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 «§
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:  Volume I:   Summary Report," EPA-600/2-76-222a,
     September 1976.

32.  Shaheen, D.G., "Contributions of Urban  Roadway Usage to Water Pollution,"
     EPA-600/2-75-004, April 1975.
                                     522

-------
         Eighth' Session

      STORMWATER MANAGEMENT
Moderator:
Myron Tiemens
EPA, Washington, D.C.
               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. Welter
                             Dwight A. MacArthur
                       O'Brien & Gere Engineers, Inc.
                             Syracuse, New York
                                     and
                              Raymond P. Canale
                           University of Michigan
                             Ann Arbor, Michigan
Abstract

      Urban storm runoff has been determined by a number of investigations to
be a significant portion of the water pollution problem and abatement of this
source has been recognized as a necessary consideration in achieving the
national water quality goals of PL 92-500.  However, due to the highly ir-
regular nature of the runoff phenomenon, abatement'measures that address this
problem tend to be very expensive.  Thus, in the EPA construction grants
program great importance has been placed on careful  planning demonstrating the
cost-effectiveness of proposed projects (Program Guidance Memorandum 61).

      The benefits of runoff pollution abatement measures must be evaluated in
terms of projected receiving water quality conditions relative to some defined
quality standards.  There is considerable experience in the establishment of
water quality standards and resultant effluent limitations for non-transient
municipal and industrial discharges, generally through the use of mathematical
models.  Generally, these models are applied against a "design" receiving
water hydrologic regime (for instance, 10-year, 7-day low flow) and effluent
limitations selected which will meet the desired water quality standards.
This steady-state analysis assumes a toleration of water quality contraventioa
from the discharge of a low frequency corresponding to the designated low flow.

      This approach is not wholly adequate to analysis of the problem posed
by intermittent storm runoff discharges.  These analyses require alternate
modeling approaches and a restatement of pertinent standards to reflect the
short-term high variability of the storm runoff event.

      Using ongoing studies of combined sewer systems in Rochester, New York
and Washington, D.C., as case studies, a procedure is presented for the
application of receiving water analyses in urban runoff planning.  The studies
in both areas have included the development of water quality models based on
familiar concepts of mass balance and calibrated against detailed field sur-
veys and laboratory experiments.

      In a Rochester study, steady state models of dissolved oxygen in the
                                      524

-------
Genesee and fecal coliform concentration in the Rochester Embayment of Lake
Ontario are used to project receiving water conditions under dry-weather loads
and an envelope of expected impacts of the combined sewer overflows under
various system configurations.  A time-variable model  of Rock Creek and the
Potomac and Anacostia Rivers has been used to project transient fecal  coliform
nutrient and dissovled oxygen concentrations resulting from overflows  from
the District of Columbia sewer system.

      In an analysis of marginal costs and marginal benefits associated with
various combined sewer overflow abatement alternatives, a series of model
runs are made to project the impact of various waste discharges from the
alternate system configurations on the receiving waters under several
hydrologic regimes and ambient temperatures.  The conjoint probabilities of
the receiving water conditions and storm loads are determined on the basis
of historical records and used to project the expected water quality under
the alternative system configurations.  These projections can be quantified
as water quality improvements in terms of expected degree of contravention
of stream standards with regards to frequency, duration, area! extent  and
peak concentration.  On the basis of this analysis of project cost estimates
drawn from generalized cost curves, marginal .costs and benefits can be dis-
played graphically for each alternative.

      The analysis of the Syracuse, New York, Rochester, New York and  the
District of Columbia combined sewer facilities planning activities are dis-
cussed relative to this marginal cost-marginal benefit approach as case
studies.
  Introduction

       Urban storm runoff has been determined by a number of investigators
  to represent a significant portion of the water pollution problem and
  abatement of this source has been recognized as a necessary consideration
  in achieving the national water quality goals of PL 92-500.  However,
  due to the highly irregular nature of the runoff phenomenon abatement
  measures that address this problem tend to be expensive and difficult to
  size.  Thus, in the EPA construction grants program great importance has
  been placed on careful planning demonstrating the cost-effectiveness of
  proposed projects (PRM 75-34).

