coyote Creek were
examined as the creek passed through San Jose* between Lake Anderson and the
confluence of Silver Creek. With the exception of urban runoff, in this
reach of Coyote Creek there are no pollutant discharges. Each sampling area
included a stretch of stream several hundred yards long.
The following factors were analyzed in this study:
o runoff (various water quality parameters)
o sediment (BOD5, COD, TOC, volatile solids, NH3, N03, organic N,
ortho PO*, SO/, complete elemental scan by SSMS including heavy
metals, complete organic scan by MSGC including pesticides and
PCB's, and particle size distribution)
o tissue analyses (for lead and zinc)
o fish
o benthic organisms (including aquatic insects, crustaceans, and
molluscs)
o attached algae
o rooted aquatic vegetation (where present).
While the sediment and water column samples were only collected once,
most of the biological sampling was conducted in two consecutive periods dur-
ing the spring and early summer. The following paragraphs briefly describe
some of the water quality and biological changes observed in Coyote Creek.
* This urban area includes the Keyes Street and Tropicana study areas that
were investigated as part of the San Jose street cleaning demonstration
project.
386
-------
Runoff Water Quality
The ranges and average runoff pollutant concentrations for several
monitored storms in San Jose were presented in Table 1. In the urban areas,
Coyote Creek had water quality concentrations similar to those of runoff dur-
ing storms. These water quality conditions only occur during storm runoff;
the water quality in the urban areas of Coyote Creek improves somewhat during
dry weather.
In Table 10, observed runoff concentrations are compared with recom-
mended water quality criteria for various beneficial uses. The water quality
criteria values for these uses are recommended maximum limits designed to
protect the beneficial uses with a reasonable amount of safety. If a moni-
tored concentration exceeds these criteria, it does not mean that a problem
exists but that one could occur,,and that additional monitoring may be neces-
sary (to define the relationships between water quality and impairment of the
beneficial uses for the specific receiving water). Beneficial uses that are
actually impaired by urban runoff must be carefully considered. For example,
no one would swim in the creek during storms, but aquatic life could be
severely impacted (intermittently) during storms.
The following list summarizes the parameters that exceeded the recom-
mended beneficial use criteria:
Livestock: Pb*
Wildlife: None '-. ,
Aquatic life: Cr, Cd*, Pb*, Hg*, BOD5, turbidity*, suspended solids*
Marine life: PO^*, Cd, Cu, Zn
Recreational uses: PO^*,
Freshwater public supply: Cd, Pb*
Irrigation: Cd.
These data indicate that those parameters most potentially responsible for.
water quality impairment are solids, cadmium, lead, and mercury for aquatic
life uses; orthophosphates for marine life; orthophosphates for eutrophica-
tion; and lead for freshwater public supply.
Limited data are available concerning the water quality of Coyote Creek
in urban areas during dry weather and in non-urban areas during wet weather.
These data are summarized in Table 11 and indicate that the high suspended
solids and phosphate concentrations in non-urban areas during wet weather can
greatly exceed the aquatic life and marine life criteria. The phosphate con-
centrations may also exceed the criteria for recreation.
*The maximum monitored value was greater than 10 times the minimum recom-
mended criterion.
387
-------
3
H
H
f*5
O
w
CO
i
o
H
En
i>*
sS
g
Q
1
S
s
u
H
o
&
H
•5
^
t*t
1
P
«
o
0)
H
1
J3
kl
£J
w
«
a
•H
•
2
S
£
h>
Freihvate
Public
Supply
1
*M
O
I
0
IK
Is
t)
2
«J
IM
•-«
S
i
*>
§
4J
s>
tJ
l-l
^
5Ei?
ill
JM
« i
1 1 1
I 1 1
lit
m *i
1 o*S
§ *
900
* a a
Q
•4 ^ •«•
«n «
*
•0 *
* IN tQ
&
•H !!
•SooKS
«l • «n « •) •
« t ft S 2 o
:!
!i-
•^.
MOO
1 *§ 1 - 1 1 1
0
Ij
« rt ff
0 p
^)
• ^ 4J
• IM
« S
0 o
1 ! ! ! 1 . *
1 1 1 1 1 4
0
1 1 1 1 1 1
i^
M! ' 1 ' !
sl ^
^
i " i *i
1 > ' ' i ol
a a o
f- •» O
o CN CM «o m o
in *^ *^ •
vrt ^M rN «M CM 4
i M i i i i?
0 0 « » N 0°
81 £* S ^ «1 S
m « in
010
0 — 0
i i :
rt 5 S
* * i
0 0 X
b
*j
* M » rt
TJ "^ l> "^
*r> B *
4 0 J-0
0*0
1 1 1
> ff ff
O M-l .-
-^00
0 K® K^
^.^S^o
o o " in
N f>
oo-*
000
Ot
• .
4
000 2
o o o o
S 3 £
•H
8 ^ S
°. * s
0 v> 8
1 i 1
< 51 5
60 DC U
• • S
5 5 5
.<
ff «
^ iH
§4J M
•? 1 s
o S
1
pi !
t* *J
5E
I » 1
e5 a
K
.H 1 1 1
•
_ ^
1 2
^ o a
m
»
4 ° *4
llo *
000 ~
ff A! I
e
s
s
. iJ
i H
U M
< « O
OM *H
1 ||
(^ -0
Ot to V
— • i<
BJ ° ""
B 515
s ;|
s 7*
£ Is
g Ǥ
s ss
i «S
s
s I i
B SI
M * ^
388
-------
Table 11. WATER QUALITY CONDITIONS IN COYOTE CREEK BY LOCATION AND
WEATHER CONDITION
Parameter (mg/1 , unless
otherwise noted)
pH (pH Units)
Temperature (°C)
Ca
Mg . . '
Na
K
HCO,
CO,
SO,
Cl
Total Hardness
Total Alkalinity
Total Solids
TDS
Sus. Solids
Vol. Sus. Solids
Turbidity (NTU)
Spec. Cond. (umhos/cm)
Dissolved Oxygen ' ' , •
BODc
COD
Kjeldahl N '•_
NO,
NO,
NH,
0-POt .-" ;
TOC
Pb
Zn
Cu
Cr
Cd
Hg
Total Coliform Bacteria (MPN/lOOml)
Fecal Coliform Bacteria (MPN/lOOml)
Fecal Strep. Bacteria (MPN/lOOml)
Urban
Wet
Weather
6.7
16
13
4.0
0.01
2.7
54
0.019
18 •
12
_*
- • •'
310
150
210
23
49
160
8.0
24
200
7
0.7
-
_
2.4
110
0.4
0.18
0.03
0.02
<0.002
2400
>2400
>2400
Area
Dry
Weather
7.8
15
-
• -
—
-
130
<24
55
60
250
240
— f ,-!.
' -
-
-
14
520
7 .6
-
—
-
0.84
0.'024
', 0,95
-
-
-
'
—
-
-
>1900
-
<•»
Non-Urban
Wet
Wether
_
- • '
_
-
—
- •
-
-
-
—
•-
• - : ;:
—
•-
600
90
.
, - • _
•- " -
4 '
—
2
-
-
^
0.8
-
-
•; . - •
'.- . '; - • • '
. ' — • •
-
-
-
-
—
Area
Dry
Weather
7.9
15
40
25
19
1.9
170
3.0
37
15
200
140
— • •'•
280
;. -
-
18
400
11
-
—
- •
; 1.2,
: • <0.002
0.33
« 0.6
-
-
:'•:.. - '• '•
— ' .
-
- .
>1300
- •
—
The blanks signify where no data is shown.
389
-------
A comparison between secondary sanitary wastewater effluent concentra-
tions and urban runoff concentrations for the study area is presented in
Table 12. The average and peak 1-hour runoff concentrations and the average
secondary sanitary wastewater effluent concentrations are shown. This sani-
tary wastewater treatment facility is a modern, advanced, secondary treatment
plant for the study area. The short-term effects of urban runoff on receiv-
ing waters occur (by definition) during and immediately following a runoff
event (i.e., short-term effects are associated with instantaneous concentra-
tions). A comparison between urban runoff average concentrations and the
sanitary wastewater treatment plant effluent average concentrations shows
that the concentrations of lead, suspended solids, BODc, COD, TOG, turbidity,
zinc, chromium, and cadmium are all higher in the runoff than in the sanitary
wastewater effluent. In addition to the previously listed parameters, copper
and Kjeldahl nitrogen have greater runoff peak concentrations than the ave-
rage wastewater concentrations. Therefore, urban runoff may have more impor-
tant short-term effects on receiving waters than average treated sanitary
wastewater effluent concentrations.
Annual yields of these sources provides an indication of potential long-
terra problems. Table 12 also shows the annual sanitary wastewater treatment
plant effluent yield expressed as tons per year (derived from monthly average
concentrations and effluent quantities) and the calculated street surface
portion of the annual urban runoff yield expressed in tons per year for a
similar service area. On an annual basis, the total orthophosphates and
Kjeldahl nitrogen yields associated with street surface runoff alone are less
than 4 percent of the total yield from sanitary wastewater treatment plant
effluent and street surface runoff. Total solids, cadmium, and mercury in
the street surface runoff contribute from 5 to 10 percent of the total yield
(respectively), while COD, BODc, and copper contribute from 10 to 50 percent
of this total yield. Suspended solids, chromium, zinc, and lead in the
street surface runoff contribute more than 50 percent of the total yield.
These data show that, for receiving waters in which both secondary
treated sanitary wastewater and untreated urban runoff are deposited, addi-
tional improvements in the sanitary wastewater effluent may not be as cost-
effective as some urban runoff treatment (except for nutrients). That would
be especially true for lead, where more than 95 percent of the total waste-
load is from street surface runoff. For example, if all of the lead were
removed from the sanitary wastewater effluent, the total annual lead dis-
charge would only decrease by about 4 percent.
Grab samples of Coyote Creek water were analyzed for major parameters
during dry weather. The water was very turbid, hard to very hard, and had
high ammonia and coliform bacteria concentrations. Figure 2 shows a marked
increase in nitrites, ammonia, turbidity, chlorides, and specific conduct-
ance as the creek passed through urban areas of the watershed. The concen-
trations are expected to be 'greater and the trends more evident during and
immediately after rains, as previously shown in Table 11.
390
-------
H
S
W
3
Pn
fn
W
H
a
PM
H
a
§
CM
pH
(U
0) >w o t-t
v S § w
US C/l DS O
§•3!
o o o o <
OOOOOOOOOOOOOOrHOrHO-tfOO
S to «w
•< >*-i
O O O O i
o o o o «
O CM r- O <_ _ _
co cn -* • * * •>
o m o CM
cn -a* CM <*i
en
CM
cu,o • I o rH r-» r- o I i I
rH *-> CJ - CM ^O
„ OPJ OOOOOOOOrHOi-HOOOCMOCMi-nOvOO C3-O
^ (^ 4J r • — -• " -
O (
o oou-j fc * OrHr-nr>.'d-cMcn«* com-*in r- m vo
•fH>*HtMCj CMrHrHOOrHCMtHrHvOCMiHcnrHCM cncn rHO Oin
*J •< O Ui C *•*•••*»••••••• •* •• ••
CMCnr-ICM OO ' ~
> » ^2^ rtrH
IcSsS^
S I OxcO^Ot'-CO!'-. OrHOOmcaOOOOOOrHOO OOr
t-i rH »H cMrHcvim fnincMrHmcom vocn _
c1 (ncM-*mcMco-^cD- m •* = V CM-4"
CM CO rH CM rH
£0
C/5
-
OOOOO
i eu
o) o.
C/l (0
O 3 -
a cj to w S
•H 3 O -H
>W Tj J3 TJ 33
•H (3 g -H O.
a o 9 .a X^*H
^D
U O O
(U CO 4-1 4J
co co a
* «' n
*J CM t) JS
p ^ C c8
flt O
I-H
-------
0.05
AMMONIA,
MG/L
2.00-
1.50-
1.00-
0.50-
n .
\
\ ^
m
CCD
?b
175
150
125
100
75
50
25
0
ui
ui
Q
CC
O-I
300
250
200
150
100
SO-
0-
1
is
"by
1800
1600
1400
1200-
1000
800
600
400
200-
0
APPROXIMATE
QOYOTE CREEK
MILEPOSTSAND
MONITORING
STATION LOCA-
TIONS
10 12 14 16 18 20 22 24 26 28 30 32 34 36
(7) (?) (?) (7) (a) (o) rto)
ANDERSON
DAM
URBANIZED AREA
OF WATERSHED
Figure 2. WATER QUALITY TRENDS ALONG
COYOTE CREEK (March 31, 1977)
392
-------
Sediment Analyses
Sediment samples were obtained by carefully scooping bottom material
into glass jars. These jars were then sealed underwater. The samples were
frozen and delivered to a laboratory for analyses. As shown in Figures 3
and 4 and Table 13, samples from urban areas contained higher concentrations
of many of the parameters in comparison with samples from non—urban areas.
Sulfates (33 to 60 times greater), lead (10 times greater), arsenic (9 times
greater), BODr (up to 4.4 times greater) and orthophosphates (up to about
4 times greater) are examples of these higher concentrations. In addition,
much more silt was found in samples from urban areas. The urban samples
also had significantly greater concentrations of high molecular weight oxy-
genated hydrocarbon compounds.
Tissue Analyses
Selected samples of fish (Gambusia affinis), filamentous algae (Clado-
phora sp.), crayfish (Procambarus clarkii), and cattail plant segments
(Typha sp.) were collected from both the non-urban and urban reaches of
Coyote Creek. Each sample was chemically digested and analyzed for whole
organism lead and zinc concentrations. As shown in Tables 14 and 15, lead
concentrations in urban samples of algae, crayfish, and cattails were 2 to
3 tines greater than in non-urban samples. Zinc concentrations in urban
algae and cattail samples were about 3 times those of the non-urban sample
concentrations. Fish, lead, and zinc concentrations did not increase notice-
ably. Concentrations of lead and zinc in the organisms were greater than the
concentrations in the sediments for many of the samples and stations (up to a
maximum factor of about six),. The accumulation of zinc in the crayfish,
attached algae, and fish was at least 300:1 when compared with zinc concen-
trations in the water. The accumulation of lead in attached algae in the
urban areas of Coyote Creek was at least 500:1. For crayfish, the accumula-
tion of lead was at least 100:1 compared to lead concentrations in the water.
Biological Characterization
Fish, benthic inacroinvertebrates, and attached algae were collected from
representative habitats throughout the study area. Seines were used to cap-
ture fish from both riffles and pools. Captured fish were identified and
counted and the total length of each individual was recorded. Replicate ben-
thic macroinvertebrate samples were collected from natural substrates in both
pool and riffle habitats with an Ekman dredge and Surber sampler. In addi-
tion, artificial substrates (replicate pairs of Hester-Dendy multiplate samp-
lers) were used at each sampling area. The benthic samples were washed
through a 500\i sieve, and the organisms retained on the screen were picked,
sorted, and preserved in a 70—percent ethanol solution for later identifica-
tion and enumeration. Attached algae were sampled from both natural and
artifical substrates throughout the various reaches of the stream. Qualita-
tive samples of attached algae were collected by scraping uniform areas of
393
-------
I
TJ
8
o
O-
O
7-
6-
5-
4-
3-
2-
1-
0
ORTHOPHOSPHATES
urbanized
6 5
40,000-
§ 30,000-
20,000-
10,000-
TOTAL ORGANIC CARBON
urbanized -
6 S
s
Q
O
CD
2,000
1,800-
1,600
1,400-
1,200
1,000
800
600
400-
200
0
BIOCHEMICAL OXYGEN DEMAND
(5-day)
urbanized •
654 3
STATIONS (relative locations)
Figure 3. SEDIMENT QUALITY CONDITIONS
ALONG COYOTE CREEK
394
-------
65 4
2 1
to
I
1000'
800'
600
400-
200-
0
SULFUR
urbanized
654
2 1
adiment —
2
a.
£
400-
" 300-
200-
100-
0
V i LEAD
\
\ urbanized I
I
1 — ' ' r ~~" 7~, — 1~~ ~ . . l ; — l — '
65 4
*2 1
400-I
300-
e 200
3.
I 100
urbanized
MEDIAN SEDIMENT
PARTICLE SIZE
654 3 . 21
STATIONS (relative locations)
Figure 4. SEDIMENT QUALITY CONDITIONS
ALONG COYOTE CREEK
395
-------
Table 13. SEDIMENT CONCENTRATION INCREASES BETWEEN STATION 2 (NON-URBAN)
AND DOWNSTREAM STATIONS
Greater than lOx
Scat Ion 2 Values
Between 3.0 and
lOx Greater than
Station 2 Valued
Between 2.0 and
2.9x Greater than
Station 2 Values
Between 1.3 and
1.9x Greater than
Station 2 Values
Stations
3 (Non-Urban) 4 (Urban) 5 (Urban) 6 (Urban)
Nickel (8.0)
Chromium (6.5)
Cobalt (2.9)
Manganese (2.6)
Tantalum (2.3)
Scandium (1.9)
Sulfate (60)
Sulfur (3.8)
Hafnium (2. A)
Chromium (2.0)
Ytterbium (2.0)
Erbium (1.9)
Barium (1.5)
Tantalum (1.3)
Sulfate (33)
Lead (10)
Arsenic (8.7)
Hafnium (4.7)
Gallium (2.4)
Tantalum (2.3)
Thorium (2.3)
Sulfur (2.1)
Antimony (2.1)
Niobium (2.1)
Cadmium (2.0)
Ytterbium (2.0)
Yttrium (2.0)
Chlorine (1.9)
Erbium (1.8)
Neodymium (1.8)
Silver (XL. 7)
Zinc (1.7)
Ortho P04 (1.7)
Phosphorous (1.7)
Europium (1.5)
BOD5 (1.5)
Lanthanum (1.4)
Thallium O1.3)
Holmlum (1.3)
Selenium (1.3)
Sulfate (44)
Lead (11)
Arsenic (8.7x)
Thallium (4.8)
Hafnium (4.7)
BOD5 (4.4)
Praseodynium (4.4)
Ortho P04 (3.9)
Silver (>3.5)
Erbium (3.5)
Ytterbium (3.5)
Tantalum (3.0)
Sulfur (2.7)
Cadmium (2.6)
Tungsten (2.6)
Lanthanum (2.5)
Bismuth (>2.4)
Thorium (2.3)
Yttrium (2.3)
Antimony (2.1)
Lutecium (2.1)
Gallium (2.0)
Europium (1.9)
Gadolinium (1.9)
Niobium (1.9)
Uranium (1.9)
Chlorine (1.7)
Germanium (1.7)
Tin (1.7)
Titanium (1.7)
Mercury (1.6)
TOC (1.5)
Thallium (>1.3)
Holmium (1.3)
Selenium (1.3)
COD (1.3)
396
-------
Table 14. LEAD CONCENTRATIONS IN BIOLOGICAL ORGANISMS*
(mg lead/kg dry tissue)
Specimen
Fish
Attached Algae
Crayfish
Higher Aquatic Plants
Sediment
Ron-Urban
Station
1
<40
<20
14
<20
28
Reach
Station Station
2 3
_
<30
-
<30
37
_
<30
<30
<30
16
Urban Reach
Station
4
<30
200
29
<30
37
Station
5
<40
170
<36
<50
370
Station
' 6
<50
70
40
60
400
*The lead concentration in the urban sections of Coyote Creek during storms
averaged about 0.4 mg/1. Dry weather and lead concentrations from non-urban
areas are expected to be much less.
Source: Pitt and Bozeman 1979
Table 15. ZINC CONCENTRATIONS IN BIOLOGICAL ORGANISMS*
(mg zinc/kg dry tissue)
Specimen
Fish
Attached Algae
Crayfish
Higher Aquatic Plants
Sediment
Non-Urban Reach
Station Station
1 2
135
6.
80
9
70
" -
5 24
-
78
70
Station
3
' -'
17
90
26
14
Urban Reach
Station Station
4 5
100
160
89
40
30
120
135
140
150
120
Station
6
130
69
62
210
70
*The zinc concentration in the urban sections of Coyote Creek during storms
averaged about 0.2 mg/1. Dry weather and zinc concentrations from non-urban
areas are expected to be much less.
Source: Pitt and Bozeman 1979
397
-------
natural substrates such as logs, rocks, etc. Quantitative collections were
made with artificial substrates (diatometers equipped with glass slides) that
were suspended in the water column. Qualitative samples were preserved in a
5-percent buffered formalin solution for later identification. Diatometer
samples were scraped, cleaned with 30-percent hydrogen peroxide and potassium
permanganate, identified, and counted.
Fish. Seine collections from the non-urban sections of the study area indi-
cated the presence of 12 species of fish, half of which are native to the
Coyote Creek system. Similar collections in the urban sections of the study
area yielded only one native and three introduced fish species. As seen in
Table 16, the non-urban section of the stream supports a comparatively di-
verse assemblage of fish that includes native species such as the California
roach, hitch, Sacramento blackfish, Sacramento sucker, threespine stickle-
back, and prickly sculpin. Collectively, those species comprised over 60
percent of the 366 fish collected from the upper reaches of the study area.
In contrast, hitch", the only native fish collected from the urban sections of
the study area, represented less than ,1 percent of the,, 1124 fish captured in
the lower section of Coyote Creek. Hitch generally exhibit a preference for
quiet water habitat and are characteristic of warm, low elevation lakes,
sloughs, sluggish rivers, and ponds (Calhoun 1966; Moyle 1976). In streams
of the San Joaquin River system in the Sierra Nevada foothills of central
California, Moyle and Nichols (1973) found hitch to be most abundant in warm,
sandy-bottomed streams with large pools where introduced species such as
green sunfish, largemouth bass, and mosquitofish were common., Similarly, in
the lower portions of Coyote Creek hitch were found to be associated with
green sunfish, fathead minnows, and mosquitofish. However, mosquitofish
completely dominated the collections from the urban sections of the creek,
since they represented over 98 percent of the total number of fish captured.
Mosquitofish are particularly well adapted to extreme environmental condi-
tions including those imposed by stagnant waters with low dissolved oxygen
concentrations and elevated temperatures. In the foothill streams of the
Sierra Nevada, Moyle and Nichols (1973) found mosquitofish to be most abun-
dant in disturbed portions of intermittent streams, especially in warm,
turbid pools.
Benthic Organisms. The taxonomic composition and relative abundance of ben-
thic macroinvertebrates collected from both natural and artificial substrates
in Coyote Creek are presented in Tables 17 and 18. The benthos in the upper
reaches of the creek consists primarily of immature dipt'erans (midges and
blackflies) as well as certain clean water taxa such as mayflies and caddis-
flies. The benthos of the lower reaches of the creek is dominated exclusive-
ly by pollution-tolerant oligochaete worms (tubificids).
In general, the abundance and diversity of taxa appear to be greatest
in the non-urban sections of Coyote Creek. Figure 5 shows the trend of the
overall decrease in the total number of benthic taxa encountered in the urban
sections of the study area.
398
-------
Table 16. TAXONOMIC COMPOSITION AND RELATIVE ABUNDANCE OF FISH .COLLECTED
IN SEINE SAMPLES FROM COYOTE CREEK DURING FALL 1977 AND
SPRING 1978
Species
Cyprinidae - minnows and carps
California roach
(Hesperoleucus symmetricus)
Hitch
(Lavinia exilicauda)
Sacramento blackfish
(Orthodon microlepidotus)
Fathead minnow
(Pimephales promelas)
Catostomidae - suckers
Sacramento sucker
(Catostomus occidentalis)
Poeciliidae - live bearers
Mosquitofish
(Gambusia affinis)
Gasterosteidae - sticklebacks
Threespine stickleback
(Gasterosteus aculeatus)
Centrarchidae - sunf ish •
Green sunfish
(Lepomis cyanellus)
Bluegill ., ^
(Lepomis macrochirus)
Largemouth bass
(Micropterus salmoides)
Black crappie
(Pomoxis nlgromaculatus )
Cottidae - sculpins .
Prickly sculpin
(Cottus asper)
Total Number of Fish Collected
Non-Urban
Relative
Abundance
(Z)
2.7
3.6
2.2
3.8
18.3
20.5
: 33.1
9.0
4.4
0.5
0.5
1.4
Reach
Length
Range
(mm)
32 to 100
78 to 292
160 to 340
27 to 66
30 to 406
18 to 52
34 to 50
36 to 123
35 to 96
89 to 350
64 to 219
38 to 90
366
Urban Reach
Relative Length
Abundance Range
(Z) (mm)
0.4 48 to 142
- 1.2 35 to 65
.
98.3 15 to 52
- . ' •
0. 1 30
-
1124
Source: Pitt and Bozeman 1979
399
-------
Table 17. TAXONOMIC COMPOSITION AND RELATIVE ABUNDANCE OF BENTHIC MACRO-
INVERTEBRATES COLLECTED IN THE NON-URBAN REACH OF COYOTE CREEK
DURING SPRING OF 1978
Taxon
Ollgochaeta**
lUrudlnc*
Crustacea
Aaphipoda
Talltrldae
Hyalella azteca
Insccta
Ephctteroptera
Baetidae
Baetis sp.
Centroptlllua ap.
Epheaercllidae
Epheaerella sp.
Leptophlebildae
Habrophleblodet ap.
ttcalpcera
Corixidac
Colcoptera
Dyciceldae
Trichoptera
Hydropsychldac
Cheuma topsyche sp.
Hydroptllldae
Dlptcra
Ceratopogonldae
Chlroncatdae
Eapldldae
Huicldae
SlBUliidae
Tabaoldae
Tlpulldae
Gastropoda
Lyraaeldae
Lyanaea sp.
Phyiidae
Physa sp.
Planorbidac
Proaenetua ap.
Pclccypoda
Sphaectldae
Pltidtuji ap.
Total Huaber of Organlsss/B
Relative Abundance (Z) of Each Taxon Within The Sample
STATION 1
Ekman Surber Artificial*
Dredge Sampler Substrate
48.7 23.2 2.3
0.7
_ _ _
10.8 0.9
- - —
0.6
4.5
_ _ _
— - —
- - -
0.8
0.4
46.8 2.7 36.0
0.4
_ _ —
56.4 59.9
_ _ -
0.8
0.4
0.8
0.7 - 0.3
1.9
5836 602 1428
STATION 2
Ekman Surber Artificial
Dredge Sampler Substrate
89.8 0.4 1.4
2.2 0.4
• - 1.2,
4.9 5.2
1.2
6.1 3.8
- - .
1.7
0.4 0.8
- - -
- - -
- - -
0.7 2.4 18.7
_ _ _
0.7 2.9
70.2 27.9
0.7
*" ~ ~
0.7 0.4
1.5 3.3 30.1
3.7 4.5 12.1
— •" ~
2952 1323 555
STATION 3
Ekman Surber Artificial
Dredge Sampler Substrate
45.3
_ _
_ _ _
4.7
23.5
15.1
_ _
1.1
4.7
4.7
- - -
- - -
93.6 15.1
1.1
_ _ _
10.4 76.5
4.2
_ _ _
_ — —
- - -
- - -
_ — «
w * —
2046 106 17
Source: Pitt and Bozeaao 1979
*Method of Collection at each location.
**Thc Majority of these worms belonged to the family Luabrlculldae.
400
-------
Table 18. TAXONOMIC COMPOSITION AND RELATIVE ABUNDANCE OF BENTHIC MACRO-
INVERTEBRATES COLLECTED IN THE URBAN REACH OF COYOTE CREEK DURING
SPRING OF 1978
Taxon
Oligochaeta**
Hirudlnea
Crustacea
Amphipoda
Talitridae
Hyalella azteca
Insecta
Ephemeroptera
Baetidae
Baetls sp.
Centroptllium sp.
Ephemerellidae
Ephemerella sp.
Leptophlebildae
Habrophleblodes sp.
Hemiptera
Corixidae
Coleoptera
Dytiscidae
Trichoptera
Hydropsychidae
Cheumatopsyche sp.
Hydroptilidae
Diptera
Ceratopogonldae
Chironomidae
Empldidae
Muscidae
Simuliidae
Tabanidae
Tipulidae
Gastropoda
Lymnaeidae
Lymnaea sp.
Physidae
Physa sp.
Planorbidae
Promenetus sp.
Pelecypoda
Sphaeriidae
Pisidlun sp.
Total Number of Organisms/m
Relative Abundance (Z) of Each Taxon Within The Sample
STATION It
Eknan Surber Artificial*
Dredge Sampler Substrate
100.0 99.1 79.7
- - _
_ _ _
-
-
_ _ _
- - -
_ _ _
_ _ _
- - -
- - _
- - -
15.5
_ _ _
_ _ _
0.1 4.8
_ _ _
_ _ _
~ _ —
- _ _
_ _ _
_ _ -
- 0.8
926 3432 84
STATION 5
Ekman Surber Artificial
Dredge Sampler Substrate
100.0 94.5 54.3
- - -
_ - -
_ _ _
- - -
_ _ _
- - -
_ - -
- - -
-
- - —
_ _ _
- _ _
5.5 45.7
_ _ _
_ _ _
_ _ _
- _ _
— — ~
_ _ _
_ _ _
_ _ _
_ _ _
1335 290 138
STATION 6
Ekman Surber Artificial
Dredge Sampler Substrate
100.0 99.5 34.9
_ - -
- - -
_ _ _
_ _ _
_ _
- - -
_ _ _
- - -
- - -
_ - -
_ _ _
0.5 65.1
_ _ _
_ _ _
_ - -
_ _ _
_ — —
- - - -
- - -
_ _ _
_ _ _
1787 3362 83
Source: Pitt and Bozenan 1979
"Method of Collection at each location.
**The majority of these worms belonged to the family Tubificidae.
401
-------
15-
o
I
H
iu
CD
U_
O
tr
HI
m
S
z
10-
5-
Urbanized -" »- Non-urbanized
65 4
STATIONS (relative locations)
Source: Pitt and Bozeman 1979
Figure 5. Abundance of benthic taxa collected from natural and
artificial substrates in Coyote Creek during Spring of 1978.
402
-------
Attached Algae. Qualitative samples from natural substrates indicated that
the filamentous alga, Cladophora sp., was found throughout the study area.
However, growth of this alga reached its greatest proportions in the upper
sections of Coyote Creek. The taxonomic composition and relative abundance
of diatoms collected from artificial substrates placed at each sample loca-
tion are presented in Table 19. The periphyton in non—urban sections of the
creek was dominated by the genera Cocconeis and Achnanthes. The genera
Nitzschia and Navicula, generally accepted to be more pollution—tolerant
forms, dominated the periphyton in urban reaches of Coyote Creek.
These preliminary biological investigations in Coyote Creek have indi-
cated distinct differences in the taxonomic composition and relative abun-
dance of the aquatic biota present in various reaches of the stream. The
non-urban sections of the creek have been found to support a comparatively
diverse assemblage of aquatic organisms, including 12 species of fish and
various benthic macroinvertebrate taxa such as mayflies, caddisflies, aquatic
beetles, midges, blackflies, snails, and fingernail clams. In contrast, how-
ever, the urban portions of the creek have been shown to comprise an aquatic
community that is generally lacking in diversity and is dominated by
pollution-tolerant organisms such as mosquitofish and tubificid worms.
403
-------
Table 19. TAXONOMIC COMPOSITION AND RELATIVE ABUNDANCE OF DIATOMS COLLECTED
ON GLASS SLIDES IN COYOTE CREEK DURING THE SPRING OF 1978
Tucon
CcQtralcs
Coicinodiicaccae
Meloslra »pp.
Pennales
Dlatoaaceae
Diatoma vulgare
Fragllariaceae
Synedra «p.
Aehnanthaceae
Achnanthes laneeolata
Rhoicosphenia curvata
Cocconelc pedlculus
Cocconeis placentula
Kavlculaceae
Havicula «pp.
Dlplonels sp.
Fruttulia rhoaboldes
Cyrogigaa sp.
CoaphoneBataceae
Coaphonema »pp.
Cyabcllaccae
Cyabella spp.
Khopalodla cpp.
Ktczachiaceae
KitztchU app.
Benticula elegans
Surlrellaceae
Cysatopleura *olea
Surlrella app.
total Ntabec of Fru»tulcs/im2
Relative Abundance (Z) of Each Taxon Within the Sample
Kon-Urban Reach
Station Station Station
123
0.4
0.4 - 1.5 -
20.6 37.8 56.1
0.4 - -
15.0 18.2 0.4
62.4 44.0 41.2
0.8
0.8
5545 4950 1874
Urban Reach
Station Station
4 5
1.2
0.8 0.9
49.8 0.9
1.2
10.5
2.4
0.4
2.8 6.9
2.0
43.4 67.5
2.4
0.9
2.0 4.0
4488 1189
Station
6
0.8
0.4
1.6
23.8
0.4
0.8
0.4
0.4
70.6
0.4
0.4
4575
Source: Pitt and Bozeaan 1979
404
-------
REFERENCES
Calhoun, A.C. , Ed. 1966. Inland Fisheries Management, California Depart-
ment of Fish and Game. Sacramento, California. 546 pp.
Farmer, J.G., and T.D.B. Lyon. 1977. "Leads in Glasgow Street Dirt and
Soil." The Science of the Total Environment, 8:89-93.
McKee, J., and H.U. Wolf. 1963. Water Quality Criteria, 2nd ed.: State
Water Quality Control Board. Sacramento, California.
McMullen, T.B., and R.B. Faoro. 1977. "Occurrence of Eleven Metals in Air-
borne Particulates and Surficial Materials." Journal Air Pollution
Control Association. 27:12:1198-1202. December.
Metcalf and Eddy Engineers. 1978. "Surface Water Management Plan for Santa
Clara County - Technical Appendices" for Santa Clar-a-Valley Water Dis-
trict. San Jose, California. December.
Moyle, P.B., and R.D. Nichols. 1973. "Ecology of Some Native and Intro-
duced Fishes of the Sierra Nevada Foothills in Central California."
Copeia, 1973(3):478-490.
Moyle, P.B. 1976. Inland Fishes of California. University of California
Press. Berkeley, California. 405 pp.
Olson, K.W., and R.K. Skogerboe. 1975. "Identification of Soil Lead Com-
pounds from Automotive Sources." Environmental Science and Technology,
9:3:227-230. March.
Pitt, R.E. , and G. Amy. 1973.. Toxic Materials Analyses of Street Surface
Contaminants. EPA-R2-73-283. U.S. Environmental Protection Agency.
Washington, D.C., August.
Pitt, R.E. 1979. Demonstration of Non-Point Pollution Abatement through
Improved Street Cleaning Practices. EPA Grant No. S-804432.
Pitt, R.E., and M. Bozeman. 1979. Water Quality and Biological Effects of
Urban Runoff on Coyote Creek - First Phase. EPA Grant No. R805418010.
U.S. Environmental Protection Agency. (undated). 1975 .Interim Primary
Drinking Water Standards:; Subchaipter D, Part 141, Subpart A.
. 1973. Proposed Criteria for Water Quality: Vol. 1. October.
. 1973. Water Quality Criteria, Environmental Studies Board: NAS-NAE,
EPA-R3-73-033. March.
405
-------
r
Seventh Session
METHODOLOGIES FOR STORMWATER IMPACTS ASSESSMENT
Moderator: Wayne C. Huber
University of Florida
Gainesville, Florida
406
-------
NATIONWIDE ASSESSMENT OF URBAN
STORMWATER POLLUTION IMPACTS ON
RECEIVING WATER BODIES
by
James P. Heaney, Wayne C. Huber,
and Melvin E. Lehman
University of Florida, Gainesville, Florida 32611
ABSTRACT
Urban stormwater runoff has been recognized in recent years as a poten-
tial major contributor of pollution to receiving water bodies. Assessment of
urban stormwater runoff pollutant quantities and characteristics have been
made for several areas throughout the United States, the most ambitious being
the Environmental Protection Agency's 208 Areawide Wastewater Management
Planning Program. Price tags for abating urban stormwater pollution (through
elimination or reduction of discharges) range in the billions of dollars.
Projections of high costs have forced a look beyond abatement of discharges
to the receiving water bodies for insight as to what are the impacts, where
are they and are they significant?
First-year results of a nationwide search for documented case studies of
impacts of urban runoff receiving waters indicate that well-documented cases
are scarce. Impacts previously attributed to urban stormwater runoff may be
point source impacts in disguise, or they may be masked by greater contribu-
tions from other sources. In some cases they are offset by hydrological,
biological, or geological attributes of the receiving water body.
The lack of documentation and clear definition of urban stormwater im-
pacts makes the task of assessing importance of this pollution source even
more difficult. Efforts to address this aspect include relating sources of
pollutants and pollutant types to receiving water characteristics and effects
on desired water uses. Characteristics such as stream or lake bed hydraulics,
present and potential water uses, established stream standards, ecological
data and water quality information are being summarized for the documented
cases to determine how the urban runoff pollutants might behave or react in
the receiving water and what potential use they might affect most adversely.
Results of these analyses are to be used as a basis for devising simple
criteria for analyzing an urban area to determine whether a potential impact
does or would occur.
407
-------
INTRODUCTION
A previous nationwide assessment indicated that urban runoff can be
viewed as causing problems since, on a nationwide average, the quantity of
urban runoff (13.4 in/yr) is approximately equal to the quantity of sewage
(12.8 in/yr), and the annual BOD5 per acre from a sewage treatment plant
with a removal efficiency of 90 percent is 59.4 pounds as compared to
43.6 pounds from urban runoff (1). Furthermore, the cost of controlling
urban runoff appears to be competitive with the cost of additional removal of
pollutants in sewage. Consequently, if further reductions in pollutant loads
are needed, then urban runoff controls as well as further waste treatment
should be evaluated carefully. The anticipated high price tag for such
control programs has prompted decision makers to take a harder look at how
serious the problem really is. This paper presents the results of a year-
long effort to search published and unpublished literature, 201 and 208
project documents, EPA-furnished project materials, agency data and permit
files, and other miscellaneous data sources to characterize stormwater runoff
impacts on receiving waters.
IMPACTS DEFINED
Several interrelated views on impact assessment may be gleaned from a
review of the literature. Traditionally, two perspectives, public health and
sanitary engineering, were of prime importance. The public health approach
focused on prevention, whereas sanitary engineers took a cost-effectiveness
approach (2). An example from the turn-of-the-century is the controversy
over whether cities should be required to treat their waste to reduce down-
stream water treatment costs. Sanitary engineers argued that the assimila-
tive capacity of the rivers should be considered and that it is much more
cost-effective to treat the intake water than to spend larger sums (approxi-
mately ten times more) on upstream waste treatment. Cooperative efforts
between these two groups led to the present regulatory approach, wherein
effluent and/or receiving water standards are promulgated. Within this
context, "impacts" can be defined in terms of whether the "standards" have
been violated.
This approach prevailed until 1972 when the Federal Water Pollution
Control Act Amendments established the following basic water quality goals
and policies for the United States (3, 4):
1. The discharge of pollutants into navigable waters should be
eliminated by 1985.
2. Wherever attainable, an interim goal of water quality, which pro-
vides for the protection and propagation of fish, shellfish, and
wildlife and for recreation in and out of water, should be achieved
by July 1, 1983.
3. The discharge of pollutants in toxic amounts should be prohibited.
These amendments represented a shift towards the early public health philoso-
phy of anti-degradation with relatively little consideration being given to
the cost of attaining these goals.
408
-------
In this same general period, the National Environmental Policy Act
(NEPA) of 1969 ushered in a new field of environmental impact assessment.
Numerous methodologies were proposed by professionals from various
disciplines. Canter (4) outlines the following 12-step procedure for the
water environment.
1. Determine types and quantities of water pollutants emitted from all
alternatives for meeting a given need during both construction and
operational phases.
2. Determine the existing water quantity and quality levels for the
surface watercourses in the area. Examine the frequency distribu-
tions and the median and mean data for both water quantity and
quality. If possible, consider historical trends of water quality.
Note particularly the low flow utilized by the local regulatory
agency for maintenance of water quality standards.
3. Document unique pollution problems that have occurred or are
existing in local surface watercourses.
4. If relevant for the project alternatives, describe groundwater
quantity and quality in the area, noting the depth of the ground-
water table and direction of groundwater flow. Identify major local
uses of groundwater, and delineate historical trends for groundwater
depletion and pollution.
5.
6.
7.
8.
Assemble summary of key meteorological parameters for the area,
noting particularly the monthly averages of precipitation, evapora-
tion, and temperature.
Procure the applicable water quality standards for local surface
watercourses and groundwater supplies if relevant. Specify appli-
cability of effluent: standards and required treatment technology and
state whether the receiving stream is water-quality limited or
effluent limited. Consider time schedules required for attaining
applicable water quality standards.
Summarize the organic waste load allocation study for the area.
Also procure extent information on inorganic, thermal, sediment, and
bacterial waste loads. Identify known point sources of pollution,
focusing specifically on unique discharges or wastewater consti-
tuents. Also enumerate the types of water uses in the area and
summarize the quantities involved.
Determine the mesoscale impacts by calculating estimated daily
quantities of water pollutants from the alternatives during both
construction and operational phases and comparing these to existing
409
-------
waste loads in the drainage area. Determine the percentage increase
in these waste Ib-ads. Note existing xater quality parameters that
are good or poor relative to current or potential standards.
9. Consider construction phase impacts in terms of the following
factors:
a. Time period of construction and the resultant time period of
decreased water quality. Specify stream discharges and quality
variations that would be anticipated during the construction
phase.
b. Anticipated distance downstream of decreased water quality.
c. Implications of decreased water quality relative to downstream
water users. If there are users that require certain water
qualities, identify the required raw water quality characteris-
tics and discuss the effects of decreased quality during the
construction phase.
d. Specific construction specifications directed toward pollutant
minimization.
10. Determine the microscale impacts by calculating specific downstream
concentrations resulting from conservative pollutants, dissolved
oxygen concentrations resulting from nonconservative (organic)
pollutants, and temperatures resulting from thermal discharges.
Consider these microscale impacts for both construction and opera-
tional phases. Compare calculated downstream concentrations with
applicable water quality standards. Check if applicable effluent
standards are met for existing facilities, or consider how they will
be brought into compliance. In the case of new sources, identify
necessary technology for compliance with new source performance
standards.
11. If water quality or effluent standards are exceeded, consider miti-
gation or control measures.
12. Consider operational impacts of alternatives in terms of the
following factors:
a. Frequency distribution of decreased quality and quantity.
b. Effects of sedimentation on the stream bottom ecosystem.
c. Fate of nutrients by incorporation into biomass.
d. Reconcentration of metals, pesticides, or radionuclides into the
food web.
410
-------
e. : Chemical precipitation or oxidation/reduction of inorganic
chemicals.
f. Anticipated distance downstream of decreased water quality and
the implications for water users and related raw water quality
requirements.
g. General effects of any water quality changes on the stream
ecosystem. •
h. Unique water quality changes that occur as a result of water
impoundment and thermal stratification.
This method goes well beyond the strict determination of whether stan-
dards have been violated. The output from this assessment is sometimes
summarized in terms of a numerical quality index [see Dunette (5) or Ott (6)
for a summary of the literature on this subject].
Ecologists provided yet another perspective on impact assessment. They
called for a more holistic view of the problem and a recognition that mea-
sures of value other than man's are relevant. Ehrenfeld (7) summarizes the
viewpoints on how "non-resources" can be valued in an appropriate manner. He
lists nine categories of "economic" values:
1. recreation and aesthetic-values,
2. undiscovered or undeveloped, " '
3. ecosystem stabilization, ~~ - - • '
4. examples of survival, • •
5. environmental baseline and monitoring values, : '• •-• ''
6. scientific research values,
7. teaching values, • : ' " '
8. habitat reconstruction values, arid
9. conservation values: avoidance of irreversible change.
In addition, he lists two "non-economic" values:
1. natural art value, and
2. Keligious value.
411
-------
Lastly, economists have contributed a significant body of literature on
this subject. Freeman (8) summarizes the categories of impacts according to
the channels through which environmental changes affect man.
1. Through living systems—biological mechanisms
Human health—morbidity, mortality
Economic productivity of ecological systems—agricultural
productivity, forestry, fisheries.
2. Other ecosystem impacts
Recreational uses of ecosystems—fishing, hunting
Ecological changes—diversity, stability
3. Through nonliving systems
Materials damages, soiling, production costs
Weather, climate
Other—odor, visibility, visual aesthetics.
To avoid the accusation of parochialism in adopting, a priori, any one
or combination of the above systems for assessing stormwater "impacts," the
literature search was approached with an open mind. However, the need for a
more precise definition of an "impact" became apparent early into the litera-
ture search. Definitions of "impact" are almost as numerous as there are
investigators, congressmen, regulators, and citizen review committees of
urban stormwater problems. The range of impact definitions includes specific
cause-effect statements, comparisons of constituent concentrations to
numerical standards, sensory perceptions such as odor and color problems, and
"perceived" impacts from citizens. All are applicable and valued, with
respect to the level of action or understanding desired. However, for a
"standard" definition by which to conduct comparative studies at the environ-
mental regulatory agency (EPA) level, a stormwater impact was defined as one
which resulted in "loss of beneficial use." Beneficial uses considered are
those listed in local, state, and federal water laws, which include drinking
water use, fishing and she11 fishing, swimming, boating, manufacturing process
water use, etc. The list of beneficial uses could be extended to include
loss of a receiving stream to wastewater disposal as a result of already
overloaded assimilative capabilities or as loss of biological integrity, but
these uses are not clearly defined in many water laws.
This definition states that an impact occurs when beneficial use is lost
as a result of stormwater runoff. The key question for determining
documentation of impacts is at what point does loss of beneficial use
actually occur? Does loss occur when a. receiving water body does not meet a
governmental water use classification, when constituent levels violate water
quality standards (or criteria), when 30-year population counts show
reductions in fish populations, or when people get sick or fish kills occur?
When is loss of use implied, when is it imminent, and when is it obvious?
The following summary relates "impact level" to "loss of use" in a general
overview:
412
-------
Impact Level
Policy or Management Planning
Standards or Criteria Violations
Documented Cases of Cause-Effect
Loss of Use
Is considered possible or is
implied.
Is implied, may be imminent, or
can actually occur through
restriction of use.
Actually occurs.
Policy or Management Planning—At this level the loss of beneficial use is
implied but has not actually occurred. Use of a receiving water body for
stormwater discharge may violate (or be in contrast to) a comprehensive plan,
environmental agency philosophy, coastal zone management policy, area-wide
water use classification system, or some other indicator of intended use.
Standards or Criteria Violations—This is the level at which impact typically
has been assessed. The usual approach is to measure constituents of storm or
receiving waters, compare measured values with local, state, or federal stan-
dards (criteria, or guidelines), and then directly equate impact with the
number of constituent standards violated, or the number of days a constituent
standard is violated. This same approach is used in cases where key concen-
trations have not been identified or developed. The presence or absence
(occurrence) of a constituent (e.g., EPA list of 129 priority pollutants) is
often considered an impact.
The loss of use at this level can be implied, may be imminent, or can
actually occur. Whether or not an impact actually occurs is relative to the
basis on which a standard is promulgated. In the absence of supporting data,
standards are usually set conservatively, so many documented violations of
standards imply impact, rather than actually indicate impact (e.g., oxygen
standards may be violated, but fish kills may not occur). Standards or
criteria violations may be considered "paper" impacts.
Documented Cases of Cause-Effeet—This was considered the impact level where
loss of beneficial use was actually documented. Site-specific evidence of
fish kills, beach closings., loss of water supply, trends in aquatic systems
deterioration, citizens' complaints of odor, floating debris, medical records
of water-use related diseases, and other effects (impacts) caused by or
related to stormwater runoff were reviewed. The nationwide survey focused on
this level.
Documentation of extreme or unusual events is often easier than identi-
fication of trends or subtle changes, especially where complex systems of man
and nature are concerned. Questions arise as to the cause of an impact (such
as a fish kill). Was it due to an event such as a toxic substance'being
flushed, or was it due to bioaccumulation (to lethal levels) of a substance
over 30 years? Thus, the cause of an impact may be an event, or a series of
events establishing a trend. In general, pulsed (event-driven) systems are
under examination when addressing stormwater issues, so causes of impacts
413
-------
I
considered were short-term. The impact itself,, being event- or trend-induced
can be manifested quickly or over a long period, and finally, the manifested
impact may have short or long-term significance. This study attempted to
categorize impacts in terms of:
Event—type - short-term impacts such as fish kills and temporary beach
closings.
Trend type — long-term impacts such as evidence of biological deteriora-
tion or loss of species diversity.
SEARCH FOR IMPACTS
Literature Riview
A systematic category-by-category search and review of literature
sources was conducted. Major sources include:
1. EPA Cincinnati in-house files on impacts of urban runoff (9).
2. Review of 208 studies throughout the United States. As part of
EPA's National Urban Runoff Program (NURP), EPA Headquarters person-
nel and a team of consultants visited every regional office at least
once. During the initial visit, the regional personnel were asked
to identify which of these 208 studies indicated that urban runoff
was a "problem." Next, the group was asked which areas had receiving
water data documenting that an impact existed. They were asked to
indicate whether the "problem" re.lated to violation of water quality
standards, impairment of beneficial use(s), aesthetics, or other
cause(s). Lastly, each regional group was asked to suggest candi-
date cities for further study. Ideally, these cities should have a
clearly identified urban runoff problem, and sufficient interest in
solving it to finance 25 percent of the cost of the study. Based on
this procedure, 30 cities were selected for further study.
3. Computerized Literature Searches
a. Water Resources Scientific Information Catalog (WRSIC) system of
U.S. Department of Interior, Office of Water Research and
Technology.
b. Smithsonian Scientific Information Exchange (SSIE) lists of
on-going research projects.
c. University of Florida's State Technologies Application Center
(STAC) information retrieval system which is tied into the NASA
system of about 20 million publications.
4. U.S. Environmental Protection Agency Fish-Kill Data. Approximately
10,000 individual fish-kill reports were surveyed [see (10) for a
general summary].
414
-------
5. Studies for the National Commission on Water Quality on beach
closings (11) . '•-
6. 1978 Needs Survey (12).
7. EPA National Urban Runoff Program files= of case studies of
30 cities.
8. Sutron Corp. case studies on relationships of rainfall and dissolved
oxygen (13).
9. Abstracts for EPA National Conference titled Urban Stormwater and
Combined Sewer Overflow—Impact on Receiving Waters, November 1979.
10. Environmental Protection Agency, Nationwide 201 and 208 Technical
Documents and Wastewater Management Plans.
11. National Eutrophication Survey Documents (14).
12. 1974 EPA National Water Quality Inventory (15).
The literature survey was the'main tool used to identify case study
areas. Other sources included review of newspaper articles, attendance at
scientific and technical conferences, telephone conferences with regional
Environmental Protection Agency offices, and conversations with other inves-
tigators in the field.
Handling of Receiving Water"Informat ion
The documents from the preliminary screening as a result of the litera-
ture review were catalogued and incorporated into the existing EPA Urban
Stormwater Runoff Data Base geographical files at the University of Florida.
The information is summarized,: by state, for-each of the 248:urbanized areas
listed in (1). Case studies for smaller cilfies are included in'the
general information regarding each state. For each urbanized area, the
following information is presented: , t
Demographic Data:
—1970 population and urbanized area acreage from (1)
—Percent combined sewers from (12)
Hydrologic Background:
—General information regarding the water resources of the area. Most
of this information was taken from a report by Schneider (16).
Waste Sources:
—The annual volume of sewage and urban runoff from the area. The
data are taken from (1).
415
-------
Receiving Waters:
—The primary receiving water is indicated. Minor receiving waters or
special study sites are also mentioned. Where appropriate, a dilu-
tion ratio is calculated as follows:
Dilution Ratio
Primary Receiving Water Flow (in/yr)
Sewage flow (in/yr) +• Urban Runoff (in/yr)
All values are in inches per year over the developed urbanized
area.
Specific Studies:
—Summaries of receiving water impact studies.
Maps:
-For each city, a 1:500,000 map was drawn from the U.S.G.S. State
Hydrologic Maps. The same scale was used for all cities. Hydro-
logic boundaries are indicated. Receiving waters are shown exactly
as indicated on the original map. The name of the receiving water
is included only if it was included on the original map. This
provides the user with a better perspective on the relative
importance of the local receiving waters.
Latitude and longitude are included.
U.S.G.S. sampling stations is shown.
Streamflow Data:
The location of significant
—The 1950-1960 summary of U.S.G.S. monthly flow data are included.
This summary includes the long-term average flow and extreme
values.
A sample result for Tampa, Florida is shown in Appendix A.