       The benefits of runoff pollution abatement measures must be evaluated
  in terms of projected receiving water quality conditions relative to
  some defined quality standards.  There is considerable experience in the
  establishment of water quality standards and resultant effluent limitations
  for non-transient municipal and industrial discharges, generally through
  the use of mathematical models.  Frequently, these models are applied
  against a "design" receiving water hydrologic regime (for instance, 10-
  year, 7-day low flow) and effluent limitations are selected which will
  meet the desired water quality standards.   This steady-state analysis
  assumes a toleration of water quality contravention for a low frequency
  discharge corresponding to flows less than Q7, 10.
                                      525

-------
     This approach  is not wholly adequate  to analyze  the problem
posed by intermittent storm runoff discharges.   These analyses require
alternate modeling  approaches and a  re-statement of pertinent standards
to reflect the short-term, high variability of the storm runoff event.

     Using ongoing  studies of combined sewer systems  in Rochester, New
York and Washington, D.C. as case studies, a procedure is presented for
the application of  receiving water analyses in urban  runoff planning.
The studies in both areas have included the development of water quality
models based on familiar concepts of mass  balance and calibrated against
detailed field surveys and laboratory experiments.  In the Rochester
study, steady state models of dissolved oxygen in the Genesee River and
fecal coliform concentrations in the Rochester Embayment of Lake Ontario
are used to project receiving water  conditions under  dry-weather loads
and an envelope of  expected impacts  of the combined sewer overflows
under various system configurations.  Results of the  dissolved oxygen
model for the Genesee River and fecal coliform model  for the Rochester Embay-
ment of Lake Ontario are presented.  Time-variable models of Rock Creek
and the Potomac and Anacostia Rivers are being developed to project
transient fecal colifornu algal nutrient,  and dissolved oxygen concentrations
resulting from overflows from the District of Columbia sewer system.

     In an analysis of marginal costs and marginal benefits associated
with various combined sewer overflow abatement alternatives, a series of
model runs are made to project the impact  of various waste discharges
from the alternate  system configurations on the receiving waters under
several hydrologic  regimes and ambient temperatures.   The    joint
probabilities of the receiving water conditions and storm loads are
determined on the basis of historical records and are used to project
the expected water quality under the alternative system configurations.
These projections can be quantified  as water quality improvements in
terms of expected degree of contravention of stream standards with
regards to frequency, duration,  area! extent,  and peak concentration.
On the basis of this analysis of project cost estimates drawn from
generalized cost curves, marginal  costs and benefits can be displayed
graphically for each alternative.

Case Study
Rochester,  New York

     As part of the Combined Sewer Overflow Facility Planning Activity
conducted for the Monroe County Division of Pure Waters,  O'Brien & Gere
Engineers,  Inc.  and Limno-Tech,  Inc.  conducted a water quality evaluation
program design to:

     A.    Establish the impact of existing combined sewer overflow
          pollutional  loads on the Genesee River and the Rochester
          Embayment of Lake Ontario.
                                    526

-------
     B.   Project the effect of improvements to the existing combined
          sewer system on the receiving waters.

     The combined sewer system of the Rochester Pure Waters District
serves nearly 65% of the city, with overflows which combined into 15
discharges to the Genesee River, two discharges to Irondequoit Bay and
one discharge to the New York State Barge Canal (Figure 1).  An overflow
monitoring and sampling program was initiated in early 1973 which involved
the measurement of flow and quality at 13 major overflow discharges and
three, collection system locations in conjunction with a rain gauge
network of 12 stations.  Table 1 presents the average constituent concentra-
tions for 5 of the parameters.                             ,     •',,,,