RESULTS
Types of Receiving Waters
Table 1 shows the percentage of urbanized areas that discharge into each
of 15 categories of receiving waters. The results indicate that about
75 percent of the discharge is to rivers, 5 percent to lakes, and 20 percent
to estuaries or oceans. However, this distribution can be misleading because
many of the impacted areas of major significance are not the primary
receiving water.
416
-------
TABLE 1. DISTRIBUTION OF PRIMARY RECEIVING WATERS FOR
URBANIZED AREAS IN THE UNITED STATES
Category
Percent of
Total
A. Rivers
1. Creeks and shallow streams [depth (d) <2 feet] 8.8
2. Upstream feeders (2_30)
4. Open ocean or beach
Sub-Total, Estuaries and Oceans 20.1
Total, A,B, and C 100.0
*Source: Raw Data From 1978 NEEDS Survey (12).
417
-------
Documentation of Impacts
Fish Kills-
Urban receiving water bodies, for a variety of reasons other than
stormwater runoff (e.g., habitat alteration, multi-source pollutants,
temperature), may not have abundant fish populations. In fact, abundance of
aquatic life in general is most often inversely proportional to the degree of
urbanization. However, many fish kills have been related to stormwater
runoff, combined sewer overflows (CSO's), storm sewers, or rainfall events.
Table 2 is a summary of storm-water/storm-sewer related fish kill reports in
the EPA files (17) for the period from January 1970 to May 1979. It gives
the number of fish kills reported for a given year relative to the seven
categories listed. In order of appearance in the table, a brief description
of each category is presented.
Pollutant Spilled into Storm Sewer—The pollutant in this category may
or may not have been flushed by a rainfall event (not indicated on fish kill
report) . Typically this is a human event-type impact where someone dumps a
toxic substance into a storm drain, or a toxic substance accidentally spills
and drains into a storm sewer.
Storm Sewer Discharge from Rainfall Event—This is distinguished from
the previous category in that the person filling out the fish-kill report
stated that the kill was due to discharge during a storm event. This
category overlaps the previous one in instances where residuals of a spilled
toxic substance remain in the storm sewer until adequate flushing occurs.;
Combined Sewer Outfall—Fish kills occur downstream of identified :
combined sewer outfalls (CSO's), but are not necessarily storm-event
related.
Rainfall-Runoff Related to Land Use—This fish kill category includes
reports which do not specify drainage ways, storm sewers, or sheet flow but
document fish kills following rainfall-runoff events. The attempt to
classify relative to land use is a very rough estimate. Fish-kill reports
are not detailed enough to identify the specific sources of runoff, or how
the runoff actually gets into the receiving water body.
Acid Mine Drainage-Storm Event Related—Acid mine drainage, a nonurban
stormwater problem, was included in the table for two reasons: the mining
activity may be borrow material for roads, or a town built around a mine; and
it is useful to compare the number of acid mine related fish kills to those
related to stormwater/storm sewers. Acid mine drainage is a fairly well
documented problem; the numbers of reported kills are comparable.
Landfill Leachate Storm-Event Related—Most landfills are products of
urban activity, even though many are located in agricultural or fringe
areas.
Other—This category includes fish-kill reports not suitable under the
other headings but of interest due to location (a receiving water body under
418
-------
s
i
I
I
CM
a
1
s
CO
•g -
•H C
.1 3
•y -p
S 9
'O
CO
8 &
a
•r-f O
Pt JJ
W CO
CO f^ i-l «*• I
CM .-I I CO
co CM 1/1 m CM I I CM i i
-------
stormwater runoff study, e.g. Trinity River, Texas), or apparent storm-water-
related kill (indicated, not directly stated).
Some general comments pertinent to the survey of the fish-kill data
are:
1. Because many urban streams are grossly polluted from a variety of
sources, no fish remain. Hence, no fish kills are reported.
2. Very few of the reports (a total of 20 for the period 1970-1979)
state storm-water runoff directly as the cause of fish kills.
3. There are several instances of stormwater flushing pesticides,
sewage deposits, herbicides, dumps of oil, etc.
4. There are several instances of accidental dumps into storm sewers of
toxic substances and then flushed prematurely of rainfall events—a
good example is a hotel or drug store fire where fire department
runoff goes into a storm drain.
5. Many citations indicate fish kills due to "eutrophication," or
"natural causes." These were not included in Table 2. General
water quality deterioration or nutrient enrichment, could in part,
be due to storm water runoff.
6. The EPA instructions for filling out fish-kill reports group
sewerage, storm water, and CSO's into a single category (sewerage
system). If the person filling out the form does not specify the
type of system, it is anyones "best guess" as to which type of '"
system caused the kill. The Sewerage System category contains the
largest number of fish kills (10). Unless CSO's or storm sewers
were specifically cited, they were not counted and recorded in
Table 2. For this reason, Table 2 probably understates the actual
number of CSO or storm sewer-related kills.
7. In a true urban stream, the resident fish population may be adapted
to pollutant loads, and/or pulses; or the population may become
dominated by more pollutant tolerant species. The fish-kill data
do have some information concerning species of fish killed, but no
historical information is available to assess impacts of adaptive or
community changes.
8. Fish-kill frequency data are relative. If an area experiences a
severe kill that wipes out the fish population, subsequent events
that occur do not record kills. Conversely, a severe kill could
wipe out a resident population, but immigration from a nearby water
body would mask reduction in the local population on a longer term
basis.
9. Fish-kill reports are prepared by people with a variety of positions
and backgrounds. The inherent variability in such a nationwide
420
-------
reporting system makes numbers and statistics for this type of data
base very subjective and less reliable.
The monthly distribution of fish kills as a percentage of the total
number of fish killed and the total number of reports is shown in Figure 1.
As expected, relatively more kills occur during the warmer months of the
year.
Beach Closings—
The National Commission on Water Quality placed heavy emphasis on
attempting to evaluate the benefits associated with water pollution control
(11). A total of 3,521 beaches throughout the United States were surveyed.
Of these beaches, 449 had water quality problems. Table 3 shows the
proportion of the closings due to various causes. While urban runoff is not
identified as a separate category, the majority of the closings may be
attributable to this cause. For example, within the coliform related problem
category, almost 50 percent of the total closings are due to undefined sewage
contamination or unknown causes.
Low Dissolved Oxygen—
Keefer, Simons, and McQuivey (13), in an EPA sponsored national assess-
ment, related the magnitude of dissolved oxygen (D.O.) deficits and the
presence of storm runoff downstream of urban areas. Based on an initial
screening of over 1,000 D.O. monitors located throughout the United States,
over 100 water quality monitoring sites in and downstream of urban areas were
selected. Approximately one-third of these monitors indicated at least a
60 percent probability of a higher than average dissolved oxygen deficit
occurring at times of higher than average streamflowJ Of the areas where a
low D.O., high streamflow correlation was observed, a more detailed hourly
analysis was performed. These results were striking. During steady state
low flow conditions the D.O. fluctuates diurnally between 1 and 7 mg/1.
However, after a storm begins, the diurnal fluctuations are completely
dampened. The minimum wet-weather D.O. is 1 to 1.5 mg/1 lower than the
dry-weather minimum, and remains that way for 1 to 5 days. As the impact of
the storm dissipates, the D.O. resumes its original cyclic behavior. This
relationship, for the Scioto River at Chillicothe, Ohio, (downstream of
Columbus, Ohio) is shown in Figure 2. It suggests the need to reexamine the
traditional approach to defining "critical" conditions in receiving waters.
In this study (13), the two most severe cases of low D.O. were the
Trinity River near Dallas, Texas and Wilsons Creek near Springfield,
Missouri. The authors' independent analysis indicates that both of these
receiving waters have large deposits of sludge from primary sewage treatment
plants. Thus, a significant part of the problem is attributable to
resuspension of this benthal material.
National Urban Runoff Program Case Studies—
This detailed evaluation of urban runoff problems in cities throughout
the Unites States yielded 30 case studies. Each of these applicants
421
-------
<
-J
40-
30
20-
10
LEGEND
FISH KILLED
REPORTS
J F
A S 0 N D
MONTH
FIGUBE 1. MONTHLY DISTRIBUTION OF FISH KILLS AS A PERCENTAGE
OF TOTAL FISH KILLED
422
-------
TABLE 3. CAUSES OF WATER QUALITY RELATED BEACH CLOSINGS
IN THE UNITED STATES
Cause
Algae, Scum
Turbidity
High Coliform or Fecal Coliform Count Due To:
Flood, Wind, Heavy Rainfall
Agricultural Runoff
Sewage Treatment Plant Malfunction,
Spills, Outfall
Undefined Sewage Contamination or Unknown
Other
Unknown
Total
Number
11
13
41 '
10
42
337
71
80
605
Percent of
Total
1.8
2.2
6.8
1.7
6.9
55.7
11.7
13.2
100.0
*Source: Raw Data From Battelle (11).
423
-------
SATURATION
D.O.LEVEL-
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
3.0
2 I 20
I- *—
< 2
K g
W £
1 E '-°
o °
S2 n»
Q a.
= ^ O 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
TIME (day of month )
(I)DAILY DISCHARGE/AV6. ANNUAL DISCHARGE-
(2) PRECIPITATION (inches)
FIGURE 2. OBSERVED RELATIONSHIP BETWEEN DIURNAL DISSOLVED OXYGEN
CONCENTRATION AND STORM EVENTS FOR THE SCIOTO RIVER AT
CHILLICOTHE, OHIO
424
-------
indicated that they have a "problem" and that they were willing to do some-
thing about it. However, little definitive evidence was presented to support
the contention of a receiving water problem. Summary information regarding
these projects is shown in Table 4. Descriptions of the receiving water
problem for some of these projects are presented below for illustrative
purposes.
1. Baltimore, Maryland—Studies by Olivieri and his co-workers (18)
appear to be the best available source of information on the
bacteriological quality of urban stormwater. Their results indicate
that "urban runoff" is really a composite of all unaccounted for
residuals leaving an urban area via the watercourses. It includes
illicit industrial waste, cross connections with the sanitary sewer
system, septic tank seepage, landfill leachate, etc. Thus, sanitary
surveys of local receiving waters are needed to characterize the
actual problems that exist.
2. Myrtle Beach, South Carolina—;Bacteriological contamination of the
city's beaches occurs after heavy rains. Urban runoff is the
alleged cause although it might be illicit interconnections with the
sanitary sewer system.
3. Tampa, Florida—Deterioration of the City of Tampa surface water
supply in the Upper Hillsborough River appears to be partially
attributable to urbanization of the riparian lands. Water quality
in the Lower Hillsborough River and Hillsborough Bay is degraded by
landfill leachate, contaminated spring water, sanitary sewer system
overflows, zoo runoff, as well as the more general forms of urban
runoff. These receiving waters have been seriously impacted by
sewage and water treatment plant sludges, industrial wastes,
incinerator leachate, and other problems. The relative importance
of urban runoff is to be determined.
4. Milwaukee, Wisconsin—Milwaukee has very serious combined sewer
overflow problems and a large accumulation of sewage sludges. The
problem is relatively well documented and the City of Milwaukee is
committed to take remedial actions.
5. Austin, Texas—Town Lake, a water supply source for the city, has
received urban runoff for a number of years. Urbanization is pro-
ceeding upstream along the lake. There is general evidence that
water quality in the lake has deteriorated but the exact extent is
unknown. Are stringent controls on urban runoff needed and/or
should the City of Austin move its water supply further upstream?
6. Bellevue, Washington—The City of Bellevue seeks to preserve the
salmon runs. Efforts are being made to prevent deterioration of the
local streams. This requires effective ordinances and extensive
monitoring. What would such a program cost and how effective will
it be?
425
-------
TABLE 4
CHARACTERISTICS OF NATIONWIDE URBAN RUNOFF PROGRAM PROTOTYPE PROJECTS
»« Ian
I
11
111
It
V
VI
VII
Vlll
IX
X
llXtSOl
Applicant
Agency
Ceao. o( Mass
C«eo>. of Mass.
S.H. «ater Supply
U»|{ Island RFS
Kev York. DEC
Stu York DEC
I'AlMr.Jton COG
Balttaore RFC
ttaexnuu KrC
y, Carolina DR
Taepa BUT
Trl-County RfC
(C>.coe
stxcec
UFA ;
N'IfC
uin. DSH/SEunrc
City of Austin
Mtcroslan
Denver «COC
Sale LaVe County
6tfe District COC
Alaacda CFC 1 UCD
SCAC
City of Bellevue
Ui» County COC
*,«!.„* HSB
Recelvtnp Uater Type Impacts Beneficial Uses
£5
8 « ll ., e c"e *
Project ^S " * * ' 3* CM-HU ai'o.2£V£jr;5 *£
Location 3 « ^ « u 2 ^ "" t^ S tS t§ m^JiJow-S^ S
Lake Qulnslgaaond, HA • • 0 » 0 t 0
Mystic River, HA 0*0 t 0 §00
Durhaa, KH 0* • 0 0 0*
Veraottt
Long Ziland, NY • 0 0 • 0 0 • 0 0 •
Lake George, NY • • 0 • 0
IrondequoU Bay, NY • • 0 0 • 0 0
Hetro Uachlngton, DC 0000 900 tOOOO
Baltlaore, HD 00* 0*0 Ot
Myrtle Beach, SC • 0 • 0 •
Wins ton-Sal eta, HC 0 0 t •
Tampa, FL 000 0* OOOOO
Lansing, HI 0 0 0
Oakland County, HI 0 0 0 0
Ann Arbor, HI • o 0 0 0
Chaapalgn-Urbana, IL 0 • 00
Chicago, IL 00 6 0 0 0
Milwaukee, WI • 0 0 * 000
Austin, TX « 0 0 0 0 0
Little Rock. AR 0 • 00 0 0
Albuquerque, KH
Kansas City, HO ' "
Denver, CO 00* 00 0000
Salt Lake City, UT 0 • • 0 t 0
Rapid City, SD
Cantro Valley, CA t
Santa Ana, CA t • 0 00*0
Freano, CA
Bellevue, WA • 0 • 0 • 0
Eugene, OR 000 00 000
Portland, OR 0 o 000
* A»|r«t:ta of program receiving ujor cnpHante. For receiving waters, this means an emphartte
e»n ttonttorlng and Impact asBcsntaent. For B'IP'G, this mentis an emphasis on determining
•.oncrol effect 1 virnei* and co«t*
426
-------
TABLE 4
CHARACTERISTICS OF NATIONWIDE URBAN RUNOFF PROGRAM PROTOTYPE PROJECTS
(CONTINUED, PAGE 2 OF 2)
Applicant
ReRlon Agency
I Conm. of Mass
Coon, of Mass.
N.H. Water Supply
II Long Island RPB
New York DEC
Baltimore RPC
N. Carolina DR
Tampa DP'-'
SEMCOG
SEMCOG
IEPA
NIPC
Wise. DS'R/SEURPC
VI City of Austin
Hetroplan
VII
VIII Denver RCOC
6th District COG
IX Alameda CFC & WCD
SCAC
X City of Bellevue
Lane County COG
Portland MSD
P bl P 11
* t
. § JT I i s S- E Average Mf I. i* |||
pr°J"' D: £ I 55*2 Rainfall S fr S 1 "g 1 1 J I t
Locatlon S 5 5 £ S 5 5 ' "»'**> SI SI 33 5- IS I
Lake Qulnslgatnond , HA • 0 0 • ' 46
Mystic River, HA 00 0 • swirl cone.
Durham, NH OOO • 000
Vermont
Long Island, NY 0 0 * 44 0 • - • pena. seuer
Lake George, NY t 0 0 0 35
O4n
Baltimore, MD 0 0 0 __ 0 0 = _0 45
•
Ulnston-Salem, NC 000 0 44
Tampa, FL .00 0 49
Oakland County, HI 0 0 t runoff ordinance
Ann Arbor, MI '-000 • • •
Champa ign-Urbana, IL • • -
Chicago. IL 0 • 0 0 0 .34 t ' '-'•••
Milwaukee, UI .0000 . ... 31 »
Austin, TX 000 0 runoff ordinance
Little Rock, AR 000 0 50 '
Albuquerque, MM , . ,. . . .
Kansas City, MO . . „ .. . .
Denver, CO 0 0 • 0 16 0 e runoff ordinance
Rapid City, SD 0000 0 , .22 • ,
Castro Valley, CA
Santa Ana, CA ,0000 13 O ' oil/grease traps
Fresno, CA . . . . ,
Bellevue, WA 0000000 t*
Eugene, OR 0 0 0 O 0 " 0
Portland, OR 0 0 0 - O
427
-------
National Water Quality Inventory—
The National Water Quality Inventory (15) is a compilation of reference
level violations in major waterways. The causes of these violations were not
segregated so the contribution of stormwater is not known. The frequency of
violations is of interest in that many of the water bodies have been
identified relative to stormwater/receiving water studies.
Sulfates
Alkalinity
pH*
Dissolved oxygen
BOD5*
COD (.025N)*
Total Coliforms*
Fecal Coliforms*
Phenols
Odor
Parameters addressed were:
Suspended solids*
Turbidity*
Color*
Ammonia*
Nitrite
Nitrate (as N)
Nitrate (as ^3)*
Nitrite plus nitrate*
Organic nitrogen
Total Kjeldahl nitrogen
Total phosphorus*
Total phosphate
Dissolved phosphate
Dissolved solids (105°C)
Dissolved solids (180°C)
Chlorides
Table 5 shows the percentage of these parameters exceeding reference
levels. These receiving waters are relatively large and do not provide a
good representation of the mix of urban area receiving waters shown in
Table 1. However, seasonal flow analysis indicated higher pollutant levels
in periods of high flow for parameters denoted by asterisks. Thus,
wet-weather impacts are evident.
SUMMARY AND CONCLUSIONS
A study was undertaken to inventory documented receiving water impacts
from urban runoff on a nationwide basis. The search for documentation
included published and unpublished literature, Section 201 and Section 208
projects (PL 92-500), EPA-furnished project materials, EPA fish kill data
files, National Urban Runoff Program Proposals, and miscellaneous water
quality reports and permit files. Impacts due to stormwater runoff can be
defined at three levels: (1) policy or master plan violations; (2) criteria
or standards violations; and (3) actual environmental degradation or loss of
beneficial use, such as fish kills and beach closings.
Findings to date indicate that documented case studies of impacts of
urban runoff on receiving water are scarce. Several reasons may be given for
this situation.
428
-------
TABLE 5. MAJOR WATERWAY RANKINGS-PERCENT PARAMETERS
EXCEEDING REFERENCE LEVELS*
0 to 7
7 to 17
Over 17
Upper Missouri River
Columbia River
Lower Tennessee River
Snake River
Willamette River
Boston Harbor
Rio Grande River
Alabama-Coosa Rivers
Upper Ohio River
Susquehanna River
Upper Red River
Lower Colorado River
Upper Mississippi River Potomac River
Yukon River Detroit Area-Tributaries
Chicago Area-Lake Michigan Sacramento River
Upper Tennessee River
Detroit Area-River
Lower Red River
Brazos River
Upper Colorado River
Hudson River
Delaware River
Middle Mississippi
River
Lower Arkansas River
Lower Ohio River
Lower Mississippi
River
Middle Ohio River
Lower Missouri River
Chicago Area-
Tributaries
Mississippi River near
Minneapolis
Upper Arkansas River
Middle Missouri River
*Based on the number of parameters having medians which exceed reference
levels selected for comparative purposes.
Source: Data from EPA (15)
429
-------
1. Under the anti-degradation philosophy espoused by PL 92-500 in 1972,
there was less need to devote resources to receiving water impact
assessment. Urban runoff did not become recognized as a problem
until after 1972. Thus, little attention was given to this problem.
2. Impacts of sewage effluent, industrial wastes, and other discharges
mask the impacts of urban runoff. Even when other sources have been
reduced or eliminated, their residual impacts in terms of benthal
deposits are often still evident.
3. The increased reliance on mathematical models for assessing receiv-
ing water impacts reduced the level of effort , in field sampling
programs.
4. The greatly enhanced emphasis on broad-based environmental impact
assessments diverted effort from the more traditional sanitary
survey approach to assessing impacts. These studies produced
relatively little hard information on impacts from urban runoff.
5. The cost of sampling programs is relatively high due to the
intermittent nature of storm events, wide variations in flow and
concentration, and general inexperience with this type of activity.
6. Expected impacts from urban runoff are relatively subtle and do not
cause obvious large scale problems. Thus, more refined and
longer-term sampling efforts are needed to develop reliable
cause-effect information. Indeed, if experience in the related area
of sediment transport in receiving waters is any indication, it may
be many years before these cause-effect relationships are
understood.
Most cases of documented impacts were made for minor receiving water
bodies. Larger water bodies tended to have sufficient dilution, adequate
flushing, or multiple source inputs which obscure impact differentiation.
The complexity of the stormwater/receiving water systems is such that general
correlations were not possible between characteristics of urban areas and
receiving water impacts. Cause-effect comparisons of data show that storm-
water impacts at any given site may contradict those at another. It was
determined that a case-study approach was probably best for future work and
that detailed site-specific data collection is needed to adequately document
storrawater runoff impacts on receiving water bodies.
Few, if any, trend-type impacts, or causes of impacts can be identified
at this time. Ecological deterioration, eutrophication, reduction of quality
of life, alterations of trophic state indices, depletion of animal popula-
tions, and reductions in diversity are all equated to some degree with
stormwater runoff, but are not well correlated to specific losses of
beneficial use. Event-type impacts-are the easiest to document and will
continue to receive greater emphasis. Trend-type documentation will take
years to develop even though there are indicators that they exist.
430
-------
The significance of stormwater impacts on receiving water bodies cannot
be assessed at this time. The sparseness of documented cases, the lack of
detailed data, and the general focus of.stormwater investigations into water
quality and system dynamics (and away from actual 'impacts) do not provide a
substantial basis for determinations.
Nationwide attention has not been focused on impact documentation.
Thus, the data bases are just now beginning to be extensive enough to address
scientific correlation of constituents with respect to impacts, and because
some possible sources of impact documentation have not been searched at
site-specific levels, the anount of documentation identified is probably much
less than what exists. In other words, documentation is only beginning to be
found because the search has just begun.
ACKNOWLEDGMENTS ., • . • • -
This paper is based on research sponsored by the Municipal and
Environmental' Research Laboratory, U.S. Environmental Protection Agency.
Mr. John English, the Project Officer, provided valuable in-house information
on receiving water impacts. Mr. Dennis Athayde, his staff, and consultants
provided information on EPA's Nationwide Urban Runoff Program. Dr. Tim
Stuart and his staff were very helpful in providing access to EPA's fish-kill
data. Mr. Richard Field and Mr. Doug Ammon kept us current on the activities
of EPA's Storm and Combined Sewer Branch; in "Edison, New Jersey.
REFERENCES , " ' '
Heaney, J.P., _et^ al_. Nationwide Evaluation of Combined Sewer Overflows
and Urban Stormwater Discharges, Vol. II: Cost Assessment and
Impacts., EPA-600/2-77-064, U.S. Environmental Protection Agency,
Cincinnati, Ohio, 1977. 364 pp.
Tarr, J.A., and_McMichael, F.C. Historic Turning Points in Municipal
Water Supply and Wastewater Disposal, 1850-1932. Civil Engineering,
Vol. 47, No. 10: 82-86, 1977.
Federal Water Pollution Control Act Amendments of 1972. PL 92-500,
92d Congress, S. 2770, October 18, 1972.
Canter, L.W. Environmental Impact Assessment. McGraw-Hill Book Co.,
New York, 1977. 331 pp.
Dunette, D.A. A Geographically Variable Water Quality Index Used in
Oregon. Journal Water Pollution Control Fed., 51; 53-61, 1979.
Ott, W.R. Water Quality Indices: A Survey of Indices Used in the United
States, EPA-600/4-78-005. U.S. Environmental Protection Agency,
Washington, D.C., 1978.
431
-------
7. Ehrenfeld, D.W. The Conservation of Non-Resources. American Scientist,
Vol. 64: 648-656, 1976.
8. Freeman, A.M. The Benefits of Environmental Improvement. The Johns
Hopkins Press, Baltimore, Maryland, 1979. 272 pp.
9. English, J.A. In-House Files on Receiving Water Impacts.
U.S. Environmental Protection Agency, Cincinnati, Ohio, 1978.
10, Anonymous Fish Kills Caused by Pollution. Fifteen-Year Summary:
1961-1975. EPA-440/4-78-011. U.S. Environmental Protection Agency,
Washington, D.C., 1979. 78 pp.
11. Battelle Memorial Institute. Benefits from Water Pollution Abatement-
Beach Closings and Reopenings, Final Report to National Commission
on Water Quality. NTIS No. PB-251 221, Washington, D.C., 1975.
107 pp. plus appendices.
12. CH2M-Hill. 1978 Needs Survey, Cost Methodology for Control of Combined
Sewer Overflow and Stormwater Discharge. EPA 430/9-79-003.
U.S. Environmental Protection Agency, Washington, B.C., 1978.
196 pp. plus appendices.
13. Keefer, T.N., Simons, R.K., and McQuivey, R. S. Dissolved Oxygen Impact
from Urban Storm Runoff. EPA Report (Draft), Cincinnati, Ohio,
1978.
14. Seyb, L., and Randolph, K. North American Project. A Study of
U.S. Water Bodies. EPA-600/3-77-086. U.S. Environmental
Protection Agency, Corvallis, Oregon, 1977. 537 pp.
15. Environmental Protection Agency. National Water Quality Keport to the
Congress. EPA-44019-74-001 and 002. Vol. 1 and 2. Office of
Water Planning and Standards, Washington, D.C., 1974. 305 pp. and
279 pp.
16. Schneider, W.J. Water-Data for Metropolitan Areas. U.S. Geological
Survey Water Supply Paper 1871. Washington, D.C., 1968. 397 pp.
17. Biernacki, E. In-House Files on Fish Kills. U.S. Environmental
Protection Agency, Washington, B.C., 1979.
18. Olivieri, V.P., Kawata, K., and Kruse, C.W. Microorganisms in Urban
Stormwater. EPA-600/2-77-087. U.S. Environmental Protection
Agency, Edison, New Jersey, 1977. 182 pp.
432
-------
TAMPA
Demographic data
1970 population - 369,000; Urbanized area - 44,000ac; % combined
sewers - 0
Hydrologic background (Schneider, 1968)
The municipal and industrial supplies for the Tampa area come
chiefly from the Hillsborough River. Because of seasonal distribution
of rainfall and the limited storage capacity of the city reservoir, this
source is inadequate during dry periods, and a supplemental supply from
a large spring (Sulphur Spring) is utilized. Adequate quantities of
water are available from the Floridan aquifer to meet the future water
requirements of the area.