     From the flow data measured during the 1975 monitoring program it
had been estimated that the average overflow event involves a discharge
of 52.8 MG of CSO generated wastewater to  the Genesee.  Utilizing the
analytical data presented in Table 1, it has been determined that an
average summer overflow event results in a.,discharge of the following
constituent loadings to the Genesee ,River (1):.
     Biochemical Oxygen Demand
     Total Suspended Solids
     Total Inorganic Phosphate
     Zinc
     Lead      .  '  ,.
     Cadmium
 64,920 pounds/storm
136,860 pounds/storm
  ,  540 pounds/storm
    110 pounds/storm
     90 pounds/storm
    1.3 pounds/storm
     Modeling the/existing combined sewer network utilizing the Simplified
Stormwater Model  (SSM) projects that the combined sewer system will
contribute 42.7 overflow events during an average year.  Utilizing this
frequency factor and the derived average storm loading, it is possible
to estimate the following annual combined sewer loadings'(1):
     Biochemical Oxygen Demand
     Total Suspended Solids
     Total Inorganic Phosphate
     Zinc
     Lead
     Cadmium
 2,772,050 pounds/year
 5,844,040 pounds/year
    22,850 pounds/year
     4,870 pounds/year
     3,750 pounds/year
        54 pounds/year
     To compare the annual impact of the Rochester combined sewer system
with the total load carried by the Genesee River, the following annualized
load of the Genesee River is presented as follows:
                                    527

-------
528

-------
             00

             O
             Q£
             LU OO

             O O
                        CO
                        Q.
Q
O
                1-1      O
             oo 2:
LU    lf>       U-l
—1    t~*    Q O
en    cr>    LU z:

             I- co oo'oo ro oo ro co CM    CM
ooooooooooo    o
ooooooooooo    o

ooooo'oooooo'    o




cr>cr>ocr»cT»cricr>cy>oooor^    «-<
OOt-tOOOOOOOO    t-l

o'ooooooo'o'oo    o


         t^i^r^cooovDLOLh    oo
CMOOCMOOOOOOOO    O
O O O O O O O O O O O    C5

oo" oo oo o o'o o" o    o

                                   cOi— CM
t-HCMlT!i«3tr)LOCOr^CMOVDJ2Cy>
OOOOOOOOOOOOCOOOCMCMi—t  reJCM
      co oo oo CT»
      en cr> en cr»
                               IO  > OO

                              - «vt-  tO If)
                                            ta
           i oo oo
               oo co tn oo CM o
               coooinL
              it— ICMCM
                                            ta
                                           -a
                                            oio
                                                                i-<
                                                     o
                                                          •V
                                                     s-  w
                                                     

                          •a:
          S-
          O
         +->
          O
          ta
         u_
•o  c  cu
O  O -O T3 -E
O •»->••-
                            oo n3
                                    e
                                    T-4->r—
   -cr           cu >> cu
   4->           vi ca cu
    cu           =5    s-
   oo  cuEnJ
 O -I-2 4-^  ^ S- CU CO E
                                                   CMCMCMCMCMCMOO
                                                                             C3J
                                                                             to
                                                                             cu
                                                                             3
                                                                            LU


                                                                            CD
                                                     cu  c
                                                  «J C LU
                                                  +i -I-
                                                    E
                                                  i— uj  cu
                                                  jQ CU -I-
                                                  =?^:  s-
                                                  a. o ca
                                                  c: o -
                                                  ra a: o
                                                                                   LU
                                                                                   o
                                                                                   oo
                                                        529

-------
     Biochemical Oxygen Demand
     Total Suspended Solids
     Total Inorganic Phosphate
     Zinc
     Cadmium
  5,271,700 pounds/year
683,606,680 pounds/year
    181,040 pounds/year
     91,980 pounds/year
     11,680 pounds/year
     The projections indicate that the combined sewer system contributes
only a fraction of the annualized loading of the Genesee River to Lake
Ontario for parameters measured except for Biochemical Oxygen Demand.
Even though most of the loading projections  are quite low, the CSO
impact is of a very short and intense nature.  This transient wet-
weather impact creates significant adverse stress on the aquatic ecosystem
which occurs fairly frequently on an annual basis.  However, its mark is
also left in a less intensive but still significant magnitude in the
form of particulate pollutants contributed to the benthos in the quiescent,
lower reaches of the Genesee River.

     In light of this setting, it was determined that the likely transient
water quality impacts induced in the lower reaches of the Genesee River
involve short-term dissolved oxygen and fecal coliform water quality
violations.  The most likely transient water quality contravention
within the Rochester Embayment involves elevated fecal coliform levels
in the vicinity of near shore beach areas. This is supported by an
incomplete historical water quality data base which illustrates a condition
of variable and elevated indicator organism densities in both the Genesee
River and Rochester Embayment as well as frequently depressed dissolved
oxygen concentrations within the water column of the lower reaches of
the Genesee River.

     The seven historical samplings of the sanitary conditions of the
Genesee River dating back as early as 1912 support the initial water
quality impact assessment.  However, the historical data are not adequate
with respect to spatial detail and lack associated discharge information
needed to make a quantified water quality assessment.  Therefore,
O'Brien & Gere Engineers performed four extensive water quality surveys
during 1975, two of which were during storm events (2).

     The impact of combined sewer overflow discharges on the water
quality of the Genesee River and the Rochester Embayment of Lake Ontario
is illustrated by the water quality data collected during the November
10-12, 1975 wet-weather survey. The overflow event was initiated by
rainfall of 0.38 inches over a span of 8 hours and was estimated to
result in peak CSO discharge rates of 154 cfs, receiving water flows of
2300 cfs and water column temperatures of 13°C.
                                   530

-------
     Figures 2-4 illustrate the minimum, average and maximum water
column parameter concentrations at each of the sampling stations for the
duration of the survey. As an aid in the interpretation of the water
column response data, one should realize that the significant CSO discharges
are located along the region from milepoint 5.3 to milepoint 10.7.

Genesee River Models

     The wet and dry-weather data sets were used to calibrate and verify
three water quality models developed for the Genesee River and the
Rochester Embayment. These involved steady state dissolved oxygen and
fecal coliform models for the Genesee River and a steady state fecal
coliform model for the Rochester Embayment of Lake Ontario.   The basic
model framework used to simulate water quality in the Genesee River was
a Limno-Tech developed modification of the U.S. Environmental Protection
Agency steady state AUTO-QUAL modeling system.  The model produces a
steady state uni-directional description of rivers using second-order,
finite difference approximations.   The structure of the model is based
on the simple concepts of continuity, including terms for advection,
dispersion, reaction, and point or non-point sources or sinks.   The
model can calculate the steady-state spatial distribution of any water
quality parameter whose kinetic behavior can be described by zero or,first-
order kinetic formulations or described as conservative.  In the case of
the Genesee River, the parameters modeled included chloride, fecal
coliform, carbonaceous oxygen demand, nitrogenous oxygen demand, and
dissolved oxygen.  Details concerning the model framework are presented
in Table 2.

Rochester Embayment Model

     A review of the historical data base developed on the Rochester
Embayment of Lake Ontario indicated the necessity to collect wet-weather
field data to serve as the basis for developing a fecal coliform model.
Fecal coliform was selected as the water quality reference parameter
because of a long history of variable sanitary conditions exhibited at
existing and potential beach areas within the Embayment.
     Because extensive data did not exist for the Embayment Fecal Coliform
model, a three day investigation of coliform and chloride conditions in
the Embayment was conducted.  During a 72 hour period following the
November 10, 1975 storm, O'Brien & Gere Engineers, Inc. monitored the
water quality at ten stations as well as all known sources of Embayment
contamination.  Figures 5-8 present the fecal coliform and chloride data
measured on November llth and 12th.   The total measured loading to the
Rochester Embayment from the Genesee River for November llth and November
12th are 2.48xl014 cells/day and 7.84xlOid cells/day, respectively.
                                   ,531

-------
                       FIGURE 2
00*0081    OO'OOSI    00*0021     00*006    00*009     OO'OOE
     (E-OIX)  HW OOt/SlNnOO)  NOIiUyiNBONOG 1d03J
                                532

-------
                                                                  o
                                                                  o
                        FIGURE 3
                                                                 ..(O
                                                                  o

                                                                 ..<\J
       ee.