Much of the metropolitan area is subject to hurricane damage because
of its location near sea level and because of extensive residential
development along the waterfront. Tampa, which is in the lower reaches
of the Hillsborough River, is subject to flood damage during periods of
excessive rainfall. However, this problem will be alleviated in the
near future as flood regulation reservoirs and bypass channels are
completed upstream from the city. Other problems of major importance
are encroachment of saline water on fresh ground-water supplies, disposal
of municipal waters, and the effects of the metropolitan complex on the
coastal waters.
Precipitation - 52.0 in/yr
Waste sources
Sewage - 11.2 in/yr.; Urban runoff - 19.3 in/yr
Receiving waters
Primary - Hillsborough River
Mean annual flow - 21.1 in/yr
Dilution ratio - 0.69
Other - Hillsborough and Tampa Bays
Special studies
The Hillsborough River is the primary river draining the Tampa
area. The upstream portion of this river is used as a water supply
source for the City of Tampa. It receives urban runoff. The lower
portion of the river moves through the City of Tampa where it receives
433
-------
inputs from a variety of sources including landfill leachate, water
treatment plant alum sludge, and overloaded sanitary sewers. The discharge
from the Hillsborough River enters Hillsborough Bay. The bay has serious
water quality problems and extensive sludge accumulations due to the
discharge of primary treatment plant effluent until very recently.
Tampa is one of the EPA National Urban Runoff Program study areas.
Tampa was one of three cities selected as case studies of estimating
the impact of improved water quality on beach closings. (Battelle,
1976). The results are summarized below
Tampa-St. Petersburg, Florida. The population region for this study
case included Pasco, Pinellas, Hillsborough, and Manatee Counties with a
combined 1970 population of 1,185,664. In addition to the resident
population, there are an estimated 4,432,000 businessmen, vacationers,
and other travelers coming to this area each year. Many-of these individ-
uals came expressly for the purpose of swimming on their vacations. It
is estimated that over 80 percent of the tourists participate in the
winter and about 60 percent in the summer.
In recent years, six smaller beaches in Pinellas County have been
closed intermittently as a result of high coliform counts following
heavy rainfall. In Hillsborough County, a beach along the Hillsborough
River has been closed permanently since 1972 as a result of bacterial
contamination. These areas total 2,450 frontage feet or 1.4 percent of
the estimated 178,320 feet of beach frontage in the four-county area.
Because of the availability of abundant high quality .beaches, the
effect from storm runoff on estimated total resident and tourist swimming
activity days is negligible. For the specific beaches that were inter-
mittently affected by high coliform counts, assuming that swimmers avoid
them completely, an annual increase of 1 percent in resident and .12
percent tourist activity days were estimated. Because these estimates
assume complete seasonal closure, actual loss in activity days would be
lower assuming individuals use these beaches intensively when water
quality is acceptable. Also, no determination of the specific type and
level of use of the affected beaches was obtained during the course of
data collection.
References
Federal Water Pollution Control Admin. 1969. Problems and Management
of Water Quality in Hillsborough Bay, Florida. NTIS PB-217 147,
U.S. Dept. of Commerce, Springfield, Va. 94 pp.
Lopez, M.A., and D.M. Michaelis. 1979. Hydrologic Data from Urban
Watersheds in the Tampa Bay Area, Florida, USGS Water-Resources
Investigations 78-125. 51 pp.
434
-------
435
-------
HILLSBOROUGH RIVER BASIM
3045. Hillsborough River near Tampa, Pla.
Location.—Lat 88°Q1'E5", long 82°25'40", in sec.29, T.88 S., R.19 E., on left bank just
upstream from spillway of Tampa reservoir dam, at Thirtieth Street, 5* miles northeast
of Tampa, Hillsborough County.
Drainafle area.—650 sq mi, approximately.
Records available.—October 1938 to September 1960.
0§Se. —Water-stage recorder. Datum of gage Is at mean sea level, datum of 1929 (city of
Tkmpa bench mark). Prior to Oct. 1, 1945, at site 1.4 miles upstream "at "datum 0.66 ft
higher.
Average discharge.—22 years (1938-60), 685 ofs (495,900 acre-ft per year), adjusted for
diversion.
Extremes.—1938-60: Maximum discharge, 14,600 cfs Mar. 21, I960: maximum gaze height,
tOSS ft Aug. 2, 1960; no flow Nov. 30 to Dec. 2, 1945.
Maximum stage known, S5.6 ft Sept. 7, 1933, at former site and datum, from flood-
narks, affected by backwater prior to failure of Tampa power dam, 1.4 miles below
former gage. A discharge of 16,500 ofs was measured Sept. S, 1933.
RDBiPka.—Flow regulated by Tampa reservoir since Oct. 1, 1945. Capacity of reservoir
insufficient to affect monthly figures of runoff. Diversion at point lj miles above
station ftr water supply by city of Tampa. Records of chemical analyses for the period
November 1956 to September 1,958 are published in reports of the Geological Survey.
Monthly and yearly mean discharge, In cubic feet per second a/
Vatci
year
19M
19SS
1353
19S4
1SS5
me
1357
19EB
law
13C-
Oet.
SIS
151
187
Z.79S
268
191
US
1,340
144
1,957
Nov.
236
221
391
C40
109
197
172
157
no
80S
Dec.
364
358,
141
1.79S
129
120
46.1
112
134
217
Jan.
3S3
185
188
850
125
83.1
34.8,
414
887
231
Feb.
23,4,,
109
3E1
250
216
174
49.4
466
4i7
464
Mar.
199
521
171
177
76.6
42. 9
366
1,975
3,082
4,926
Apr.
406
598
1,065
103
65.1
13.9
759
647
2,022
1,358
Hay
186
flft,l
217
97.8
33.1
, 14.8
388
204
740
154
June
52.4
257
170 '
327
34.3
8.48
318.
74.5
1,853
22O
July
144
226
469
723
215
31.5
528
494
2,7CF
1,200
Aug.
498
641
1,965
1,199
89°
3!?9
1,834
811
2,738
4,713
Sept.
097
429
4,371
308
1,099
324
1,790
238
3,597
4,270
r ffl.s
The year
333
371
852
781
204
,102
^oi
, "1,546
1,718
a Unadjusted for diversion by city of Tampa.
Monthly and yearly discharge, in acrp-feet
Vatir
year
19SI
im
19M
1914
19S.S
19S*
13S7
1316
19!»?
1SCO
Oct.
37,830
46,180
48,410
17 1.900
is.soo
11,730
46,840
62,370
a.aso
120,400
Nov.
14,040
13,160
23,840
38,070
G.S10
11,730
10,250
9,350
10,110
47,900
Dec.
22,400
22,000
8,680'
110,400
7,940
7,370
2,830
7,320
8,230
13,370
Jan.
21,710
11^390
"tt;s9o
55,260
7,660
5,110
2,140
25,440
54,530
14,230
Feb.
13,010
9,700
17.820
1S,880
11,980
10,030
2,75O
27,OOO
23,180
26,700
Mar.
12,240
32,030
10,490
10,860
4,7lO
2,640
22,490
121,500
189,500
302, 900
Apr.
24,160
35,560
63,360
6,150
s;a7o
830
45,140
50,380
120,300
80,840
Hay
11,410
4,920
13,320
6,010
2,040
908
23,880
12,520
45, 520
9,440
June
3,120
15,270
10,140
19,440
2,040
505
18,920
4,440
110,200
13,070
July
8,870
13,910
28,830
44,450
13,190
,1,94,0
32,470
30,iSO
166,300
73,770
Aug.
30,600
39,620
120,900
73,720
49,190
2,210
U.2,800
49,890
168,300
289,800
Sept.
41,450
2S,550
260, 100
18,310
65,400
19',2'BO
106,500
14,170
214,000
254,400
The year
240,800
209,300
616,900 - „ • , ,i i,,1;,, . „
565,500 •• ......
191,000 / . , ' i „„ a'.
74, 280 '
425,800
435, 300
1,119,000
1,247,000
!!2Jt£. --Figures given herein prior to October 1956, not previously published. l; " " J °
discharge, in cubic feet per second
Year
1950
IS*?
1953
1954
1955
• ' —
WSP
„
1204
1234
1274
1334
1384
1958; 1454
1957, 1504
1958 1554
1959) 1G24
1960)1794
PMat.er year endln^'Sept.1 30"'
Observel
|~~^~ltoxlmuiirdTy HTnT3
31aiharg«
_
1,500
1,900
6,830
112,890
1,900
810
3,810
3,180
7,390
C14.GOO
Da*.fc day
,' _ , '
Sept. 24,25,1951
Apr. 5,1952
Sept. 30, 1953
July 31,1954
Sept. 15,1955
Sept. ?,B, 195G
Aug. 10,1957
Oct. 6,1957
Har.23,24,1959
Mar. [11,1900
^
24
28
37
2a
30
6.4
2O
31
84
17
Mean
_
333
371
852
781
264
102
sea
601
1,546
1,718
Acre-feet
_
(240,600
t269,300
*616,9OO
*565,500
"$191,000
*74,280
42S,800
•43i>,300
1,119,000
1,247,000
A'3 Justed, a/
Mean
n
361
399
882
810
296
138
623
638
1,586
1,760
Per
jquar*
mile
_
0555
.614
1.36
L25
.455
.212
.958
.982
'.44
2.71
lunoff
In
Lnche!
i
7.53
8. ,3S
18,42
16.91
6.16
2.91
13.01
13.34,
33.11
36.84
1 " TCafen.
Observed
1
Mean 1 Acre-f
¥98
342
370
1,183
381
,,264
141
645
501
i,^bS
*360,
*247,
t268,
tss'e,
*276,
,"9°J
102,
466,
362,
1,274,
Jar yeaY
^ f't-lju
„,
|
700, 526
900 371
3dt 3yfl
900)1,'<:K
3C 1 "' 111-
901 ! 29f
200 l~"l.
600 " . t' I
900 t J
000 U, 799
,|" „
i
3te^j/
lunorr
In
Inches
10.99
7.74
8.33
25:33
8.57
6.18
3.73
'1.20
.11.27
3/.57
a Adjusted for dlveral
-------
STATISTICS OF RECEIVING WATER RESPONSE TO RUNOFF
By
Dominic M. DiToro*
Environmental Engineering and Science Program
Manhattan College, Bronx, New York
Abstract
An analysis is presented for the mean and variance of a one dimensional
advective dispersive system that is subjected to random inputs of runoff.
Analytical solutions are available for which the mass input is represented as
a Poisson process of delta functions. The effects of event to event vari-
ability of runoff mass discharges are included in the formulation as are the
random times between overflows. The method of solution evaluates the over-
lapping effects in the receiving water due to the persistence of discharged
mass. Both conservative, first order reactants, and sequentially reacting
substances are considered.
The solutions have certain unexpected properties. In particular, the
normalized variance of BOD and DO are symmetric about the discharge point,
whereas the means of the concentration are not. This is explained in terms
of the effects of advection and dispersion of fluctuations. The analytical
solutions are compared to simulated results using an observed hourly rainfall
sequence. The results indicate that the within event variability is not
significant, if the receiving water dispersion is large enough. It is also
pointed out that treatment devices such as retention basins which remove a
certain average fraction of the overflowing mass are less effective in re-
moving variance in the receiving water. This phenomena is of importance in
the evaluation of the probable benefit of runoff treatment.
1.
Introduc tion
The analysis of receiving water responses to intermittent
discharges from separate and combined sewer overflows is a
problem of substantial difficulty. The problem setting is in-
herently time variable due to the transient mass discharges, and,
although a time variable integration of the mass balance
equations describing the receiving water response is possible,
it is an unwieldy and laborious undertaking.
*Consultant, Hydroscience, Inc., Westwood, N.J., Associate
Research Professor Environmental Engineering and Science Program.
Manhattan College, Bronx, N.Y. 10471.
437
-------
Consider the results of such a simulation: receiving water
concentrations are calculated for each volume element for the
duration of the integration, e.g. three summer months. And
since "the purpose of computing is insight not numbers" [1] it
is necessary to perform this calculation many times with differ-
ent mass discharge sequences, treatment alternatives, and re-
ceiving water parameters in an effort to determine the important
controlling mechanisms. Faced with this large amount of infor-
mation, it is probable that the computational results would be
summarized in terms of the relevant statistics: the mean,
variance, and the percentiles of the receiving water concentra-
tion.
The purpose of this paper is to present a statistical
analysis of the response of a water body to a sequence of runoff
mass discharges. The statistics of the receiving water concen-
tration are related to the statistics of the runoff process and
the parameters of the water body. Certain quite general form-
ulas are presented which apply to any receiving water type.
These are specialized for an advective-dispersive system and
analytical solutions are derived for the mean and variance of
the response.
2. Runoff as a Pulse Process
Consider a particular discharge event i with duration d.,
centered at time t., for which the mass discharge rate within
the event is constant, w.. Then the mass discharge rate enter-
ing the receiving water ior the period of the analysis, W(t), is
the superposition of these events, suitably displaced in time:
W(t) =
£
i
r(di,t-ti)
(D
where:
r(dirt) = 1
= 0
- d./2 <
-L —
elsewhere
d./2
-L
(2)
a rectangular pulse with unit height and duration d., centered
at time zero. 1
If the substance being considered is conservative or decays
following linear kinetics, e.g. total suspended solids, coli-
forms, or BOD, then the principle of superposition applies and
the receiving water response, c(x,t), at any location, x, can be
written as the sum of the responses to the individual mass dis-
charges suitably displaced in time:
438
-------
c(x,t) = Y,
(3)
where h is the unit response of the receiving water to a dis-
charge centered at time t., of unit magnitude (w.=l), duration
d., with discharge flow, q., augmenting the flow1in the receiv-
ing water. The situation is illustrated in Figure 1. The
sequence of mass discharges produces a sequence of individual
concentration responses which possibly overlap and their super-
position is the receiving water response. The problem is to
compute the statistics of c(x,t) from the statistics of the run-
off and the parameters that determine,the unit response of the
receiving water.
The analysis is tractable if the occurrence times, t.,
satisfy certain assumptions. These are best described in1terms
of <5., the time between the midpoints of the events:
"6i- = W - t± (4)
It will be assumed in what follows that: ~
(a) 6} = exp (-6/A) - (5)
where A is the average time between the midpoints of the events,
This distribution has the property, that the standard deviation
of 6 equals the mean so that the coefficient of variation: ..
'
V6= cr6
/A = 1;
(b) the time between events are serially uncorrelated
and in fact independent, that is:
6.-A 6 . -A
(6)
These assumptions imply that t. is a Poisson process. In
addition the other runoff parameters must satisfy certain
assumptions:
(c) The times between events are uncorrelated with the
other runoff variables, that is
d.-D 6.-A
(7)
439
-------
r
W(t)
*
1
1
1
1
1
w
wl
-*-d, •*•
1
T
1
i
! w,
c(x,t)
RUNOFF MASS
DISCHARGE
RECEIVING
WATER
CONCENTRATION
RESPONSE
Figure I. Schematic representation
of the problem framework. Mass discharges
corresponding to runoff events at t|,t2>
enters the receiving water at X = 0. The resulting
concentration response is the superposition
of the individual concentration responses.
440
-------
where D = E{d.}; and similarly for p ~ and p g.
(d) The runoff quantities: w., d. and q. are serially
uncorrelated: i i i
d.-D
-D
= 0
k > 0
(8)
and similarly for w. and q..
(e) The probability distributions of all the random
variables are independent of time (stationarity).
(f) The principle of superposition applies and the re-
ceiving water response is a function only of the time
difference t-t., and not an explicit function of the
time t.. That is, the coefficients in the relevant
mass balance equations are constants in time (time
invariant linear system).
Although it is difficult to verify assumptions (a)-(e) with
observed runoff data since they are usually of insufficient
quantity, it is straightforward for rainfall data. The compari-
sons can be made using standard goodness of fit tests. As an
example of the procedure, the hourly rainfall data for the
summer of 1967 (an average summer) for Boston's Logan Airport
gauge are converted into event variables [2]. The choice of
the minimum length of dry hours that separate storm events is
based on assumption (a). In particular it is chosen so that
v
-------
r
o>
o
-0.5
-1.0
-1.5
-2.0
EMPIRICAL
THEORETICAL
(DISPLACED)
50
100
150
S. (HOURS)
200
250
300
Figure 2. Empirical log survivor function of the midpoint
Intervals. The ordinate is log|0 of the plotting position;
the qbcissa are the values of the ordered observed
midpoint intervals. The theoretical curve (displaced
downward by 0.5 for visual clarity) represents the
result for an exponentially distributed random variable.
442
-------
1.0 ,-
-I.O1—
MIDPOINT INTERVAL
Y 1 T *"""
• 9 ,, 1 ,, | 1 o
A , "It H
fj ? ' T i
— -in
DURATION
, l'f 1 1 }
1 1 Y 9 T
ft
If
lag k
2 4 6 8 10
lagk
02 4 6 8 10
1.0
^ 0
a.0
— 1 O
r- , 1.0
INTENSITY
,-'*'' hi
i I ' T ^ •* 0
1 1 T ' 1 ~>
l}' T{
VOLUME
, il,,}.
\ i T 1 T T
1 O
figure 3! Sample serial auto correlations for
storm midpoint interval, duration, intensity, and volume
with approximate 80% confidence limits.
443
-------
there is a slight tendency for large values to be followed by
smaller values of the event variables, this effect is small and
the independence assumption appears reasonable.
Assumption (c), the lack of cross correlation between the
midpoint interval and the other event variables is tested in
the same way. Table 1 presents the results: p.g and PV(r are
not significantly different from zero; however p,* = 0.448 is
significant (99%). This would affect the analysis if the
receiving water response is strongly affected by the runoff
duration. As shown subsequently the more important quantity is
runoff volume which is not significantly correlated to the mid-
point interval.
Perhaps the most difficult assumption to verify is that
p . «= 0 since this requires actual runoff data. As shown in
Taole 1, the rainfall volume is independent of the preceding
midpoint interval. However, lack of correlation between w^
and 6. requires that the concentration, c.^, of the runoff also
be uncorrelated with 6.. This is a site specific requirement
that depends on the mechanisms which influence the resultant
pollutant concentrations. If the principle mechanism is a
steady buildup of pollutants on the land surface between runoff
events then a strong correlation might be expected. However, in
one case at least [5] only a slight relationship was observed.
In any case this assumption provides a convenient starting
point for the analysis of receiving water response.
The stationarity assumptions (e) and (f) can be violated
if the period of analysis spans too long a period of time since
rainfall has definite seasonal patterns, as does receiving water
temperature and flow. However, stationary periods for runoff
can be chosen from an inspection of the monthly distributions
of the relevant rainfall statistics. An example [2] is shown in
Fig. 4. In particular it can be seen that the change in the
means and coefficients of variation over a three month summer
period are not too severe so that stationarity is a reasonable
assumption for certain periods of time. This period should
also coincide with the period for which the receiving water
parameters are constant, assumption (f).
In summary, these assumptions guarantee that the runoff
events occur purely randomly in time; that their properties
are uncorrelated event to event; and that the seasonal effects
are negligible. It is for this simplified yet realistic
situation that analytical results are available.
444
-------
TABLE 1
RAINFALL STORM STATISTICS
LOGAN AIRPORT, BOSTON, MASSACHUSETTS
May 15 - September 15, 1967
Mean
V
S
(day)
2.70
1.04
(in/hr)
0.0706
1.32
d
(hr)
4.73
1.22
V
(in)
0.358
1.98
Cross Correlation Coefficients
1.0
i
d
-0.149
1.0
0.448
0.044
1.0
0.257
0.592
0.735
445
-------
MEAN
VARIATION
0.12
0.10
0.08
0.06
0.02
12
10
8
6
4
2
0.6
0.5
0.4
02.
01
120
100
80
60
40
20
_STORM INTENSITY '
I i i i i i i i i i i I
1 2 3 4 5 6 7 8 9 10 II 12
MONTH
-DURATION
i i i i i i i i i i i i
12345 678 9 10 II 12
MONTH
_ DEPTH
i i i i i iii ti it
12345 678 9 10 II 12
MONTH
-TIME BETWEEN STORMS
i i i i i i i i i i i i
1 2 3 4 5 6 7 8 9 10 II 12
2.5
2.0
-..5
1.0
0.5
2.5
2.0
5* ''5
1.0
0.5
2.5
2.0
> ( 5
5'
1-0
0.5
2.5
2.0
~> 1.5
A
1.0
Q5
-_v— v^
i i I i i i I i I i I i
1 2 3 4 5 6 7 8 9 10 II 12
MONTH
12 3456 7 89 10 II 12
MONTH
' -^^
i i r i i i i i i i i i
1 2 3 4 5 6 7 8 9 10 II 12
MONTH
-
i i i i i i i i i i i i
1 2 3 4 5 6 7 8 9 10 II 12
MONTH
MONTH
Figure 4. Seasonal variation of the means and
coefficients of variation for storm intensity,
duration, volume, and midpoint integral from the
Central Park, New York City rain gauge for the
years 1948-1975.Values are computed from
all storms occurring within each month.
446
-------
3. Mean and Variance of the Concentration Response
The expressions for the mean and variance of the concen-
tration, Eq. (2), are available from an application of the
Generalized Campbell Theorem for Poisson processes [6],[7].
Applying Campbell's theorem to Eq. (3) yields these gener.al
results for the receiving water concentraton mean and variance:
-,
E{c(x,t)} = J E{w± / h(x/t,di,qi) dt} (9)
— OO
OO
V{c(x,t)} = ~ E{w? / h2(x,t,di,qi) dt} (10)
— OO
These remarkable formulas are a complete' solution for the
statistics of the receiving water concentration, the problem
posed in the introduction. It is only necessary to compute the
time integrals of the unit response and its square, and then
to perform the expection (i.e. the statistical average) of these
quantities with respect to the runoff variables: w., d. and
q,. This is an important simplification of the alternative: to
compute long term simulations for the period of interest and
statistically analyze the result. Instead, what is required is
the unit responses for a sufficient range pf durations and run-
off flows so that the expectations can be computed. For the
idealized case in which the duration and runoff flow do not
significantly affect the unit response, then only a single unit
response is necessary. This case is discussed below.
These equations provide a method for investigating the
importance of various phenomena on the mean and variance. The
question can now be posed in terms of what effect these
mechanisms have on the time integral's and expectations in the
formulas. In particular, it is immediately apparent that in-
creasing the average time between events decreases both the mean
and the variance as I/A. This is the case regardless of the
relative decrease in the overlapping effects of successive run-
off events as the time between events increase.
Such insights are valuable even if complete simulations
are computed since if the simulations behave in a significantly
different way, then the analyst can inquire into,the cause of
this unexpected behavior. In fact these formulas provide a
basis for the intuitive understanding of receiving water
response to random, pulse loadings. Such information is invalu-
able when dealing with complex numerical simulations since the
question: are the simulations correct, is of constant concern,
and methods of confirming the computations are constantly
447
-------
sought. Eqs. (9) and (10) provide the basis for a validation
of the simulation.
4. Application to Advective-Dispersive System
The equations for the mean and variance, Eqs. (9) and (10)
can be evaluated directly if the unit response of the receiving
water is independent of the runoff duration and flow. The
latter is the case in rivers and estuaries for which the base
freshwater flow is large relative to the runoff flow. The effect
of duration is small if the dispersion in the receiving water
is large. For moderate dispersion the effect of finite duration
runoff events is to spread out the loading over time and reduce
the peak receiving water concentration. Using the approximation
discussed below will tend to overestimate the receiving water
impact in this case and provides an upper bound.
Consider the limiting case for which the duration approaches
zero and the mass discharge rate approaches infinity in such a
way that the mass discharged at each event, m., is finite. The
result is a mass discharge rate of the form: 1
W(t) =
6(t-ti)
(11)
where S(t) is a delta function: an impulse of infinite height,
zero width, and unit area centered at t = 0. The receiving
water response is:
c(x,t) = E
h(x,t-ti)
(12)
where h(x,t) is the unit impulse response of the receiving
water. These equations are limiting versions of Eqs. (1) and
(3) for duration approaching zero and no runoff flow effect.
The version of Campbell's theorem that applies in this case is:
E{c(x,t)> = -/ h(x,t)
A 0
dt
(13)
M'
/ h (x,t) dt
0
(14)
where M = E"tm.}, the mean mass discharged, and
r* O
v =3 v{m.)/E {m.}, the normalized variance of the mass dis-
cnarged a"t each1event.
448
-------
Eq. (13) for the mean has a direct atnd general interpreta-
tion. The time integral of the impulse response is the steady
state response to a continuous unit mass discharge rate. E{m.}/
A can be thought of as the mean mass discharge rate (mean masl
discharged during an event/mean time between events). Hence,
Eq. (13) implies that the average receiving water concentration
can be calculated from the steady state response to a continuous
constant mass discharge corresponding to the mean runoff mass
discharge rate. This general result applies to any receiving
water configuration - the form of h(x,t) is as yet unspecified -
so that the results of steady state numerical model calculations
are directly applicable in computing the average concentration
response, if runoff duration and flow effects are negligible.
Unfortunately the integral of the square of the impulse
response has no general interpretation and it must be computed
or approximated in each case. The expression involving the
mass discharge statistics is interesting since it indicates
that the variance of the receiving water concentration is due
to two effects: the intermittent discharges, and the
event-to-event variability of the mass discharged. Even if the
mass discharged at each event were of equal magnitudes so that
v = 0, there would still be receiving water concentration
variance due to the intermittent nature of the discharges.
Event-to-event variation of mass discharges augments the
variance as shown in Eq. (14).
The magnitude of this effect can be estimated in the
following way. Since m. = c.v., and if it is assumed that c.
and v. are independent, then:
M
TT
= v v +
c v
v
+ v (15)
c
=0 and
= which from Table 1 is seen to be v = 1.98 for this
For the limiting case of constant concentration, v
M = which from Table 1 is seen to be v = 1.98
xample" so that v,, = 4 and the effect of ¥vent-to-event
.M
variation is to increase the variance by a factor of five,
which is quite substantial.
5.
Constant Parameter Case
In order to compute the square of the impulse response a
specific receiving water configuration is required. For the
case of an advective-dispersive system with constant parameters,
the impulse response to a unit mass discharge is well known to
be [8] :
h(x,t) =
(x"Ut)
t > 0
(16)
449
-------
where A, E, U, K are the cross-sectional area, dispersion,
velocity, and decay coefficients respectively. The evaluation
of the mean and variance involves the definite integral [9]:
/
0
v-l
dt = 2 (T^.)