     LOUJ


     II
       LU
       ts
OQ'3l
                                                                    UJ
                                                          LU
                                                          CJ
                                                                    tfl
                                                                    I—«
                                                                    a
                                                                  O
                                                                  o
                                                                   S
00*91      00"Et       00'6       00'9

         n/3W)  NOIiHUiN33NOO  £009
CO'E
00 0°
                                 533

-------
                                          o
                                          o
FIGURE  4
                                          Tocn
                                          om

                                         -sri
                                     O)
                                     c in
                                     •r- 
-------
Type:
                        TABLE 2

6ENESEE RIVER WATER QUALITY MODEL TECHNICAL INFORMATION


  Steady State
Predictive Capability:  Chloride, Fecal Coliform, Carbonaceous Oxygen  Demand,
                        Nitrogenous Oxygen Demand, and Dissolved Oxygen

Basic Model Framework:

          i£ = n = 1  1   AE1£    1  JL  (Qc) + r
          3t       A  ax    3x  ~ A  9X

Dissolved Oxygen (DO) Differential Equation
  9DO -0=I1
  3t        A 3X
                               9 DO - I  1  (QDO)-KrcCBOD-KrnNBOD
                               3X    A  3
                  + Ka(DOs-DO) + I  (P-R-B)
                                 H

Fecal Coliform (N) Differential Equation

      :  .  M = n = I  -L_   AEM _ I • J_  (QN)-Kf (N)
          3t       A  3x     3x " A  3x

Conservative Substance (C) Differential Equation
  oC „ r\ _. 1  _o	   r\L- ov*   x
  ¥t ~   ~ A"  ax     ix " A"  3x
                                        (Qc)
     where
               c = concentration of pollutant
               A = cross sectional area of channel
               E = longitudinal dispersion coefficient
               Q = flow rate
               r = net rate of biochemical reaction
               t = time
               x = longitudinal distance
               P = rate of photosynthesis
               R = respiration rate
               B = benthal demand
              D0s= oxygen solubility
               H = depth
               Ki = first order reaction rates
                                    535

-------
536

-------

              ^•ff| ^^..^^SSfl:!;     f^SEJt^r-S:*
                £. ^^

              f\ • ». J fe .  .?---. -+- • e «-^%-^ =M '•>- V^T^wRz.^:. -..  *•
p^-^-y ;i i^.a4t.-"l ^-t
- I    «i = L -'- .'TI--- ' , 11 fSf^
= *    ;1  r«  •  .i.    ifli^
                            537

-------
                                                             -*>U&'. S.M < =T ;-.•*•?
-».  •          »   ;;   •   i

^H^^Ti  ;  V 31"
=   "     «'           '
                                    538

-------
539

-------
     The conceptual basis for the embayment model is very similar to
that of the Genesee River model.  The concept of conservation of mass is
intrinsic in the model. The two mechanisms which determine the fate of
material in the system are transport and reaction.  For the general case
these effects can be described by the equation:

accumulation = net input of mass + net input of mass + rate of disappearance
of mass        from convective     from dispersional   or appearance of
               transport           transport           mass due to reaction

Mathematically the continuity of material can be expressed as:

     jte =  V -(Eve)- ? (uc) +r
     3t

where  u is the velocity

     The embayment model differs from the Genesee River model in that
the transport terms for the embayment necessitate a three dimensional
description. Therefore it was necessary to approximate the continuous
system with a system of completely mixed segments which allow for advective
and dispersive flow among adjacent cells.  A system of 132 cells was
used to describe the embayment. The model has 87 surface segments which
describe the lake to a depth of 20 meters and 45 bottom layer cells
which describe all waters below  the 20 meter depth contour.   Six
additional one-mile segments for the lower reaches of the Genesee River
were included in the embayment model to adequately handle river-embayment
interactions from upstream loads to the river.  The modeling techniques
used are similar to those described in reference (3).