Kv(/aX)
(17)
where K (Z) is the modified Bessel function [10] with the
specialvcase Kx (Z) = /ir/2Z e . For the case of the mean
the result is:'2
E{c(x,t)} =
-(1 + m)
(18)
where p = Ux/E, n= KE/U , and m /l+4r|. The minus sign
applies for x > 0 and the positive sign for x < 0. As
expected, this is the steady state solution for an equivalent
continuous mass discharge rate of M/A.
The variance of the concentration follows directly from
Eqs. (14) and (17) :
V^c(x,t)> =
ep K (mp)
(19)
where since K (-Z) = K (Z) Eq. (19) applies for all x.
The behavior of the variance at the point of discharge,
x = 0, is unrealistic since it is infinite, K (Z)~ -ln(Z) as
Z~*Q. The mathematical reason is that the impulse response
at x=0 is h(0,t)~e //t and the integral of the square of
this function is infinite. This is due to the impulse of zero
duration and infinite mass discharge rate which produces an
infinite concentration at the moment of discharge. The
concentration then decays slowly enough to cause the variance of
concentration to become infinite. In fact the smoothing
induced by the finite duration of the runoff event is not
included in the analysis and the unrealistic result is not
too surprising.
An unexpected consequence of these equations appears if
the normalized variance of concentration is considered:
V2{c{x,t)} =
KQ(mp)
(20)
450
-------
2 2
where v {c(x,t)} = V{c (x, t) }/E {c (x, t) ]. .;H The normalized
variance is symmetric about the point,of discharge whereas
neither the mean nor the variance are symmetric if the fresh-
water flow is significant. Fig. 5 illustrates the spatially
varying portion of the solution for a hypothetical BOD discharge
into an estuary. The asymmetry of the mean response is in
sharp contrast to the symmetry of the normalized variance. The
infinite variance at the origin is also apparent although it is
clear that it decreases rapidly as the distance from the origin
increases.
The reason for this symmetry is related to the effect of
the advective portion of the mass transport on the mean and
variance. The effect of pure advection and decay on a time
variable mass discharge, W(t), is to translate the time axis by
the travel time x/U, i.e. c(x,t) = W(t-x/U) exp(-Kx/U)/UA with
an appropriate reduction due to decay [11]., Thus the normalized
variance in the receiving water at any point downstream of the
discharge, assuming no runoff flow effect, is the normalized
variance of the mass discharge. Advection, by itself, pro-
vides no reduction in normalized variance. It is dispersion
that provides the smoothing that reduces normalized variance.
And since the effect of dispersion is increased as the distance
either upstream or downstream from the discharge point increases,
the normalized variance decreases symmetrically from the point
of discharge. In effect the advection compresses the upstream
profile and expands the downstream profile but the contraction
and expansion affects both the mean and variance. The normal-
ized variance removes this effect and exhibits only the
symmetrical effects of the smoothing caused by the symmetrical
dispersion.
The effect of increasing the average time between events,
A, is to increase the normalized variance. The reason is that
the concentration as a function of time is more variable rela-
tive to the mean since, although the events occur less frequent-
ly, they still produce the same spikes of concentration in the ,
receiving water. Conversely as the time between events is
decreased the events occur more and more frequently with more
and more overlap until the discharge approaches a continuous,
constant,- discharge rate and the normalized variance approaches
zero.
The behavior of the normalized variance as a function of
a contour
parameters
position and reaction rate is illustrated in Fig. 6:
plot of v (c (x,t) }/ [ (1+v )u A/E] . The cjimensionless j-^~,^^.^^
for position, Ux/E, reaction rate, KE/U , and the normalization
have been chosen to isolate the effect of varying the reaction
451
-------
en
E
o
• y .
UJ
15
12
M= I05lbs
A=l04ft2
— E- I.OmiVdoy
1T=0.5 mi/day
Kd = Kr=0.lday
"A=2.7Oday
-VM=I.98
I I
-20 -10 0
X(MILES)
10
20
1.50
1.25
"- 1.00
o
»-V-
M
0.75
0.5 O
0.25
-2O
I
I
I
-10 0
-X. (MILES)
10
20
Figure 5*. Computed mean and normalized
variance for BOD discharged into an hypothetical estuary.
452
-------
O.I
Figure 6! Contour plot of dimension I ess
normalized variance versus normalized distance
and estuary number.
453
-------
rate. Increasing the reaction rate increases the normalized
variance: e.g. coliform discharges produce much higher
normalized variance then do slowly reacting substances, e.g. BOD.
The reason is that the initial concentration in the receiving
water due to a runoff event is independent of decay rate whereas
the average concentration is markedly influenced. As a result
the variance caused by the high initial concentrations remains
as the reaction rate increases whereas the mean concentration
decreases.
The almost parallel contours indicate that increasing
reaction rate increases the normalized variance uniformly with
respect to distance so that the shape of the spatial profile is
essentially unchanged and only the magnitude increases as
reaction rate increases.
The effect of changing the transport characteristics can be
seen in Fig. 7 a contour plot of^v {c(x,t)}/KA(l+vM) as a
function of estuary number, KE/U ., and position 'as indexed by
the mean concentration profile. The locations considered range
from 90% to 10% of the concentration at the origin. This dis-
tance scale tracks the spatial extent of the relevant portion
of the mean concentration profile. 2A number of interesting
features are illustrated. For KE/D >1, the normalized variance
is essentially independent of KE/U , the ratio of 'dispersion
to advection. For a nine fold reduction in mean concentration,
normalized variance is reduced three fold, indicating that
normalized variance decreases less slowly than mean concentra-
tion in this spatial region.
2
For KE/U <1, dispersion is decreasing relative to advection
and normalized variance increases. This four fold spatial
decline2at KE/U =0.1 is more marked than the three fold decline
at KE/U =10.0 but the differences are not large, indicating
that, for this spatial region, the effect of changes IB trans-
port are not excessively dramatic: a two fold or.less change,
in normalized variance over a hundred fold change in estuary
number. The somewhat surprising conclusion is that reaction
kinetic effects are much more important in determining normal-
ized variance than are transpart effect so long as KE/U >0.1,
and runoff flow and finite duration effects are small.
6. Comparison to Simulations
A critical assumption in the above analysis is that the
mass for each runoff event enters the receiving water as a
narrow pulse. In fact, the mass is discharged over a time
interval that is comparable to the duration of each rainfall
454
-------
"V
10.0
CM
UJ
K
O.8 O.7 0.6 O.S
Figure 7*. Contour pilot of dimensionless various versus
fraction of concentrations at the origin and estuary number.
455
-------
event. In order to check the validity of this simplification,
a time variable simulation of a two month summer period has
been carried out [2]. The parameters of the advective disper-
sive system are chosen to be representative of the Hudson River
estuary during summer low flow. A single point source is used
in order to highlight any difficulties that might be masked if
multiple inputs were considered. The runoff mass is discharged
in accordance with the hourly rainfall observed at the Central
Park raingage. A constant coliform concentration is assumed.
However the mass discharge has both within-event and event-to-
event variability since the runoff volumes vary hourly within
the event and from event to event in accordance with the rain-
gage data. A comparison of the theoretical results and the
simulation is shown in Fig. 8. The results are in close agree-
ment., indicating that the effect of within-event variation and
overflow duration are negligible in the estuary.
7. Dissolved Oxygen Deficit
The methods developed in the previous sections can be ex-
tended to sequentially reacting substances, most noteably BOD
and DO deficit. All that is required is the DO deficit impulse
response to an impulse of BOD. This is easily obtained from
the recurrence relation that applies to sequentially reacting
substances [12]. The impulse response can be written directly
in terms of the single reactant impulse response, Eq. (16):
h(x,t,K ,K )
JL d
[h(x,t,K ) - h(x,t,K
ci -I
(21)
where K , K , and K, are the reaeration, BOD removal, and
deoxygenation rates and the notation h(x,t,K ) indicates that
Eq. (16) is to be evaluated with K =
this impulse response yields:
K
Applying Eq. (13) to
E(D(x,t)} =
K,
t E(c(x,t,K )> - E(c(x,t,K
(22)
where D(x7t) is DO deficit concentration and the notation again
indicates the reaction rate to be used in the evaluation of Eq.
(18). The variance is found from Eq. (14) using Eq. (17) to
evaluate the integrals. The result is:
456
-------
100,000
50,000
E
O
o
z
Q.
10,000
5,000
oa
ID
O
Z
O
o
£K
O
t 1,000
_J
O
u
_J
< 500
O
IOO
M = 100 tbs
A= 1.5-10s ft2
E= 15 mi2/ day
Q= I04fts/sec
K= 2.0 day1
A= 1.55 day
•VM = 1.0
THEORETICAL
MEAN
STANDARD DEVIATION
SIMULATED
• -MEAN
Q-STANDARD DEVIATION
LOCATION OF STORM LOAD
I I I I I ! I I I I I I I I I I I I I I I I
-12 -10 -8-6-4-2 0 2 4 6
MILEPOIIMT
8 10 12
Figure S! Comparison of theoretical and simulated
estuary response for conform. Receiving water parameters
are characteristic of the Hudson River estuary.
457
-------
V{D(x,t)} =
O V
£• .IY
K
d ^2
r a
) -f-
{K (mp)
(23)
(mrp)}
where the subscript notation indicates which reaction rate to
use in evaluating the dimensionless expressions: n = KE/U ; m
The normalized variance:
(24)
~ 2 K (p/l+2n +2n
\J . CL J
K (m p)
O J,
„_ r\ ~~~' ' ~' . -
exp[l-|.2-| (m +m ) l+m~ ] exp (-|p|m )
m~ exp(-|p[m ) -:
a. a.
is symmetric about the point of discharge. And unlike the BOD
normalized variance the DO deficit normalized variance at the
origin:
A(1+V2)U2 In! (1+2H +2n )/(m m ) 1
ivl O. JL Ci JL
2TTE
+ m
-2
(25)
is finite.
The mean deficit and normalized variance corresponding to
the BOD example in Fig. 5 are shown in Fig. 9. The marked
asymmetry of the deficit solution is in marked contrast to the
symmetrical normalized deficit variance which is seen to be
everywhere smaller than the normalized BOD variance. This is
intuititionaly reasonable since there is a kinetic lag time
during which BOD is converted to DOD and mixing effects can
exert their influence for that much longer.
The normalized deficit variance at the origin can be
used to investigate the effect of BOD decay rate and transport.
Fig.
of the ratio of reaeration to deoxygenation coefficient Ka/Kd/
and estuary number K E/U . For highly reactive BOD K_/K^ is
small, the kinetic time lag is reduced, and deficit
normalized variance increases. Transport effects are analogous
to the single reactant case with more marked increases for
K E/U2<1.0.
Cl
o 2
10 is a contour plot of v {D(0,t)}/K A (1+v ) as a function
458
-------
en
£
LU
-20
1 L
-10
o
X(MILES)
10
20
CM
1.50
1.25
1.00
0.75
0.50
0.25
-20
i^-v* Jc(-x.,t}[-
I II
-10
0
MILES)
10
20
Figure 9! Computed mean and normalized variance
for Dissolved Oxygen Deficit response of an hypothetical
estuary. Parameters as in Figure 5 with Ka= 0.15 day1
459
-------
10.0
!-JD(o,t) r
(I+V
M
10.0
Figure 10! Contour plot of dimensionless normalized
Dissolved Oxygen Deficit versus reaeration rate
reaction rate ratio and estuary number.
460
-------
8.
Conclusions
The methods presented in this paper can be used to evaluate
the statistical behavior "of receiving water concentrations that
result from runoff events. Certain general results are appli-
cable to a large class of receiving water types. More special-
ized results apply to receiving waters that can be idealized as
a one dimensional rectangular advective-dispersive system for
which the runoff flow is not a significant component of the
advection. These methods are especially useful for screening
many alternative treatment schemes that abate runoff mass
discharges. Their effect on the mean and variance of the
receiving water can be calculated without costly and time-
consuming simulations.
These methods have been extended for the analysis of
complex receiving water geometries and multiple discharges.
They will be presented in future publications.
9.
Acknowledgements
These methods are the results of various research and
project related investigations carried out by Hydroscience, Inc.
in the past ten years. The initial ideas were developed'
during conversations with Robert V. Thomann. The contributions
of Eugene D. Driscoll, Mitchell Small, Donald J. O'Connor,
James Fitzpatrick, John A. Mueller, and Charles Dujardin are
specifically acknowledged. The contributions by Dennis Athayde
of the Environmental Protection Agency is also gratefully
acknowledged.
REFERENCES
1. Hamming, R.W., Numerical Methods for Scientists and
Engineers. McGraw-Hill, Inc., 1973, p. 1.
2. Hydroscience, Inc., "A Statistical Method for Assessment
of Urban Stormwater" Office of Water Planning and Standards,
EPA 440/3-79-D23, Environmental Protection Agency,
Washington, D.C., 1979.
3. Cox, D.R. and Lewis, P.A.W., The Statistical Analysis of
Series of Events, Methuen & Co., Ltd., London, 1966. p 7-11.
4. Wine, R.L., Statistics for Scientists and Engineers,
Prentice-Hall, Englewood Cliffs, N.J., p. 551-552.
5. Mueller, J.A., Anderson, A.R., "Combined Sewer Overflow
Quality From Treatment'Plant Data", J. Water Poll. Control
Fed., Vol. 51., May 1979, p. 958.
461
-------
6. Rice, S.O., "Mathematical Analysis of Random ^Noise", in
Selected Papers on Noise and Stochastic Procedures, ed. N.
Wax., Dover Publications, N.Y., 1954, pp. 133-294.
7. Papoulis, A., Probability, -Random Variables, and Stochastic
Processes, McGraw-Hill, N.Y., p. 359, 560-570.
8. O'Connor, D.J., "Analysis of Diffusion Data of the
Delaware River Model." Int. J. Air Water Pollution, Vol. 7
pp. 1073-1089, 1963.
9. Erdelyi, A., Tables of Integral Transforms, McGraw-Hill,
N.Y., 1954, p. 146.
10. Abramowitz, M., Stegun, I.A., Handbook of Mathematical
Functions, National Bureau of Standards, Applied Mathemat-
ics Series 55, Superintendent of Documents, U.S. Government
Printing Office, Washington, D.C., 1954, p. 417-429.
11. Di Toro, D.M., O'Connor, D.J., "The Distribution of Dissol-
ved Oxygen in a Stream with Time Varying Velocity", Water
Resources Research, Vol. 4, N6. 3, June 1968, pp. 639-646.
12. Di Toro, D.M., "Recurrence Relations for First Order
Sequential Reactions in Natural Waters", Water Resources
Research, Vol. 8, No. 1, Feb. 1972, pp. 50-57.
APPENDIX I
NOTATION
Ci
cross sectional area
concentration of,i runoff event
c(x,t) = concentration at location x and time t
di -
D
D(x,t)
duration of i runoff event
E{d.}; mean runoff duration
dissolved oxygen deficit at location x and
time t
462
-------
E = dispersion coefficient
E{ } = expectation of the quantity
h(x,t,d.,q.) = receiving water Response to unit (w.. = 1)
pulse runoff,mass discharge at x = 0 and1centered
at t = 0, of duration d., flow rate q..
h(x,t) = receiving water response to unit impulse of
mass at x = 0, t = 0
i.
i
K
Kr
K
K
d
rainfall intensity of i event
first order reaction rate
BOD removal rate
reaeration rate
BOD deoxygenation rate
K (Z) = modified Bessel function of order v, argument Z
K (Z) =, modified Bessel function of order zero, argument Z
m.
M
m
ma
m
mass discharged at i runoff event
mean mass discharged =
VI + 4n
/I +. 4n
cl
/l + 4n .
p = Ux/E; dimensionless distance
q. = flow rate of i runoff' event
i
r(d.,t) = unit pulse function of duration d. centered at
1 t = 0 1
.th
t
U
midpoint time of occurrence of the i event
time
advective velocity
463
-------
vi
v{
wi s
W(t) =
X
A
n
n.
volume of the i runoff event
variance of bracketed quantity
.th
e.g.; mass discharge rate of the i runoff event
mass discharge rate into advective dispersive
system at x = 0
longitudinal distance
t. , - t.; midpoint interval between runoff
events i+1 and i
o*
E{6 . };-mean midpoint interval
o
KE/U ; estuary number
K E/U2
Cl
V
V
V
V
M
V
V
V ' {6.}/A; coefficient of variation of midpoint
interval 6.
1/2
' {d.}/D; coefficient of variation of duration d.
V
V {i.}/E{i.}; cbefficient of variation of intensity
V {m.)/E{m.}; coefficient of variation of runoff
mass/ m.
V {v.}/E{v.}; coefficient of variation of volume,
V {c(x/t)} = v{c(x,t)}/E {c(x,t)); normalized variance of
concentration c(x,t)
v2{D(x7t)} = V(D(xft)}/E2{D(xft)}; normalized variance of
dissolved oxygen deficit
pg(k)= autocorrelation function of 6. at lag k; eq. (6)
464
-------
p, (k)=
autocorrelation function of d. at lag k; eq. (8)
P-(k)/ Pv(k) = autocorrelation function of i. and v. at lag k
'as
>qfi'
cross correlation coefficient between d. and .6.;
eq. (7) 1 x
= cross correlation coefficients between q., w. and
6, respectively """ """
V
1/2
{6.}; standard deviation of 6.
V
1/2
' {d.}; standard deviation of d..
465
-------
CONTINUOUS RECEIVING WATER
QUALITY MODELING FOR URBAN
STORMWATER MANAGEMENT
by
Miguel A. Medina, Jr.
Department of Civil Engineering
Duke University
Durham, North Carolina 27706
ABSTRACT
A simplified continuous Deceiving water quality model has
been developed as a planning guide to permit preliminary screen-
ing of areawide wastewater treatment strategies. The model sim-
ulates the hypothetical response of the stream or tidal river
system to the separate and combined effects of waste inputs
from: 1) upstream sources, 2) dry weather urban sources, and
3) wet weather urban sources. The total hours of runoff-
producing rainfall throughout a year are separated into storm
events by defining a minimum interevent time. For a given storm
event, the runoff and pollutant loads are summed and critical
dissolved oxygen concentrations are estimated as a function of
several hydrodynamic and biochemical parameters. Alternative
control strategies are evaluated in terms of relative impacts by
determining the probability of occurrence of water quality vio-
lations. Model output includes the downstream dissolved oxygen
sag curves computed per each event, and the dissolved oxygen
profile computed at a user-specified location downstream for all
simulated events. An application to the Des Moines River at Des
Moines, Iowa, is presented.
INTRODUCTION
In a 1.67 square mile (433 hectare) urban watershed in Dur-
ham, North Carolina, it was found that the dissolved oxygen con-
tent of the receiving watercourse was independent of the degree
of municipal waste treatment beyond secondary during storm flows
466
-------
( 3 ). Approximately one-half of the stream miles in the United
States are water quality limited and 30 percent of these stream
segments are considered polluted to a certain degree with urban
stormwater runoff ( 5 ). The implication is that, generally,
secondary treatment of dry-weather wastewater flows is insuffi-
cient to meet desired receiving water quality standards: there-
fore , control of runoff pollution must be considered in areawide
wastewater management plans and abatement programs. The results
of a nationwide assessment of costs and related water quality
impacts derived from non-point sources ( 7 ) were, among others,
that:
, • wet-weather flows represent at least 50 percent of the
total wastewater flow from urban areas;
• a generalized optimization model, assuming linear costs,
predicted primary type facilities are preferable only up
to a 10 percent level of BOD removal for wet-weather
flows, with a secondary type facility preferable for
higher levels of control; and
• on a national average basis using BOD removal as the ef-
fectiveness parameter, approximately 39 percent of the
combined sewer problem and 10 percent of the other wet-
weather flows should be controlled before initiating
tertiary treatment of point sources.
The study also confirmed that gross inadequacies exist in our
present data base and conclusion^ are highly sensitive to sim-
plifying assumptions necessary for successful simulation of com-
plex physical processes occurring throughout.our watersheds.
Nevertheless, mathematical models are needed to predict variable
responses to stochastic hydrologic phenomena.
The 208 planning effort (Section 208, PL 92-500) estab-
lished the need for various levels of urban water management
analysis ( 5 ) to permit, preliminary screening of municipal
treatment alternatives., Four distinct levels of evaluation
techniques, ranging from simple to complex procedures which can
be integrated with one another, are summarized in Table 1. The
first three levels essentially represent various .degrees of
planning detail with models running on hourly time steps for .
long simulation periods (years). Mathematical complexity and
data requirements are kept at a minimum. At the third level,
the methods used to generate storm runoff flows and pollutant
mass rates or concentrations depend upon the continuous hydro-
logic simulation model selected by the user. Such techniques
are well documented for two models which may be applied to both
urban and non-urban watershed: 1) the Hydrologic Engineering
Center's Storage, Treatment, Overflow, Runoff Model (STORM), and
2) the U. S. Environmental Protection Agency's Storm Water Man-
agement Model (SWMM), continuous version of Runoff Block (17,9).
The urban runoff flows and associated concentrations and mass
rates, derived from storm events over the drainage area of
467
-------
TABLE 1. LEVELS OP URBAN WATER MANAGEMENT
ANALYSIS DEVELOPED BY EPA RESEARCH
& DEVELOPMENT PROGRAMS
Level I:
Level II:
Level III:
Level IV:
a desktop calculator, statistical analysis proce-
dure, no electronic digital computer required.
* University of Florida Methodology - permits the
user to estimate the quantity and quality of
urban runoff in the combined, storm and un-
sewered portions of each urban area.
* Hydroscience, Inc., Methodology - use of a
stormwater simulator and an analytical method
based on probability distribution functions
and statistical properties of rainfall, run-
off, treatment and receiving water impact.
a simplified continuous simulation model for
planning and preliminary sizing of facilities,
developed by Metcalf & Eddy, Inc.; or the com-
puterized optimization version of University of
Florida Methodology described above.
a refined continuous simulation model approach.
Continuous hydrologic simulation models (e.g.,
STORM or continuous SWMM) which generate urban
runoff hydrographs and pollutographs are fol-
lowed by continuous receiving water impact anal-
yses (Level Ill-Receiving model) .
• Continuous SWMM - University of Florida
* STORM - Corps of Engineers by Water Resources
Engineers, Inc.
a. sophisticated single event simulation model,
e.g., EPA SWMM developed by Metcalf & Eddy, Inc.,
University of Florida, and Water Resources Engi-
neers , Inc .
468
-------
interest, represent concurrent time series which are read by
Level Ill-Receiving through the standard card reader devices, or
from peripheral storage units (disk/tape/drum). Since the pro-
gram has.the built-in capability of accessing a user-created
data set, any continuous hydrologic and water quality model may
be considered (15). Detailed models such as the single-event
version qf SWMM typically use short time steps (minutes) and
short simulation times (hours) . Their, data requirements are usu-
ally substantial.
MODEL OVERVIEW
Since it is impractical, if not impossible, to meet stream
standards all of the time it appears desirable to express al-
lowable limits for various pollutants on a probabilistic or
frequency basis. The approach guiding the development of Level
Ill-Receiving was that the cost-effectiveness of various treat-
ment alternatives can only be determined realistically by a con-
tinuous analysis of the frequency of violation of water quality
standards. Such an analysis provides an approximation of system
responses to proposed treatment measures and must never be in-
terpreted as other than a technique in the selection of the best
control strategy.
The problem of specific interest is to assess the separate
and combined effects of the major urban sources of water pollu-
tion upon river quality. The relative impact of wastewater
sources is appraised by estimating their effect on dissolved
oxygen (DO) concentrations downstream from the urban area. It
is of further interest to distinguish clearly between the rela-
tive pollutional impacts of storm sewer discharges and combined
sewer overflows. An abstraction of the urban drainage system is
presented in Pig. 1. The urban community served by a separate
sewer system will convey stormwater runoff and municipal sewage
through parallel conduits. The BOD concentration of the storm
sewer runoff is mixed with the dry-weather flow (DWF) and accum-
ulated sewer solids. An interceptor carries the sanitary design
flow to the municipal sewage treatment plant. The combined sew-
er overflow is either given treatment or allowed to discharge
directly to the receiving water. Since complete mixing is as-
sumed, the BOD concentrations of the combined sewer overflow
(Q ) and the flow (DWFCMB) intercepted for treatment by the DWF
facility are identical. Any desired degree of treatment may be
imposed at both the DWF and the wet-weather flow (WWF) treatment
plants. The concentration of the combined BOD inputs in the re-
ceiving water is given by:
BOD =
m
BODdQd
(1)
Qu + Qd + Qw
469
-------
SEPARATE SEWER SOURCE
URBAN RUNOFF
MUN.IC1PAL WASTEWATER
COMBINED SEWER SOURCE
MUNICIPAL WASTEWATER
WWF
STORAGE /
TREATMENT
UPSTREAM
SOURCES
QU,BODU
Fig. 1. Abstraction of the
Urban Drainage System
470
-------
where BOD = mixed BOD concentration in receiving water, mg/1.
m
BOD = mixed BOD concentration from sources upstream of
u
urban area, mg/1.
BOD, = BOD concentration of dry-weather flow treatment
plant effluent, mg/1,
BOD = BOD concentration of wet-weather flow treatment
w
facility effluent, mg/1,
Q = upstream flow, cfs,
Q, = DWF treated effluent, cfs, and
Q = WWF treated effluent, cfs.
The BOD concentrations of the DWF and WWF treated effluents are
given by:
[BOD
DWFSEP
DWFCMB]
BOD, =
BODw =
[BOD,
DWFSEP + DWFCMB
+ BOD . Q] (
. . . . (2)
where BODf = BOD concentration of municipal sewage, mg/1,
BOD = mixed BOD concentration in the combined sewer,
C mg/1,
BOD = BOD concentration of urban stormwater runoff, mg/1,
s
DWFSEP = DWF contribution from separate sewer area, cfs,
DWFCMB = DWF contribution from combined sewer area, cfs,
Q = urban runoff carried by the separate storm sewer,
• • • • s cfs,
Q = combined sewer flow, cfs,
c ' •
,
= fraction removal of BOD achieved by the DWF treat-
ment facility, and
R = fraction removal of BOD achieved by the WWF treat-
w
ment facility.
The initial conditions of BOD in the river are defined by Eq. 1,
and the hypothetical impact on the oxygen balance of the receiv-
ing stream is estimated by using simplified mathematical
471
-------
modeling approaches. The total hours of runoff-producing rain-
fall throughout the year are separated into storm events by de-
fining a minimum interevent time. For a given storm event, the
runoff and pollutant loads are summed and the critical DO defi-
cit is estimated as a function of several stream parameters:
temperature, flow, oxygen concentration, deoxygenation and re-
aeration rates, longitudinal dispersion, and BOD concentrations.
The minimum DO is calculated subsequently and a frequency analy-
sis i%s performed. Stream velocity is computed as a function of
the discharge and the time and distance to each critical deficit
point are obtained for each event.
The options used for the simulations include:
1. five inflow combinations of river flow plus:
a. DWF
b. DWF + separate flow
c. DWF + combined flow
d. separate flow + combined flow
e. separate flow •+• combined flow + DWF,
2. four DWF treatment rates (variable),
3. three WWF treatment rates (variable), and
4. three fractions of measured upstream flow.
Item 4 is included as a model option to investigate whether the
relative impact of urban stormwater runoff is significant in the
upstream portions of river basins. This effect may be simulated
by simply reducing the upstream flow to any desired fraction of
its actual measured value. Thus, discharge into a dry river bed
may be studied.
The impact of urban runoff on the quality of the receiving
stream is evaluated in terms of: 1) critical DO frequency
curves, i.e., by the percent of time predicted minimum DO levels
fall below a specified concentration; and 2) classical DO sag
curves for each inflow combination. During periods of wet
weather, the program performs computations on an hourly time
step. Otherwise, a daily time step is used. An economic evalu-
ation of various control strategies augments the decision-making
capabilities of such an approach.
EVENT DEFINITION
The basic approach to define a wet-weather event is to ana-
lyze the runoff time series and establish the minimum number of
consecutive dry-weather hours (DWH) that separates independent
storm events. The independence of these events is not defined
in a strictly climatologic sense; it is in fact statistically
derived. The DWH refer to periods during which no runoff was
472
-------
produced. If STORM is selected to generate the hydrologic time
series, depression storage and evaporation rates must be satis-
fied before any runoff is predicted and no runoff occurs during
periods without measurable precipitation. The continuous version
of SWMM allows runoff to decay temporally beyond intervals with
zero precipitation input. Therefore, for an identical precipi-
tation time series, runoff events generated by SWMM will gen-
erally be of longer duration.
The runoff time series is subjected to autocorrelation
analysis. For hydrologic processes, it is practical to esti-
mate the autocorrelation coefficients by an open-series approach
(6,18):
rT(k)
n-k .. j"n-k 1 f n
Z x.x. ., r Z x. Z x.
±mi i i+k n-k[.=1 xj [±=k+1 i
. . (4)
n-k
0.5
0.5
i=k+l
where r... (k) = sample estimate of lag-k autocorrelation coeffi-
cient for hydrologic process I,
x. = discrete data series (observations) of hydrologic
process I, for i = l,2,...,n,
n - total number of data points or observations, and
k = number of hourly lags.
The tolerance limits for a normal random time series which is
circular and of lag 1 is given by ( 1 ):
-1 ±t
/n-2
TL [rz(l)] =
n-1
(5)
where t = standardized normal variate corresponding to proba-
a bility level 1 - a.
A circular time series is defined as a series where the last
value is followed by the first so that the time series repeats
itself. Eq. 5 has been extended for use with an open series,
for the general lag case (18). At a 95 percent probability lev-
el, the tolerance limits are given by:
-1±1.645
TL
(6)
473
-------
A plot of the serial correlation coefficients, r(k),
against the number of lags, k, is called a correlogram. The
technique of autocorrelation analysis is essentially a study of
the behavior of the correlogram of the process under investiga-
tion (16). The model compares the value of r(k) obtained from
Eq. 4 with TL [rT(k)], computed by Eq. 6, for the corresponding
number of hourly lags k. The minimum interevent time (MIT)
which separates independent wet-weather events is defined as the
minimum value of k for which r(k) is not significantly different
from zero at a 95 percent probability level.
EFFECT ON RECEIVING WATERS
A simplified mathematical modeling approach is used in
which critical deficits and resulting minimum DO concentrations
are determined for a large number of waste input combinations,
treatment schemes, and receiving water conditions. The develop-
ment of a detailed and sophisticated model is not justified for
the problem context: to provide adequate information on the
relative effectiveness of various pollutant control strategies
in achieving selected water quality standards. The basic theory
of mathematical modeling of one-dimensional bodies of water is
applied for the spectrum of natural systems from freshwater
streams to tidal rivers. The approach is particularly advanta-
geous for a limited data base on natural system geometry, hy-
drodynamic variables, and water quality. Assumptions typical of
models limited for preliminary planning are made (10):
(1) Temporal steady-state conditions prevail, where all
system parameters and inputs are constant with respect
to time, however, a relatively short time step ,(1
hour) is used for simulation.
(2) Natural system parameters (such as flow, velocity, hy-
draulic depth, deoxygenation and reaeration rates, and
longitudinal dispersion) are spatially constant along
the flow axis throughout each time step.
(3) All waste inflows to the receiving body of water occur
at one point.
(4) The effects of various natural biological processes
(algal photosynthesis and respiration, benthal stabil-
ization) are incorporated into a background quality
which is reflected by DO deficit (if none, by satura-
tion) upstream from the waste inflow point. Any ben-
thic buildup is incorporated into the BOD decay rate.
(5) Waste treatment facilities operate at constant effi-
ciencies, independent of hydraulic and organic
474
-------
loadings, for'the entire period of simulation,
Oxygen Balance of Polluted Streams and Tidal Rivers
In view of the modeling objectives, pollutant transport
processes in these systems are approximated by the steady-state,
one-dimensional version of the classical convective diffusion
equation: '•',-.
o =
3x
f§-uc| ±
(7)
where
C = concentration of water quality parameter (pollu-
tant) , M/L3,
3 C
-E. g— = mass flux due to longitudinal dispersion along the
flow axis, the x direction, M/L2T,
UC = mass flux due to advection by the fluid containing
; the mass of pollutant, M/L2T,
S = sources ofi? sinks of the substance C, M/L3T,
U = flow velocity, L/T, and • '. •••-,..-
E = longitudinal dispersion coefficient, L2/T. • :
The equation assumes no diffusion of pollutants through the body
of water boundaries (other than what may be included in the
source-sink term) and is best • suited to predict concentrations
relatively far downstream from the point of waste injection.
Since critical DO deficits usually occur some distance down-
stream from the waste source, Eq. 7 is adequate for such pre-
dictions. The main sources of dissolved oxygen in stream sys-
tems are atmospheric reaeration and oxygen production by photo-
synthesis . The major sinks include carbonaceous oxygen demand
(CBOD) , nitrogenous oxygen demand (NBOD) , benthal demand, and
respiration of aquatic plants. In the model, the effects of
photosynthesis , respiration and benthal demand are assumed to be
incorporated into the measured upstream DO deficit. Field mea-
surements of organic and ammonia nitrogen present in all waste-
water inputs to receiving waters are seldom available, much less
during runoff events. Thus, nitrogenous oxygen demand is cur-
rently neglected. Since it is desired to solve for the DO defi-
cit and
3C
3x
3D
3x
(8)
Eq. 7 becomes:
475
-------
d2D
dD
+ K,L
1 O
- K0D = 0
2
(9)
where
m =
2E
U
for x > o
D = DO deficit, mg/1,
2
E = longitudinal dispersion coefficient, ft /hour,
U = freshwater stream or tidal river velocity, ft/hour,
K
— I
2 = atmospheric reaeration coefficient, hours ,
K, = deoxygenation constant of carbonaceous BOD, hours ,
L = remaining carbonaceous BOD concentration at x = o,
mg/1,
Critical Deficit and DO Levels
The solution to Eq. 9 as a function of time since release
is given by
D
LoKl
t _ fl at
D
,gt
(10)
where D = initial DO deficit, mg/1,
t = *r = translational time, hours,
u2
2E
1 -
1 -
4K2E
U
, hours ,
, hours" ,
4K,E
1 H -- -=— , dimensionless, and
U
4K2E
' dimensionless.
476
-------
The time at which the critical (maximum) deficit occurs is
obtained by taking the partial derivative of Eq. 10 with respect
to time, setting the resulting equation to zero and solving for
time. After simplification and convenient substitutions,
tc (j - g)
In
(11)
where t = elapsed time at which the critical deficit occurs,
hours,
f = self-purification ratio
= K2/K1, and
R = ratio of the initial DO deficit, D , to the initial
° BOD, L , dimensionless. °
Finally, the critical deficit is found by substituting the value
of t , given by Eq. 11 into Eq. 10:
D
uc
:2 ~ Kl
D egtc
(12)
where D = critical (maximum) deficit, mg/1.
c
The minimum DO level is calculated as
C . = C - D
mm s c
(13)
where C . = concentration of DO at maximum deficit, mg/1, and
C = saturation concentration of DO, mg/1.
The saturation concentration .is determined from the regression
relationship (2) ,
C = 14.652 - 0.41022 T + 0.0079910 T2 - 0.000077774 T3
s
(14)
where T = water temperature, °C.
For freshwater streams, the advective flux is significantly
larger than the mass flux due to longitudinal dispersion. In a
tidal river, the advective and dispersive fluxes are both sig-
nificant. However, it sshould be noted that even for large, deep
477
-------
rivers, the effect of dispersion generally cannot be neglected
when the wastewater inputs are time variable (13,14). If the
model user wishes to neglect the dispersive flux, the classical
Streeter-Phelps formulations are then applied.
The Langbein and Durum (11) equation for the prediction of
the reaeration coefficient was chosen because it is most closely
related to subsequent procedures applied to obtain stream veloc-
ity and depth:
= 2.303 3.3 -r^o7
L H1'-"]
(15)
-1
where K2 = reaeration coefficient at 20°C, day ,
U = stream velocity, ft/sec, and
H = stream depth, ft.
The problem lies in obtaining values of U and H, since the
streamflow varies with time. In the absence of measurements, or
if the data cannot be obtained in an expedient manner (as in the
ensuing application to the Des Moines River), an approximation
can be made (12) which uses strong correlations between velocity
versus flow and depth versus flow, namely:
U = aQ
a
12
H =
(16)
(17)
where Q = streamflow, cfs, and
a,,a2,S-j,32 = regression coefficients.
The model user also has the option of computing the deoxygena™
tion coefficient, K,, as a power function of water depth (10):
Kl -
H
(18)
where YT/ Y2 = regression coefficients,
Temperature corrections are applied to carbonaceous BOD, deoxy-
genation and reaeration rate coefficients.
PROGRAM OPERATION
478
-------
Model subprogram interactions are shown in Figure 2. The
main program (hereafter referred to as subroutine MAIN) provides
overall control and includes in its entirety the wet-weather
flow model ((WWFM). The first input data card to subroutine
MAIN allows the user to select which subprograms will be execut-
ed during simulation. The WWFM is a mathematical abstraction of
the physical system depicted earlier in Figure 1. Stormwater
runoff flows and pollutant loads or concentrations generated by
an urban, continuous hydrologic simulation model (e.g., STORM
or SWMM) are read either directly from card input or from pe-
ripheral devices (tape/disk/drum) depending on the machine con-
figuration. The WWFM in subroutine MAIN simulates the convey-
ance system, including mixing in combined sewers of wet-weather
and dry-weather pollutants during periods of runoff; the pollu-
tant removal efficiency of various treatment schemes; mixing of
the various pollutant inflow combinations with upstream sources
in the receiving waters to determine initial conditions of BOD,
DO, streamflow and other parameters; and computes the oxygen
balance of the polluted waters downstream .from the waste
sources. The procedure continues for each independent storm
event as defined by the minimum interevent time (MIT). Pollu-
tant loadings and Deceiving water quality conditions are aver-
aged over each event's total duration, which includes wet-
weather and dry-weather hours. The spatial distribution of dis-
solved oxygen concentrations along the flow axis is computed for
each event for a distance downstream chosen by the model user.
Frequency analyses are performed on either the resultant DO con-
centrations at a specified location downstream, or critical
(minimum) DO concentrations as predicted by the model for the
entire period of simulation.
Subroutine CORREL subjects the hydrologic time series (ei-
ther rainfall or runoff) to autocorrelation analysis. It auto-
matically defines the MIT used in the WWFM to separate wet-
weather events. The methodology has been discussed previously.
This subroutine may be executed independently of the WWFM, but
may be accessed only through subroutine MAIN. Subroutine
MGRAPH, a single and multiple-curve plotting subprogram, is
called by CORREL to display the correlogram of the time series.
MGRAPH, in turn, calls subroutine ROUND to set the appropriate
scale on the coordinate axes from examination of minimum and
maximum values to be plotted.
Subroutine DWFM performs the same functions as the WWFM,
during periods of no urban runoff. Thus, a model assumption is
that no combined sewer overflows occur. Therefore, there is no
"first-flush" effect, and there are no storm events over which
to average pollutant loads. It may be executed independently of
all other subroutines, except MAIN. Subroutine PLOT is called
by the WWFM (contained in subroutine MAIN) and also by subrou-
tine DWFM to display frequency histograms of receiving water DO
concentrations, optionally. Likewise, subroutine MGRAPH is
479
-------
s
u.
3:
CORREL
MGRAPH
DOSAG
ROUND
DWFM
PLOT
•CALLING
— RETURNING
Figure 2. Level Ill-Receiving Subprograms
480
-------
called by both models to plot cumulative, multiple frequency
curves of DO concentrations.
Subroutine DOSAG is called from subroutine MAIN to display
in tabular form and chronological order: the computed dissolved
oxygen concentration at a user-specified location downstream,
for each event simulated by the WWFM only, the DWFM only, or
both. Thus, the listing may include DO concentrations resulting
from wet-weather event pollutant loadings as well as DWP pollu-
tant loadings during periods of no urban runoff. Consequently,
the subprogram sorts the values according to date of occurrence
in order to produce a composite chronological record. These DO
concentrations are read by DOSAG from scratch data sets created
automatically by the WWFM and the DWFM, Subroutine MGRAPH is
called by subroutine DOSAG to provide a plot corresponding to
the tabular listing. The plot brdinate (dissolved oxygen con-
centration) is scaled according to magnitude; however, the ab-
scissa represents the total number of simulated events, numbered
sequentially and not scaled according to their time of occur-
rence since the beginning of simulation. Each event number is
identified according to date in a tabular format which precedes
each plot. , .
APPLICATION TO STUDY AREA
The City of Des Moines, Iowa, is located near the conflu-
ence of the Des Moines River and the Raccoon River. It contains
approximately 200,000 people out of the total of 288,000 for the
metropolitan area ( 4). The mean annual precipitation is 31.27
inches (795 mm) which is approximately equal to the United States
average, and the average value,for the State of Iowa. The urban
area covers 49,000 acres (19,830 ha) of land which has gently
rolling terrain. Most of the area, 45,000 acres (18,211 ha), is
served by separate sewers, while 4,000 acres (1619 ha) are
served by combined sewers.
Selection of the study area was based primarily on data
availability. Davis and Borchardt ( 4 ) conducted an extensive
sampling program of combined sewer overflows, stormwater dis-
charges, and surface waters in the Metropolitan Des Moines area
as part of a combined sewer overflow abatement planning study.
The sampling program was conducted from March 1968 to October
1969. Other considerations revolved around the fact that Des
Moines, Iowa, is somewhat typical of many urban centers through-
out the country:
(1) it has a medium-sized population;
(2) its domestic and industrial dry-weather flows receive
secondary treatment;
(3) its wastewaters are discharged into a non-tidal re-
ceiving stream; and
i
481
-------
(4) the urban area receives a mean annual precipitation
equal to the national average.
The autocorrelation function of hourly runoff events for
Des Moines, Iowa, is presented in Fig. 3. At a numbex of hourly
lags equal to zero, the correlation of the discrete open series
is unity because this point represents the linear dependence of
the data series on itself. An essentially zero correlation (with-
in 95 percent tolerance limits) between runoff events first
occurs at a lag of approximately 9 hours, defining the minimum
interevent time. The physical interpretation is that periods
without runoff for at least 9 hours separate uncorrelated, and
therefore independent, wet weather events. The hourly urban
runoff and associated pollutant loads within each event (in-
cluding DWF pollutant loads during DWH periods less than nine
hours duration) are summed, average conditions are determined,
and the model proceeds with the receiving water analysis.
Model calibration was preceded by calibration of the urban
runoff BODg loading rates for Des Moines, Iowa, as computed by
STORM. The dust and dirt surface loading factors were adjusted
to obtain an annual average flow-weighted BODj- concentration of
53 mg/1 for urban stormwater runoff. The above concentration
was the average value determined by the field monitoring program
in the separate sewer system. The developed mathematical model,
as discussed in the overview, simulates the mixing of stormwater
runoff and sanitary sewage in the combined sewer system. The
annual average flow-weighted BODg concentration of combined sew-
er overflows was computed to be 75 mg/1, including the effects
of first flush (15). The average value measured in the combined
sewer system was reported to be 72 mg/1. Model parameters were
adjusted to obtain an adequate fit between calculated and ob-
served profiles of dissolved oxygen. These curves are shown in
Figure 4, along with other pertinent information, and correspond
to a point 5.6 mi (9.0 km) downstream from the confluence of the
Raccoon and Des Moines Rivers.
Based on precipitation records obtained from the Environ-
mental Data Service, the total rainfall that fell over Des
Moines, Iowa, during 1968 was 27.59 inches (701 mm). STORM com-
puted a total runoff of 10.28 inches (261 mm) over a watershed
area of 49,000 acres (19,600 ha), for an overall urban area run-
off coefficient of 0.37. There were 65 days in the year during
which rainfall was recorded, from which 58 wet-weather events
were defined. The results are presented in the form of minimum
DO frequency curves for the wet-weather and dry-weather periods
throughout the calendar year.
As stated in the overview, a large number of wastewater in-
flow and treatment rate combinations were investigated and the
482
-------
1.0
0.8-
0.6-1
0.4-
0,0'
-o.z-
-0.4.
95 % T.L.
MINIMUM INTEREVENT TIME
9 HOURS OF DRY WEATHER
95% T. L.
0.0 10 20 30 40 50 SO 70 30 90 100
LAG k
Fig. 3. Autocorrelation Functions of Hourly Urban
Runoff for Des Moines, Iowa, 1968.
483
-------
0.2
0.1 -
-at -
-0.2
95% T. L.
/ V\
\l I \ / V \ N
95% T. U.
300 310 320 330 340 350 360 370 380 390 400
hours
0.2-
0.1-
O.O1
-0.1-
\J \r\/ \ /v\
95% T. L.
95% T. L.
V/V/ yy \ A/ \
/ \ /
4OO 410 420 430 440 450 480 470 480 490 500
LA G K, hours
Fig. 3 (continued). Autocorrelation Function of
Urban Runoff for Des Moines,
Iowa, 1968.
484
-------
o
o
If)
'"Vg'UOOOl
H3AIH £>AV
2 £i
CD
(O
cr>
o
•d CD
o 3
-P rH
g o
o o
"d o
Q) (0
H
H -H
O &
ra
0)
ro
0)
•H -H -H
vo
CD • W
o
NO co
-------
•sa*/zui
o
2
OOOI
o
o
H3A1U 3AV
O H
ffl
CD
vi •
%
a
tn
I-
UJ
>
UJ
•?
2 2
i/6uj
-a
to
o> —
§
H -P
ti
Q)
S
C!
•H
-P
C
O
u
486
-------
complete results are available elsewhere (15). A representative
sample is presented henceforth. It is appropriate to examine
the model estimates of critical DO concentration in the Des
Moines River, for all waste inputs, for conditions assumed to
exist in 1968 during periods of urban runoff; (1) secondary
treatment (85 percent BOD removal) of DWF; (2) no stormwater
treatment; (3) full river flow (100 percent of measured flow);
and (4) combined sewer area equal to 8.16 percent of the total
urban area. Fig. 5 represents the irdn.imum DO frequency curves
for these conditions. The curves indicate clearly that all com-
binations including a substantial amount of wet weather flow
(WWF) result in a drastic decrease in river minimum DO concen-
trations. For example, 63 percent of all the wet weather events
throughout the year produced conditions in the receiving water
that caused minimum DO levels below 4.0 mg/1. Combined sewers
contributed WWF from only 8 percent of the total urban area mod-
eled, yet the BOD concentration was sufficiently high to inflict
an appreciable reduction in DO levels when compared to DWF
sources during periods of runoff.
The minimum DO frequency curves in Fig. 6 compare six
treatment alternatives to reduce water pollution during periods
of urban runoff:
1. 85 percent treatment of DWF and 75 percent treatment
(BOD removal) of WWF,
2. 85 percent treatment of DWF and 50 percent treatment of
WWF, • ' .
3. 85 percent treatment of DWF and 25 percent treatment of
WWF, ;
4. 95 percent treatment o.f DWF and no treatment of WWF,
5. 85 percent treatment of DWF and no treatment of WWF,
and i
6. 30 percent treatment and no treatment of urban runoff.
Even option 3, with a small percentage of urban runoff treat-
ment, is superior to all alternatives controlling DWF only. An
economic evaluation of these treatment alternatives, on an an-
nual basis, is presented subsequently.
Dry weather was experienced for approximately 300 days
throughout 1968. The model was applied to these periods using
a daily time step. This modification is certainly justified
since conditions are more truly steady-state than during periods
of precipitation and subsequent runoff: for example, waste
loadings (DWF treatment plant effluent) and river flow do not
vary as much during the day. For the dry weather simulation pe-
riod, upstream river flow was on the average 94 percent of total
river flow, ranging from 82 percent to 99.6 percent. The re-
sults are shown in Fig. 7. A remarkable 97 percent of the dry
weather days exceed a minimum DO concentration of 4.0 mg/1 with
487
-------
3
5
03 ~ 2
to £ I
s s s
>g Z3
iSlrf
§g^|
•^ 8§g&
o
cr
o
o
'O'Q N3AI9 9N1Q330X3 S1NQA3 M3H1V3M 13M %
488
-------
w
00
•H
-------
ffl
o
00 O
(D O
cn co
& I
Q;
^- ^™*
5 co
6Q %
gs +
LJX U.
°-!5t;8
^ ..
o uj
5! fee
cr
o
ID i-
ir *^ S™ \-frn
1 Sl|
. P= to
I
r
\ 66 66 666 66-
' OTCOr>-COlO^.tOOJ"~<;
A10SSIQ
^^__°IE_^
Ml
3
•
r~
•
tn
•H
Q S.
•OJ £
a I
cr
h-
LU
Q o
•
•3
cu
•H
O
m
en
U w
Q)
>H >
U -H
G 4J
o; us
3 S
&1 J-l
(U
-------
DWF secondary treatment. The Des Moines River carries a rela-
tively high BOD load upstream of the Des Moines urban area.
Thus, even during dry weather periods only, a significant in-
crease in the DWF treatment rate does not result in a corres-
ponding increase in the critical DO levels.
COST-EFFECTIVENESS OF CONTROL ALTERNATIVES
Six alternatives to existing DWF treatment in Des Moines
during 1968 are considered in the evaluation:
1. upgrading the treatment plant from a high-rate trick-
ling filter configuration to an activated sludge-
coagulation-filtration system (85 percent BOD removal
to 95 percent) ,. ]
2. downgrading DWF treatment to primary (85 percent ,BOD
removal to 30 percent), or •. , >
3. providing urban runoff control at four levels of BOD
removal .while maintaining secondary treatment of DWF
(i) 25 percent, ' .
(ii) 50 percent,
(iii) 75 percent, or •'•. <. -.
, (iv) 85 percent WWF treatment. '•• - ;,,,
To achieve a tertiary treatment configuration, the Des
Moines DWF plant must undergo an intermediate-stage modification
to activated sludge. A detailed cost analysis is presented in
the nationwide assessment ( 7 ). Cost estimates of providing
various storage/treatment mixes for stormwater control are based
on the results of a constrained optimization model (8 ). Total
annual cost for the optimal solution can be approximated by
functions of the form: .
Z* = ke
(19)
where
Z* = total annual cost, $ per acre
k,3 = parameters
RI = percent pollutant removal,
R, _< R, _< R.. t
RI = minimum percentage pollutant control such that the
function is convex (e.g., 10 percent), and
RI = maximum percentage pollutant removal.
491
-------
For example, the cost of wet-weather control for the storm sew-
ered area of Des Moines may be estimated by
ZST - 3.61 e
0-037
(20)
where
Zf,— = annual cost, $ per acre of storm sewered area
RI = percent BOD control,
0 < R! < 85.
Similarly, the cost for the urban area served by combined sewers
may be approximated by
_ oc m
= 25.07 e
0.056
where
CO
= annual cost, $ per area of combined sewered area.
A basic assumption is that a secondary treatment technology is
applied to control urban runoff pollution (e.g., biological
treatment, physical-chemical treatment). The total annual costs
for both dry-weather and wet-weather flow control are summarized
in Table 2.
The Des Moines River stretches for 200 miles (322 km) from
the City of Des Moines to its junction with the Mississippi
River. The entire reach is classified as a warm water "B"
stream by the State of Iowa, such that the absolute minimum DO
level specified is 4.0 mg/1. Figures 8 and 9 summarize the
cost-r-effectiveness of each option during periods of urban runoff
and throughout the entire simulation, respectively. Costs of
wet-weather control include the annual cost of secondary treat-
ment of DWF.
Fig. 8 illustrates that, even during periods of urban run-
off, wet-weather control should be considered only after a sec-
ondary treatment technology of DWF has been achieved. However,
beyond that level 25 percent and 50 percent WWF treatment be-
come competitive with DWF tertiary treatment technology. It is
clearly shown in Fig. 9 that, when the entire period of simula-
tion (1 year) is considered, that 85 percent WWF treatment be-
comes asymptotic to the minimum number of violations that can be
achieved, at a very high cost. Both 25 percent and 50 percent
WWF treatment remain strong alternatives to higher than secon-
dary treatment of DWF in relative cost-effectiveness.
492
-------
TABLE 2., ANNUAL COSTS FOR CONTROL
ALTERNATIVES
Alternative
Total Annual Costl
($ /year)
Dry Weather*
a. Activated Sludge-Coagulation-
Filtration
b. High-Rate Trickling
Filter
c. Primary
Wet Weather Control!
(percent BOD removal)
a. 25
b. 50
c. 75
d. 85
3,507,000
1,843,000
405,000
816,000
2,682,000
9,293,000
15,479,000
* For plant size 35.3 mgd (1.55 cu m/sec).
t Includes amortized capital costs (20 yrs, 8%, Engineering
News Record index 2200), operating and maintenance costs.
T Based on 45,000 acres (18,211 ha) of storm sewer area;
4,000 acres (1619 ha) of combined sewer area.
493
-------
16
14
z
o
E
10
8
00
o
CJ
i
o
2:
REMOVAL OF WWF;
DWF SECONDARY
TREATMENT (85%).
INCREASING % 8 0 D —
REMOVAL OF DWF;
NO WWF TREATMENT
DES MOINES, IOWA
PRECIPITATION YEAR
OF RECORD! 1968 "
ANNUAL RUNOFF: IO.28 in.
DWF TREATED: 35.3MGD
30^
99.9 80 60 40 20
% STORM EVENTS VIOLATING
4.0 mg/I D.O. STANDARD
0.1
Fiq. 8.
Cost Effectiveness of Control Alternatives
During Periods of Urban Runoff.
494
-------
o
16
14
12
10
co
••
U)
8 8
co
8
INCREASING % BOD
REMOVAL OF WWF(
DWF SECONDARY
TREATMENT (85%).
INCREASING %BOD <
REMOVAL OF DWF;
NO WWF TREATMENT
DES MOINES, IOWA
PRECIPITATION YEAR
OF RECORD; 1968
ANNUAL RUNOFF: 10.28 in.
DWF TREATED: 35.3MGD
75
50
30
0-
120 100 80 60 40 20
TOTAL NUMBER OF DAYS DURING WHICH
4.0mg/l D. 0. STANDARD IS VIOLATED
Fig. 9. Cost .Effectiveness of Control Alternatives
During Entire Simulation.
495
-------
SUMMARY AND CONCLUSIONS
A simplified continuous river water quality model has been
developed to permit preliminary planning and screening of area-
wide urban wastewater treatment alternatives, in terms of fre-
quency of water quality violations and more traditional ap-
proaches such as dissolved oxygen profiles. It evaluates the
stream response to the separate and combined effects of waste
inputs from: 1) upstream sources, 2) dry-weather urban
sources, and 3) wet-weather urban sources. Among its capabili-
ties are:
1. Level Ill-Receiving may be interfaced through peripher-
al storage devices with various hourly, continuous ur-
ban catchment hydrologic simulation models, notably the
Hydrologic Engineering Center model STORM and the con-
tinuous version of the EPA Storm Water Management Model
(SWMM) .
2. A large number of wastewater inflow combinations to the
receiving body of water, dry-weather flow and wet-
weather flow treatment rates, and upstream flow con-
ditions may be simulated. Thus, a comparative evalua-
tion of many urban pollution control alternatives is
possible in terms of their subsequent impact on receiv-
ing water quality.
3. Continuous analysis allows representation of water
quality impacts due to the random occurrence and prob-
abilistic nature of hydrologic phenomena.
4. It computes a minimum interevent time to define statis-
tically independent storm events.
Level Ill-Receiving has been developed on a general basis
so that it may be applied to the surface drainage phase of most
urban catchments by simply changing the input data to reflect
the particular study area and hydrologic time series. There is
virtually no limitation to the size of catchment modeled. In
theory, an unlimited number of storm events may be processed;
however, practical considerations such as computer time and
costs may be limiting to some users. Data requirements are com-
mon to engineering analysis of non-point source problems and
complete instructions on data preparation are provided. Field
measurements, quantitative and qualitative, are nesessary to ad-
equately calibrate model parameters and verify predicted values.
The methodology is not applicable to stream and tidal river sys-
tems of such geometry and hydrodynamic behavior as to require
multi-dimensional transient analysis.
An application to Des Moines, Iowa, has been presented.
496
-------
Results indicate that wet-weather control should be considered
after a secondary treatment technology has been achieved to con-
trol dry-weather wastewater inputs. The results are, of course,
site-specific.
ACKNOWLEDGMENTS
The development of the model was supported by funds under
Grant No. R-802411, Municipal Environmental Research Laboratory,
Office of Research and Development, U. S.-Environmental Protec-
tion Agency, Cincinnati, Ohio. Project officers were Richard
Field and Chi-Yuan Fan, Storm and Combined Sewer Section, Edi-
son, N. J. Wayne C, Huber and James P. Heaney of the University
of Florida, provided invaluable suggestions during model devel-
opment and also final review.
APPENDIX I.—REFERENCES
1. Anderson, R. L., "Distribution of the Serial Correlation Co-
efficients," Annals of Mathematical Statistics, Vol. 13,
1942, pp. 1-13.
2. ASCE, Committee on Sanitary Engineering Research, "Solubil-
ity of Atmospheric Oxygen in Water," Journal of Sanitary
Engineering Division, Proc. ASCE, Vol. 86, No. SA4, July,
1960, pp. 41-53.
3. Colston, N. V., "Characterization and Treatment of Urban
Land Runoff," Technical Report EPA-670/2-74-096, U. S. En-
vironmental Protection Agency, Washington, D. C., Dec.,
1974.
4. Davis, P. L., and Borchardt, F., "Combined Sewer Overflow
Abatement Plan," Technical Report EPA-R-73-170, U. S. En-
vironmental Protection Agency/ Washington,D~.C., April,
1974.
5. Field, R., Tafuri, A. N., and Masters, H. E., "Urban Runoff
Pollution Control Technology Overview," Technical Report
EPA-600/2-77-047, U. S. Environmental Protection Agency,
Washington, D. C., Mar., 1977.
6. Fiering, M. B., and Jackson, B. B., "Synthetic Streamflows,"
Water Resources Monograph 1, American Geophysical Union,
Washington, D. C., 1971.
7. Heaney, J. P., et al, "Nationwide Evaluation of Combined
Sewer Overflows and Urban Stormwater Discharges, Volume II:
Cost Assessment and Impacts," Technical Report EPA-600/2-77-
0646, U. S. Environmental Protection Agency, Washington,
D. C., Mar., 1977.
8. Heaney, J. P., Nix, S. J., and Murphy,.-M. P., "Storage-
Treatment Mixes for Stormwater Control," Journal of the En-
vironmental Engineering Division, ASCE, Vol. 104, No. EE4,
Aug., 1978, pp. 581-592.
497
-------
9. Huber, W. C., et al, "Interim Documentation, November 1977
Release of EPA SWMM," University of Florida, Gainesville,
Florida, 1977.
10. Hydroscience, Inc., "Simplified Mathematical Modeling of
Water Quality," U. S. Environmental Protection Agency,
Mar., 1971.
11. Langbein, W. B., and Durum, W. H., "The Aeration Capacity
of Streams," U. 5. Geological Survey Circular 542, Wash-
ington, D. C., 1967.
12. Leopold, L. B., and Maddock, T., "The Hydraulic Geometry of
Stream Channels and Some Physiographic Implications," U. S.
Geological Survey Professional Paper 252, Washington,
D. C., 1953.
13. Liu, Henry, "Predicting Dispersion Coefficient of
Streams," Journal of the Environmental Engineering Divi-
sion, ASCE, Vol. 103, No. EE1, Feb., 1977, pp. 59-69.
14. McQuivey, R. S., and Keefer, T. N., "Simple Method for Pre-
dicting Dispersion in Streams," Journal of the Environmen-
tal Engineering Division; ASCE, Vol. 100, No. EE4, Aug.,
1974, pp. 997-1011.
15. Medina, M. A., "Level III: Receiving Water Quality Model-
ing for Urban Stormwater Management," Technical Report
EPA-600/2-79-100, U. S. Environmental Protection Agency,"
Washington, D. C., Aug./ 1979.
16. Quimpo, R. G., "Autocorrelation and Spectral Analyses in
Hydrology," Journal of Hydraulics Division, Proc. ASCE,
Vol. 94, No. HY2, Mar., 1968, pp. 363-373.
17. "Urban Stormwater Runoff: STORM," Generalized Computer
Program 723—S8-L7520, Hydrologic Engineering Center, Corps
of Engineers, July, 1976.
18. Yevjevich, V., "Stochastic Processes in Hydrology," Water
Resources Publications, Fort Collins, Co., 1972.
APPENDIX II.—NOTATION
T/ie. i
al'a2
BOD
BOD
BOD.
BOD
m
&ymboJLi> one. a&e,d AM.
= Regression coefficients;
= Biochemical oxygen demand, mg/1;
= Mixed BOD concentration in the combined sewer,
mg/1 ;
= BOD concentration of wet-weather flow treatment
facility effluent, mg/1;
= BOD concentration of municipal sewage, mg/1;
= Mixed BOD concentration in receiving water, mg/1;
498
-------
BOD
= BOD concentration of urban stormwater runoff, mg/1;
BOD
u
BOD
w,
BOD5
3(3- Q
' 1' 2
C
min
s
CBOD
D
D
c
Do .
D
u
DO
DWF
DWFCMB
DWFSEP
DWH
E
Y-
H
= Mixed BOD concentration from sources upstream of
urban area, mg/1;
= BOD concentration of treated wet-weather effluent,
mg/1;
= Standard BOD test, 5 days at 68PF (20°C), mg/1;
= Regression coefficients;
= Concentration of water quality parameter, M/L ;
•= Concentration of dissolved oxygen (DO) in the
stream, mg/1;
= Concentration of DO at maximum deficit, mg/1;
= Saturation concentration of DO, mg/1;
= Carbonaceous biochemical oxygen demand;
= Dissolved oxygen deficit = C - C, mg/1;
S
= Critical (maximum) deficit, mg/1;
= Initial DO deficit, mg/1;
= DO deficit in receiving waters upstream of inflow
point, mg/1;
= Dissolved oxygen;
= Dry weather flow, cfs;
= DWF contribution from combined sewer area, cfs;
= Dry-weather flow contribution from separate sewer
area, cfs;
= Number of dry-weather hours preceding each runoff
event;
2
= Longitudznal dispersion coefficient, feet per
second;
= Self-purification ratio, K2/K,;
= Regression coefficients;
= Stream depth, feet;
499
-------
k
k
K]
K,
Lc
n
Q
Q,
Q,
w
R-
rQ(k)
Rw
s
t
Number of hourly lags;
Coefficient;
Deoxygenation constant of carbonaceous BOD, hours
-1
-1
T
= Atmospheric reaeration coefficient, hours
— Remaining carbonaceous BOD concentration, mg/1;
= Total number of data points or observations of a
hydrologic process;
= Streamflow, cfs; . ,
= Combined sewer flow, cfs;
= DWF treated effluent, cfs;
= Urban runoff carried by the separate storm sewer,
cfs;
- Wet-weather flow (WWP) treated effluent, cfs;
= Fraction removal of BOD achieved by the DWF treat-
ment facility;
= Sample estimate of lag-k autocorrelation coeffi-
cient for rainfall;
= Deficit load ratio = D /L ;
= Sample estimate of lag-k autocorrelation coeffi-
cient for runoff;
= Fraction removal of BOD achieved by the WWF treat-
ment facility;
- percent pollutant removal;
- minimum percent pollutant removal;
= maximum percent pollutant removal;
2
= Sources and sinks of the substance C, M/L T;
= Standardized normal variate;
= Elapsed time at which critical deficit occurs,
hours or days;
= Stream temperature, °C;
500
-------
TL [r_(k)] = Tolerance limits,at a 95 percent probability level;
U
WWF
x
xi
X
Z*
ZST
ZCO
= Flow velocity in stream, feet per second;
— Wet-weather flow, cfs;
= Distance downstream, jJeet or miles;
= Discrete data series (observations) of a hydrologic
process;
= A hydrologic event;
= total annual cost, $ per acre;
= annual cost, $ per acre of storm sewered area;
= annual cost, $ per acre of e'bmbined sewered area..
501
-------
POTENTIAL OF URBAN STORMWATER IMPACTS BASED
ON COMPARATIVE ANALYSIS OF WET AND DRY WEATHER
POLLUTANT LOADS.
by
Douglas Ammon* and Richard Field**
INTRODUCTION
Urban stormwater discharges and combined sewer overflows are significant
contributors of many pollutants to receiving waters (1,2,3,4). Whether or
not these pollutant loads cause severe enough adverse impacts to justify some
degree of control must ultimately be decided with site specific evaluations.
However, characterization of urban stormwater discharges and combined sewer
overflows, in terms of concentrations and loads, provides useful rough-cut,
order of magnitude indications of potential receiving water impacts. Thus,
summaries of urban wet-weather characterization data are presented for sev-
eral groups of parameters including heavy metals, petroleum hydrocarbons,
biochemical oxygen demand (BOD) and chemical oxygen demand (COD). Heavy
metals and petroleum hydrocarbons are selected because of their abundance .in
urban wet-weather discharges and the growing concern with toxic substances-in
the environment; The oxygen demand parameters are included because dissolved
oxygen levels in surface waters remain the most important and popular indica-
tor of water quality. Furthermore, we should not become too preoccupied with
toxicity issues that we lose sight of other important concerns, such as dis-
solved oxygen depletion. There are many constituent groups that are not
included in this paper (e.g., chlorinated hydrocarbons, nutrients, solids,
pathogenic indicators, etc.). The exclusion of these constituent groups is
not meant to diminish their importance.
Comparisons are presented between the various pollutant loads from urban
wet-weather discharges and from publicly owned treatment works (POTW) efflu-
ents to indicate the relative importance of each source. The significance
of each pollutant is identified based on loading potentials or, whenever
possibles on exceedance of water quality criteria for aquatic life protection.
*Staff Engineer, Storm & Combined Sewer Section, Municipal Environmental
Research Laboratory (Cincinnati), U.S. Environmental Protection Agency,
Edison, New Jersey 08817
**Chief, Storm & Combined Sewer Section, Municipal Environmental Research
Laboratory (Cincinnati), U.S. Environmental Protection Agency, Edison,
New Jersey 08817
502
-------
OXYGEN CONSUMING MATERIAL
Urban stormwater discharges and combined sewer overflows, contribute a
large portion of the oxygen consuming material to receiving waters (1,2,3,4).
The average dry-weather flows (sanitary flows) for the urbanized area in the
United States have been estimated to be about 0.61 Mgal/yr/acre (5700 m^/yr/ha)
and 0.48 Mgal/yr/acre (4500 rn3/yr/ha) for combined and separate sewered areas,
respectively (5). The average wet-weather flows for the urbanized area in
the United States have been estimated to be about 0.45 Mgal/yr/acre (4200
m3/yr/ha) and 0.40 Mgal/yr/acre (3800 m^/yr/ha) for combined sewered and
storm sewered areas, respectively (5). The annual wet-weather flows vary .
between 0.08 - 0.95 Mgal/yr/acre (750 - 8900 m3/yr/ha) for combined sewered
areas and between 0.05 - 0.68 Mgal/yr/acre (470 - 6400 m3/yr/ha) for separate
sewered areas depending on the characteristics of the particular urbanized
area. The BOD concentrations in combined sewer overflows are approximately
one-half the raw sanitary sev/age BOD or about 100 mg/1 (3); however, concen-
trations have been reported as high as 640 mg/1 (3). The BOD concentrations
in urban stormwater discharges are about equal to those of secondary effluents
(3);. however, concentrations as high as 885 mg/1 have been reported (2). The
reader may wish to consult references by Huber et al. (2), Lager et al. (3),
and Manning et al. (4) for additional characterization information for urban
wet-weather discharges. ,
.Using the above approximate concentrations and flows, combined sewer
overflows are calculated to contribute about 380 Ib/acre/yr (425 kg/ha/yr) of
BOD which is about 28 percent of the annual BOD load from the combined sewered
area,given no treatment of the dry-weather flow. If primary and secondary
treatment are provided for the dry-weather flow, then the combined sewer
overflows will contribute 35 percent and 72 percent of the annual BOD load
to the receiving water, respectively. Likewise, in separate sewered areas,
urban stormwater discharges contribute about 70 Ib/acre/yr (75 kg/ha/yr) of
BOD which is about 45 percent of the annual BOD load if secondary treatment
is provided for the dry-weather flow.. Table 1 shows the national annual urban
wet- and dry-weather BOD and COD loads for the-developed urban area of the
United States assuming secondary treatment of the dry-weather flows.
These values represent the annual BOD loads to receiving waters based on
approximate "national" averages. Short-term effects,of -oxygen consuming
material resulting from wet-weather discharges areythe more important con-,
cerns. Obviously, annual loads are distributed, intermittently to -the receiv-
ing water by storm events. Thus, the relative magnitudes of urban wet-weather
loads compared to the POTW discharges are .much greater during and following
storm events. Furthermore, concentrations of oxygen consuming materials are
usually unevenly distributed within a storm event with greater concentrations
often occurring in the first portions of the storm, the so-called first-flush
phenomenon. Depending on the receiving water, reaeration, dilution, etc.,
this may result in shock effects that may be detrimental to aquatic popula-
tions. A recent analysis of 83 water quality monitoring sites in and down-
stream of urban areas found that 42 percent demonstrated a 60 percent or
greater probability of higher than average DO deficit occurring at times of
higher than average streamflow or on days with rainfall (6). This was found
despite the fact that the monitoring sites were not located specifically to
503
-------
observe urban wet-weather DO deficits. Roughly 25 percent of the 83 moni-
toring sites would not meet a 2.0 mg/1 standard, although the frequency
of violations would generally be less than 6 times per year.
Urban stormwater discharges contain oxygen consuming materials which are
not as readily biodegradable as combined sewer overflows and POTW effluents.
COD:BOD ratios for urban stormwater discharges and combined sewer overflows
are presented in Figure 1 (COD is taken as a crude surrogate of ultimate
oxygen demand). Secondary effluents typically have ratios around 2. Higher
ratios indicate that oxygen consumption may occur long after the storm events.
Thus, "background" oxygen demand observed during dry-weather periods are
partially due to wet-weather discharges. Another related characteristic of
urban wet-weather discharges is the high solids fractions, and associated
oxygen demand, relative to the POTW effluents. For instance, combined sewer
overflows, in addition to the solids from the stormwater portion, have a raw
sanitary portion and flushed "dry-weather" deposition which are high in
settleable solids and oxygen demand. Much of the solids from the urban wet-
weather discharges with their associated pollutants will settle in the
receiving water at outfalls and low-velocity regimes contributing to such
problems as high sediment oxygen consumption, physical blanketing (loss of
habitats) and benthic toxicity.
HEAVY METALS
Urban runoff is generally recognized as a major means of transporting
heavy metals to receiving waters. A major source of most heavy metals in
urban runoff is transportation activity. Other sources of heavy metals in-
clude atmospheric particulate fallout and washout (from industrial stacks,
incinerators, open fires and metal smelters), metal plating, soil erosion,
vegetative material, pesticides, chemical spills, and natural background.
Municipal and industrial wastewaters are additional sources of heavy metals
in combined sewer overflows and in storm sewer discharges with extraneous
connections.
Heavy metals may be classified as those essential to life and those not
known to be essential to life processes. The hypothetical relationships among
population effects (lethal and sublethal), heavy metal concentration and
duration of exposure show increasingly beneficial effects (biostimulation) up
to a certain optimum region (i.e., combinations of concentrations and dura-
tions). Beyond this region, there are tolerance regions which may be either
relatively narrow or broad, beyond which benefits decrease, injurious effects
begin, and finally lethal doses are reached. For those heavy metals not
known to be essential, the first region is missing, but there is generally
a tolerance region followed by the toxic and the lethal dose regions. Of
greatest concern are those metals known to be toxic at low or moderate con-
centrations and at relatively short duration of exposure to a variety of
aquatic species or to man. Currently, there are 13 heavy metals on the EPA
priority pollutant list, some of which may fall into this last category for
urban wet-weather discharges. The 13 heavy metals include: antimony, ar-
senic, beryllium, cadmium, chromium, copper, lead, mercury, nickel, selenium,
silver, thallium, and zinc.
504
-------
TABLE 1. NATIONAL ANNUAL URBAN WET- AND DRY-WEATHER BOD AND COD LOAD COMPARISONS.*
Type
Combi ned
Sewered
Storm
Sewered
Unsewered
Total
Percent of
Devel oped
Area
14.3
38.3
47.4
100 .
Annual DWF**
BOD
(mil. Ib.)
340
710
310
1330
COD
(mil. Ib.)
910
1390
830
3630
Annual WWF**
BOD
(mil. Ib. )
880
600
360
. 1840
COD
(mil. Ib.)
2640
2500
1800
6940
Percent WWF
BOD COD
72 74
45 57
54 68
58 67
Assuming*:
CSO
Stormwater
(Sewered)
Stormwater
(Unsewered)
Dry-Weather
BOD COD
(mg/1) (rng/1)
100
30
20
30
300
125
100
80
**lb. = 0.454 kg
20—
15-.
10-.
5-.
. Urban
Stormwater*
CSO*
*Based on flow-weighted storm average
concentrations from 7 CSO locations
and 15 urban Stormwater locations.
Center line, is the overall average
FIGURE 1. RATIOS OF COD TO BOD FOR URBAN WET-WEATHER DISCHARGES.
505
-------
Heavy metals in urban runoff have been investigated at numerous sites
across the country for more than a decade. Some of the results of these
sampling programs are summarized in Table 2 for combined sewer overflows and
urban stormwater discharges. The overall ranges of heavy metal concentra-
tions from the various locations are presented in Figures 2 and ,3 along with
the overall range of POTW effluent concentrations from a national survey (7).
As illustrated, the heavy metal concentrations are highly variable.
For discussion purposes and for comparison with other sources, "typical"
average metal concentrations are needed, although any use of typical values
must be tempered with the usual caveats (e.g., variability of urban runoff
quality). The assumed typical average metal concentrations for combined
sewer overflows and urban stormwater discharges are presented in Table 3.
The typical average metal concentrations are based on the arithmetic average
of flow-weighted storm average concentrations from locations that report
them in that manner. For example, the typical average lead concentration for
urban stormwater is 160 yg/1 which is based on flow-weighted storm averages
from 9 locations and a total of 126 storm events. For some of the metals
there are too few locations with flow-weighted data to make statistically
reasonable estimates. In these cases the typical average metal concentrations
are approximated by the logorithmic mid-point between the overall minimum and
maximum concentrations.
EPA has recently published draft water quality criteria for most of the
pollutants originally listed as toxic under the Clean Water Act (Section 304,
PL 95-217) including 6 of the 13 metals on the priority pollutant list (18,
19). These water quality criteria are estimates of the concentrations of
water constituents in ambient water which, when not exceeded, will ensure a
water quality sufficient to protect a specified water use. Under the Act,
these water quality criteria a're scientific entities based solely on data
and scientific judgment. They do not necessarily reflect considerations of
economic or technological feasibility.
Figure 4 shows examples of the 24-hour average and ceiling water quality
criteria for the protection of freshwater aquatic life for lead. These water
quality criteria and the criteria for 4 other metals are functions of the
hardness in the ambient receiving water. Using the typical average metal
concentrations presented earlier, the number of dilutions of urban stormwater
discharges and combined sewer overflows required to meet the 24-hour average
water quality criteria are presented in Tables 4 and 5, respectively. Arbi-
trary hardness values of 50, 100, and 200 mg/1 as CaC03 are assumed, which
indicates the effects over a reasonable range of receiving water hardness.
Clearly, these 6 metals are in sufficient concentrations to potentially
affect aquatic life, particularly cadmium and copper. Water quality criteria
have not been established for the other heavy metals on the priority pollutant
list; however, these metals are usually present in urban runoff in concentra-
tions at least greater than detection limits.
This analysis assumes that the metals will remain in the water column
and are in available forms. However, many of the metals are associated in
506
-------
TABLE 2. SUMMARY OF HEAVY METAL CONCENTRATIONS FOR COMBINED
SEWER OVERFLOWS' AND URBAN STORMWATER DISCHARGES.
Arsenic Cadmium
(mg/1)
JS = Urban Stormwater
a) a Flow weighted storm averages
b) = Storm averages
c) = Composite samples
d) = Grab samples
The numbers in parenthesis refer to the number of samples or the number of storm averages if reported.
m = mean concentration (flow-weighted).
507
-------
o
a
L
{ED
5 -S
«
£ .
ol
8
(l/6rt)
fc
g
O
7
°S
• t/>
O *—'
^ O
g =
^s
S!
c t-i :
SI
O ZC
O CJ
CSJ
(l/6rt)
508
-------
TABLE 3. ASSUMED TYPICAL AVERAGE HEAVY METAL CONCENTRATIONS FOR
COMBINED SEWER OVERFLOWS AND URBAN STORMWATER DISCHARGES.
Arsenic
Cadmi urn
Chromium
Mercury
Nickel
Zinc
Combined Sewer
Overflows (ug/1)
Urban Stormwater
(wg/1)
830
50*
126
5.8*
730
170*
290
40*
810
160*
1.7
2.2
70
80
460
150*
* Based on Average of Flow-Weighted Storm Average Concentrations From 5 or More Locations.
Antimony
Beryl 1i urn
•Sel entum
Silver
Thallium
Urban Stormwater (yg/1 )**
20
** 1 Grab Sample
San Jose, California (17)
10,000-,-
1,000.
Typical Maximum Urban Stormwater Concentration
Ceiling Criteria
24-Hour Avera
Criteria
Typical Urban Stornwatei/Discharqe_ Concentration
Receiving Water Concentration with /to 1 dilution
ving Water Jloncetrcration^wi_th_l_tp_10.dilution
10
Hardness in Receiving Water (mg/1 as CaC03)
FIGURE 4. PROPOSED FRESH WATER QUALITY CRITERIA
FOR PROTECTION OF AQUATIC LIFE FOR LEAD
1,000
509
-------
TABLE 4. DILUTIONS OF URBAN STORMWATER DISCHARGES NEEDED TO MEET
24-HOUR AVERAGE WATER QUALITY CRITERIA (AQUATIC LIFE).
Metal
Cadmium
Copper
Lead
Nickel
Silver
Zinc
50 mg/1*
15
44
13
4
890++
4
Freshwater Aquatic Life
100 mg/1*
8
28
4
2
890++
3
200 mg/1*
5
18
2
1
890++
2
Sal twater
Aquatic Life
6
100
-
-
31
-
* Ambient Receiving Water Hardness as CaC03
•H- Only One Grab Sample
- Criteria Not Established
TABLE 5. DILUTIONS OF COMBINED SEWER OVERFLOW NEEDED TO MEET
24-HOUR AVERAGE WATER QUALITY CRITERIA (AQUATIC LIFE).
Metal
Cadmium
Copper
Lead
Nickel
Silver
Zinc
50 mg/1*
335
159
64
4
+
17
Freshwater Aquatic Life
100 mg/1*
183
101
22
2
+
n
200 mg/1*
100
64
8
1
+
7
Sal twater
Aquatic Life
125
370
-
-
+
-
* Ambient Receiving Water Hardness as CaC03
+ Insufficient Data
- Criteria Not Established
510
-------
varying degrees with participates; therefore, bottom accumulation will occur
in low velocity regions in the receiving water. Thus, the most probable
impacts will occur in the sediments. These impacts are not reflected by the
water quality criteria. The heavy metals associated with solids, either
several sources of solids in urban wet-weather discharges or digested treat-
ment plant sludges, are presented in Tab'le 6. Metals associated with street
surface dust and dirt and with soils are two sources of metals in urban
stormwater discharges. In addition to these two sources and other sources,
combined sewer overflows can contain significant quantities of solids which
are deposited in the sewer during dry-weather and subsequently flushed out by
the higher wet-weather flows. The metal concentrations associated with this
dry-weather deposition in the Boston area are presented in Table 6 (22).
There are studies that have investigated metal accumulation in the sedi-
ments. For example, Pitt and Bozeman (24) found lead concentrations in the
sediments to be about 10 times greater in the urban reaches than in the non-
urban reaches of Goyote Creek, which receives minimal pollutant discharges,
except for urban runoff. Lead concentrations in urban samples of algae, cray-
fish and cattails were 2 to 3 times greater than in nonurban samples. Zinc
concentrations in urban algae and cattail samples were about 3 times greater.
The nonurbanized section of the creek was found to support a comparatively
diverse assemblage of aquatic organisms including at least 12 species of fish
and various benthic macroinvertebrate taxa. In contrast, the portion receiv-
ing urban runoff was found to support an aquatic community that is lacking
in diversity and dominated by pollution tolerant organisms. Although specific
cause and effect relationships are unclear, it is probably safe to conclude
that pollutional stresses from the urban wet-weather discharges are the major
contributing factor. ,•
Wilbur and Hunter (25) investigated bottom sediments in the Saddle River
upstream and downstream of Lodi, New Jersey. Metal concentrations increased
in the downstream sediment samples. The average metal enrichments are pre-
sented in Table 7. As expected in cross-sectional sediment samples metal
concentrations were the highest in the regions of relatively low velocity
and lowest in the high velocity regions.
Average effluent concentrations from secondary POTW's along with the
typical average metal concentrations for urban wet-weather discharges are
presented in Table 8. Concentration ratios of combined sewer overflows to
secondary POTW effluents and urban stormwater discharges to secondary POTW
effluents are also presented. Cadmium, chromium, lead and zinc concentrations
are typically greater in combined sewer overflows than in secondary POTW
effluents while cadmium, chromium and lead are greater in urban stormwater
than in secondary POTW effluents. In all cases, wet-weather metal concentra-
tions are significant.
Comparisons of heavy metal loads in combined sewer overflows and urban
stormwater discharges with the secondary POTW effluent loads are presented in
Table 9. The annual loads are estimated from the United States average annual
wet- and dry-weather flows for combined sewered, storm sewered and unsewered
511
-------
S
O
in
•o
Cn
J*
1
i
3
|
g
Z3>
O
fc
i
|
S
s
<0
f
II
II
o|
11
M|
1
||
.si 1
II
i
1
gi
I
1
cl
ll
ri
i
«S
DIM
11
s
r-*O
J|s
m
§
i ca
rH *
OCM— -
• tLCO
o fc—
eo
is
1
lls
r*.
CM*? CO
in E—
i
i
i
o
CM
ncentrations
s from 8 CSO
plants
O M **
K
VtCNl
0
:-^ui
h-Q^r
O
i nr tn
o ti to
en 6— '
o
.— t
• CM
OO
i •
00-— •»
SIS'
o
So-— •
in ? to
CM e*--
S--*
T—'in'
s^s.
^-* CO t-*
0 E —
CO
O *— •*
^>
ncentrations
surface
Dirt
o •*-»
o
N
<
.^
t.
in
S j
o;
u a
1— Q.
S "
g
3
l.l £
"" •?
H-« c
s a
LU
<_)
«'
^ ^ i
Q. P .?:
£ LU 6
O
O)
in
cv>
CO
CM
r^
<£J
r-t
CO
(-1
in
CM
in
4->
C
'E
c
LU
«
OJ
o>
O)
2
1
CM T-4
•«• u
co r^.
o.
in co
to cs
CV
vo
co c
CO c
o co
S*?t*
§
in en
io «r
^J" O>
d cvj
e
~ «a
S 0)^,
S"5; £ s>
*•=£ =§
|g7 |gg
^'•"< ^^Ci
^_> ^3 5 jj Jj S
QJ QJ 1_ c en c
ir o o He £ Q
512
-------
f
1—
s:
LU
=1
LU
LU
LU
O
Ol
>-
ce
a-
1
to
Q
•z.
•<
LU
C3
C£.
^
CJ
to
s
Cf
UJ
p
to
Z:
1
=3
to
e^
Ltl
ce.
LU
>
0
ct:
LU
CO
o
LU
I-H
CQ
O
to
^
[»
LU
S-
^^
5
LU
31
LU
1
LU
=c
CO
LU
£
ff
o
c
[*s
a
c
^^
>•<
s.
tj
s
•o
«3
i
S-
. r—
to co
CM LO
*-**
£~
CD
fl3
CO t. ^~
0 =5^.
•*•>
c— - -r- to E:
O O 4^ E:
to C O 4J -Q C>
d> 3 CJ CU — *
O)r— O 3!
re> <»- o c o t—
Q.
5 S cS
CO
h-
UJ
=3
u.
u,
UJ
3
^
•
1
o
UJ
CO
o
CO
UJ
CO
0
UJ
,
3
u
s_
s
1
cu'
a.
Q.
o
CJ
g
c:
|
CJ
3
's
-a
CJ'
CM CO
r— - O CO «3"
^^
CO CO
o to -a- •*
co «a-
o o •*•» '
0 J3 = C
to i. >
LU LU
"io "ia E E
c: c o . o
c E: +-> *»
«C ct to tO
_
2
CO
c
IQ
JD
&.
3
T3
^
cS
O
LU
•X
^
(O
-------
areas, as determined by Heaney et al. (5). These ratios will probably con-
tinue to increase (that is, the relative contribution of urban wet-weather
sources will increase) in the future as POTW's are upgraded and as the
federal pretreatment program reduces industrial contributions to POTW's.
However, lead may be the exception since leaded gas usage should decline.
The storm event load comparisons are based on wet-weather flows assumed to be
ten times greater than the dry-weather flows which is not an unlikely storm
event. For combined sewered areas annual loads of cadmium, chromium, lead
and zinc are greater from combined sewer overflows than in the secondary POTW
effluents. For storm sewered areas annual loads of cadmium and lead are
greater than the secondary POTW effluents.
PETROLEUM HYDROCARBONS
Petroleum enters urban runoff from many dispersed sources; however, it
may be attributed primarily to transportation activity (spills, leaks and
disposal of motor vehicle lubricants, antifreeze and hydraulic fluids). For
instance, about one-half billion gallons (2 million m ) of automotive lubri-
cants are lost to the environment annually (1972) in the United States (26).
Petroleum hydrocarbons, particularly the polynuclear aromatic hydro-
carbons (PAH's), have been shown to be carcinogenic and mutagenic in mammal-
ian and microbial systems. There is also evidence that PAH's produce
cancerous growths in some aquatic invertebrates and vertebrates, especially
early life forms (e.g., spot shrimp larvae). The various PAH's that are on
the EPA priority pollutant list are presented in Table 10.
Direct urban stormwater discharges and municipal waste disposal (waste-
water effluents and sludge dumping) have been estimated to contribute about
10 percent of the petroleum hydrocarbons entering the ocean on a global basis
as shown in Table 11 (27). This may be put into further perspective when
considering that both these sources contribute the major chronic portion to
adjacent coastal and estuarine areas which, on a mass balance basis, con-
tribute more petroleum hydrocarbons to these near-shore environments than
oil spills. It follows that urban runoff contributes a major portion of the
petroleum hydrocarbon loads to fresh waters. For instance, urban runoff has
been reported to be the major contributor of petroleum hydrocarbons to Lake
Washington (28).
The presence of specific petroleum derived hydrocarbons in urban runoff
has only been investigated by a few (28, 29, 30). What has been commonly
reported in urban runoff sampling programs are oiT and grease, usually by
hexane extractable materials (HEM's). For instance, in Jamaica Bay, New
York, about 50 percent of the HEM's are contributed by combined sewer over-
flows (31). Oil and grease cover a broad spectrum of organic material,
including fatty acids, soaps, esters, fats, waxes, and various petroleum
hydrocarbons. Thus, they should only be considered gross indicators of the
presence of petroleum hydrocarbons since they do not distinguish between
origins (i.e., biogenic versus petroleum origins). However, street solids
(dust and dirt fraction) in the Washington, D.C. area have been reported
514
-------
to contain 6400 mg/kg of oil and grease of which 3600 mg/kg are petroleum
hydrocarbons (32). Dust and dirt is a major solids component in urban runoff.
*•' ' --?•
A study in Lake Washington investigated contaminant concentrations in
sediments (11). Table 12 shows the linear correlation matrix of oil and
grease with several other parameters in sediments near combined sewer and
storm sewer outfalls and in control areas. Interestingly, the oil and grease
at the combined sewer outfalls correlate well with TOC while the correlations
are much weaker at the storm sewer outfalls. Likewise, the two metals shown,
which are principally from automotive sources, correlate well with the oil and
grease at the storm sewer outfalls, but not as well at combined sewer outfalls,
This suggests that perhaps the oil and grease in stormwater sediments are
primarily from petroleum origins. The geometric mean concentrations of oil
and grease in combined sewer overflows and storm sewer discharges were 12.9
and 4.1 mg/1, respectively. Thus, one conclusion drawn from these results
is that oil and grease are in greater concentrations in combined sewer over-
flows; however, they are better indicators of the presence of petroleum
hydrocarbons in stormwater discharges.
Oil and grease concentrations from numerous urban stormwater and com-
bined sewer sites across the nation are presented in Table 13. Also pre-
sented are the results from studies that measured specific petroleum hydro-
carbons including a site where PAH's are measured. The overall reported
ranges of oil and grease in urban stormwater and combined sewer overflows
are shown in Figure 5. Maximum oil and grease concentrations in the urban
stormwater have been reported to be greater, than-100 mg/1 while the CSO
concentrations have been reported to be greater than 1000 mg/1.
... Hunter et al. (30) measured urban stormwa.ter petroleum hydrocarbon con-
centrations from a 1520 acre (616 ha) catchment (multi- and single-family
residential land use) in Northern Philadelphia. The flow-weighted petro^-
leum hydrocarbon concentrations for 5 storms varied from 2.2 to 5.3 mg/1.
Peak concentrations were observed during the first portion of these storms.
These peak concentrations were about 4 times .greater than the flow-weighted
storm averages. Aliphatic hydrocarbons represented slightly over two-thirds
of the total petroleum hydrocarbons while the more toxic aromatic hydro-
carbons represented the remainder. Interestingly, about 86 percent of the
petroleum hydrocarbons was associated with particulates, which is encouraging
since solids separation/removal is less costly and easier than removal of
dissolved or fine colloidal matter.
A petroleum hydrocarbon mass balance estimation for Delaware Bay is
shown in Table 14. Urban stormwater runoff was estimated to contribute
significantly at about 10,500 Ib/day (4760 kg/day) or 17 percent (30).
However, after implementation of federally mandated effluent limits, urban
stormwater runoff will contribute the greatest portion estimated at 39
percent of the total.
As mentioned earlier, Wakeham (28) investigated petroleum hydrocarbons
in Lake Washington. Petroleum hydrocarbons in the sediments were found to
515
-------
§ 1
as ?S
§ ^
5 U-
o
£ !E
tu
O O
I" CT^
1 ^
§
S
i
UJ
fe
CO C
fe S
S i
rH
a
I—
j-
to
(U
^t IO O O CO m tn CM S-
CO CJ i-H rH (O
o.
in
3
u
s
c
0
i
t-1
«3
•£• s
.? 1
O-— « U' T-
i- W U- 01 CX
**r™ t *o -^ -2 2
I~*CL a) "§ •£»'& a) "o CL
O Vt O) 10 t/j Oi 4-* 3 ro
O. S* 3C W» «J "O
t/icnra CL 3c en ra 4- p r« 3 fl 3 gj nJ
r— O .C r— £- r- O. Ct S- •*-»
ggug-e^^-Cc^ .2
1- »-l CJ Q =3 3 »-« 1- Id VI *
fe~ £ 8 S ?- § •£ !t
£ OS «C -2: i-4 s: => o
s
i
«t
Wp 0
4J 6O
c 5§
t° -"H
»— * o
SC5 UJ
tn
i -1
3 CM S
i- a: t-i
(O X 15 C
^ S5 fJ
3 H* lif
S |5
ta P ""^ o.
i 52
O LU LU
O S3 p.*
""5 e*:<£
S W
rf S
i UJ
H- (/)
a
IT)
S.
00
CO
o
*,
Ol
in
1
o
1
3
t
*~* to
t5 c55
CM 0
•k
°^ M
uS ^^
t
1 B
V)
ra
t
3
01 V)
<" >p
g 1
to <->
ll
IE
*3 *S o a:
S S 'S-e
-1 -• as
8 8 o §
§ g r-'C
S 5 3,2
M- t|~ O J2
C C h-0
0 iS II II
* *
g
I
•i.
g I I
2
s
2
S § £
g fc
•o
o
s i c c a
S i 2- 2 1?
««•«—* **^ s—•
l~ «— O O O
J= N N N N
+* C C C c
c a) ej a* q;
01
¥
i
S g £ 2 S
>> OJ O O
-------
TABLE 13. PETROLEUM HYDROCARBONS AND HEM's
IN URBAN WET-WEATHER DISCHARGES.
Parameter* .
Total HC
Aliphatic HC
Aromatic HC
Aliphatic HC
Aliphatic HC
Total HC
Total HC
'PAH'S
Fl uoranthene
1,2-Benzanthracene
8enzo(a)pyrene
3, 4- Benzof 1 uoranthene
Chrysene
Anthracene
1 , 12-Benzoperylene
IndenoO ,2,3-c,d)pyrene
' Pyrene
HEM
HEM
HEM . '.
HEM.
HEM
HEM
HEM
HEM
HEM
HEM
HEM
HEM
HEM
HEM
HEM
HEM.
HEM
HEM
HEM
Average
Cone.
(mg/1)
3.69
2.57
1.12
12
1.2
10
0.9
0.098
0.050
0.170
0.052
0.050
0.027
0.041
0.086
0.076
12.5
20.0
12.4
11.5
9.5
. 9.5
18.2
14.1
17.8
34.4
11.9
11.7'
12.5
7.5
—
...
2.8
12.9 -
4.1
Range
(i»g/l)
2.18 - 5.30
1.50 - 3.5%
0.68 - 1.72
6-24
0.2 - 7.5
...
—
--
1.0 - J55.
0.5 - HO.
2.8 - JOO.
1.0 - 130.
0.2 - 52.
0.6 - 38.
1.0 - 140.
0. 1 - 110.
0.8 -1327.
3.1 - 423.
0.4 - 122.
2.2.- 63.4
6.6 - 21.5
4.5 - 17.9
1.3 -' 64.
26. - 95.
—
6.7 - 25.
'2.9 - 5.7
Type Location
Urban Stormwater Northern Philadelphia, Pa.
Highway/Bridge Runoff Seattle, Washington
Urban Stormwater
Highway Runoff Sweden
Urban Storawater
(multi-family)
Urban Stormwater Orlando, Florida
(Commercial)
Urban Stormwater. . Seattle, Washington
(residential )""
Urban Stormwater
(residential)
Urban Stoniwater
(Industrial)
Urban Stormwater
(Comnercial)
Urban Stormwater
(residential)
Urban Stormwater
(Low density residential )
Combined Sewer Overflow
(Comnercial)
Combined Sewer Overflow San Francisco, California
(Residential/Commercial )
Combined Sewer Overflow
(Residential/Industrial)
Combined Sewer Overflow
(Residential/Conercial )
Combined Sewer Overflow
(Residential )
Combined Sewer Overflow
(Residential/Commercial )
Urban Stornwater
(Residential/Conmercial)
Urban Stormwater
(Residential/Coumercial )
Combined Sewer Overflow Rochester, New York
Combined Sewer Overflow New York City. "e« York
Urban Stornwater
Combined Sewer Overflow Seattle, Washington
Urban Stormwater
Ref.
30
28
29
17
2
2
9
32
11
* HC * Hydrocarbons
HEM - Hexane Extractable Material
517
-------
10,000 -T-
p 1,000 --
i
01
in
>a
2
C3
-a
100 --
10 ..
1. _.
0.1.
cso
Urban
Stormwater
FIGURE 5. RANGES OF OIL AND GREASE IN URBAN WETrWEATHER DISCHARGES.
TABLE 14. PETROLEUM HYDROCARBON MASS BALANCE ESTIMATED TO DELAWARE BAY (30).
SOURCES
SPILLS
POTW'S
REFINERIES
OTHER INDUST.
URBAN RUNOFF
PRESENT
Mass Load*
(Ib/day)
6,000
7,800-15,000
24,000
8,900
10,500
Percent
9-10
14-23
37-42
14-16
16-18
AFTER- IMPLEMENTING
FEDERAL EFFLUENT LIMITS
Mass Load*
(Ib/day) Percent
6,000 23
2,000 8
1,900 7
6,200 23
10,500 39
* Ib = 0.454 kg
518
-------
correlate with the growth history of the surrounding metropolitan area.
Aliphatic hydrocarbons were sampled (grab samplets) in runoff from two
heavily traveled bridges (i.e., 50,000 automobiles/day). Concentrations
ranged from 6 to 24 mg/1 with an average of about 12 mg/1. An urban storm-
water runoff site had lower concentrations ranging from 0.2 to 7.5 mg/1.
Primary wastewater treatment plant effluent samples were found to contain
3 to 5 mg/1 of aliphatic hydrocarbons (wastewater treatment plant effluents
no longer discharged to the lake; however, the lake still receives combined
sewer overflows).
SUMMARY:
Urban stormwater discharges and combined sewer overflows have been
shown to contribute large quantities of oxygen consuming material, heavy
metals, and petroleum hydrocarbons which may adversely affect beneficial use
of receiving waters. Of course, actual site studies should evaluate the
impacts of urban wet-weather discharges considering other pollutant sources,
receiving water characteristics, and economic and financial analyses before
embarking on expensive control programs. Other pollutants should also be
considered in the analysis. Investigations into the impact of oxygen con-
suming material from urban areas should consider the variable nature of both
wet- and dry-weather discharges and the differences between the various
wasteload characteristics (e..g., decay rates, reaeration rates, settleable
oxygen consuming materials).
Several heavy metals have been shown to be in sufficient concentrations
in urban stormwater discharges and combined sewer overflows to exceed water
quality criteria for the protection of aquatic life even with many dilutions.
Near outfalls (before dilution and mixing can occur) metals wiTl cause bio-
logical stress areas as well. There are relatively strong indications that
the major receiving water impacts from heavy metals will occur in the sedi-
ments. Load comparisons show that urban wet-weather discharges contribute
more of several metals, such as cadmium, chromium and lead, to receiving
waters than POTW effluents.
Urban stormwater and combined sewer overflows have been shown to con-
tribute significant portions of oil and grease and petroleum hydrocarbons
to the environment. Petroleum hydrocarbons, most notably PAH"'s, are known
carcinogens; however, additional research is needed to determine the effects
of chronic exposure at relatively low levels (compared to major spills) of
petroleum hydrocarbons, particularly in the benthos. The contributions
in urban stormwater runoff are roughly equal to those of municipal wastewater
effluents. Together these sources account for at least 10 percent of the
petroleum hydrocarbons entering the ocean and a greater percent to in-land
and estuarine waters.
519
-------
REFERENCES
1. Colston, N.V. Jr., "Characterization and Treatment of Urban Land Runoff,"
EPA-670/2-74-096, December 1974.
2. Huber, W.C., J.P. Heaney, K.J. Smolenyak, and D.A. Aggidis, "Urban Rain-
fall -Runoff -Quality Data Base: Update.with Statistical Analysis,"
EPA-600/8-79-004, August 1979.
3. Lager, J.A., W.G. Smith, W.G. Lynard, R.M. Finn, and E.J. Finnemore,
"Urban Stormwater Management and Technology: Update and User's Guide,"
EPA-600/8-77-014, September 1977.
4. (Manning, M.J., R.H. Sullivan, and T.M. Kipp, "Nationwide Evaluation of
Combined Sewer Overflows and Urban Stormwater Discharges: Volume III,
Characterization of Discharges," EPA-600/2-77-064c, August 1977.
5. Heaney, J.P., W.C. Huber, M.A. Medina, Jr., M.P. Murphy, S.J. Nix, and
S.M. Hasan, "Nationwide Evaluation of Combined Sewer Overflows and Urban
Stormwater Discharges: Volume-II, Cost Assessment and Impacts," EPA-
600/2-77-064b, March 1977.
6. Keefer, T.N., R.K. Simons, and R.S. McQuivey, "Dissolved Oxygen Impact
from Urban Storm Runoff," EPA-600/2-79-156, November 1979.
7. Environmental Protection Agency, "Federal Guidelines State and Local
Pretreatment Programs, MCD-43," EPA-430/9-76-017a, January 1977.
8. Drehwing, F.J., A.J. Oliver, D.A. MacArthur, and P.E. Moffa, "Disinfec-
tion/Treatment of Combined Sewer Overflow, Syracuse, New York," EPA-
600/2-79-134, August 1979.
9. Drehwing, F.J., C.B. Murphy, Jr., S.R. Garver, D.F. Geisser, and D.
Bhargava, "Combined Sewer Overflow Abatement Program, Rochester, N.Y.:
Volume II, Pilot Plant Evaluations," EPA-600/2-79-031b, July 1979.
10. Innerfeld,H., A. Forndran, D.D. Ruggiero, and T.J. Hartman, "Dual Pro-
cess High-Rate Filtration of Raw Sanitary Sewage and Combined Sewer
Overflows," EPA-600/2-79-015, March 1979.
11. Tomlinson, R.D., B.N. Bebee, A.A. Heyward, S.G. Munger, R.G. Swartz,
S. Lazoff, D.E. Spyridakis, M.F. Shepard, R.M. Thorn, K.K. Chew, and
R.R. Whitney, "Fate and Effects of Sediments from Combined Sewer and
Storm Drain Overflows in Seattle's Nearshore Waters," Draft Report
EPA Grant No. R-805602, 1979.
520
-------
12. Betson, R., "Urban Hydrology: A Systems Study of Knoxville, Tennessee,"
TVA, June 1976.
13. Black, Crow and Edisness, Inc., et al., "Non-Point Pollution Evaluation
Atlanta Urban Area," Contract No. DACW 21-74-C-0103, May 1975.
14. Horkeby, B. and P. Malmquist, "Microsubstances in Urban Stormwater," In-
ternational Symposium on the Effects of Urbanization and Industrializa-
tion on Hydrological Regime and on Water Quality, Amsterdam, October
1977.
15. Pitt, R., "Demonstration of Nonpoint Pollution Abatement Through Im-
proved Street Cleaning Practices," EPA-600/2-79-161, August 1979.
16. Wanielista, M.P., Y.A. Yousef and J.S. Taylor, "Stormwater Management
to Improve Lake Water Quality," Draft Report, EPA Grant No. R-805580,
May 1980.
17. F.T. Brezenski, "Analytical Results - Orlando Florida," EPA Priority
Pollutant Analysis Region II Laboratory, February 1980.
18. Environmental Protection Agency, "Water Quality Criteria, Request for
Comments," Federal Register, Vol. 44, No. 62, March 15, 1979.
19. Environmental Protection Agency, "Water Quality Criteria: Availability,"
Federal Register, Vol.44, No. 144, July 25, 1979.
20. Gupta, M.K., E. Ballinger, S. Vanderah, C. Hansen, and M. Clark, "Hand-
ling and Disposal of Sludges from Combined Sewer Overflow Treatment:
Phase I - Characterization," EPA-600/2-77-053a, May 1977.
21. Salotto, B.V., et al., "Elemental Analysis of Wastewater Sludges from
33 Wastewater Treatment Plants in the U.S.," EPA-902/9-74-002, May 1974.
22. Pisano, W.C., G.L. Aronson, C.S. Queiroz, F.C. Blanc, J.C. O'Shaughnessy,
"Dry-Weather Deposition and Flushing of Combined Sewer Overflow Pollu-
tion Control," EPA-600/2-79-133, August 1979.
23. Bowan, H.J.M., Trace Elements in Biochemistry, Academic Press, New York,
1966.
24. Pitt, R. and M. Bozeman, "Water Quality and Biological Effects of Urban
Runoff on Coyote Creek," First Phase Report (In Press), EPA Grant No.
R-805418, June 1979.
521
-------
25. Wilber, W.G. and J.V. Hunter, "The Impact of Urbanization on the Dis-
tribution of Heavy Metals in Bottom Sediments of the Saddle River,"
Water Resources Bulletin, Vol. 15, No. 3, 1979.
26. Gross!ing, B.F., "An Estimate of the Amounts of Oil Entering the Ocean,"
in Sources, Effects and Sinks of Hydrocarbons in the Aquatic Environ-
ment, American Institute of Biological Sciences, August 1976.
27. National Academy of Sciences, E.B. Wilson (Ed.), "Petroleum in the
Marine Environment," Washington, D.C., 1975.
28. Wakeham, S.G., "A Characterization of the Sources of Petroleum Hydro-
carbons in Lake Washington," Journal WPCF, Vol. 49, No. 7, 1977.
29. Hallhagen, A., "Survey of Present Knowledge and Discussion of Input of
Petroleum to the Marine Environment in Sweden," Inputs, Fates and Ef-
fects of Petroleum in the Marine Environment Workshop, National Academy
of Sciences, 1973.
30. Hunter, J.V., T. Sabatino, R. Gomperts, and M.J. MacKenzie, "Contribu-
tion of Urban Runoff to Hydrocarbon Pollution," Journal WPCF, Vol. 51,
No. 8, 1979.
31. Feuerstein, D.L. and W.O. Maddaus, "Wastewater Management Program Ja-
maica Bay, New York: |