     One of the most important steps in developing a three dimensional
water quality model is the identification and quantification of the
hydraulic circulation within the system of study.   A report prepared by
Bonham-Carter, Thomas, and Lockner (4) which was generated as part of
the International Field Year for the Great Lakes investigation on Lake
Ontario presents a numerical model which calculates three dimensional,
steady-state water transport on the embayment under unstratified
conditions.   The results of the IFYGL circulation model  were not available
in numerical form but rather were depicted in graphical vector form.
                                   540

-------
Vector  analysis was used to transform the graphical information into a
numerical scheme suitable for use in the water quality model.   Circulation
pattern predictions were available for steady winds from  the north,
south,  east, west, northwest and southwest directions.

     The development of the embayment fecal coliform model required an
accurate quantification of the die-off coefficient.   This information
was derived from laboratory bacterial die-off studies performed specific
to the Genesee River -Rochester Embayment system.   The decay coefficients
used, in model simulation were consistent with the ambient water temperature.
For projections a die-off coefficient of 1.3 day"1; was used for warm
water conditions.

     The steady state embayment model was verified against four sets of
coliform data (July 18, 1912; August 2, 1912; November 11, 1975, and
November 12, 1975) and two sets of chloride data (November 11, 1975 and
November 12, 1975.).  Model verification of  the chloride data is useful
because chloride is considered to be a conservative substance and hence
provides an excellent means for testing the model's advective and dispersive
transport terms without the complication of reaction processes.  Figures
5-8 show  the chloride and fecal coliform model projections compared to
the actual data for the November 11, 1975 and November 12, 1975 data
sets.

     It is obvious that steady state models for the River and Bay cannot
accurately describe time variable changes in water quality due to overflows
of combined sewage.  Therefore model calibration is qualitative.  However,
steady state models can be used to calculate the maximum response expected
from CSOs (assuming steady maximum loads).  Therefore steady state models
can be used to compare and select planning and management alternatives for
CSOs.   The next section which deals with similar problems in the Washington
DC area will utilize dynamic models which are calibarted under wet weather
conditions.

Abatement Alternative Analyses

     An extensive sensitivity analysis was conducted in an effort to
define the critical  receiving water conditions which result in the
simulation of maximum CSO impact.   The following represents a listing of
parameters for which the sensitivity analysis was conducted:

          Genesee River Fecal Coliform Model
               receiving water flows
               water column temperatures
               dispersion rates
               discharge location
               fecal coliform loading
                                   541

-------
          Genesee River Dissolved Oxygen Model
               receiving water flows
               water column temperatures
               dispersion rates
               reaeration rates
               deoxygenation rates
               BODc and BOON loading
               benthic oxygen demand rates
          Rochester Embayment Fecal Coliform Model
               wind direction and intensity
               water column temperatures
               fecal coliform decay rates
               fecal coliform loading
     As a result of conducting the sensitivity analysis, it became quite
apparent that the receiving water impact was as dependent on the pre-
event water column conditions as it was on the actual CSO loading.  Most
of the significant determining variables for the Genesee River fecal
coliform and dissolved oxygen models are related to the seasonal conditions
generally characteristic of the Genesee River.   Similarly, the frequency
and magnitude of CSO discharges is statistically  related to the seasonal
reference.

     Figure 9 presents the Genesee River critical dissolved oxygen
projection for a single treatment plant treating at various combined
sewer treatment rates and different treatment efficiencies.   From this
analysis it can be seen that only 25 MGD of CSO treated at an 85% removal
efficiency may be discharged to the: Genesee River under the analyses for
critical low flow conditions and maximum water column temperatures
without violating the water column dissolved oxygen level of 4.0 mg/1.
Figure 10  presents the same CSO loading  and treatment efficiency
response under average summer flows.  Under average summer flows 25 MGD
of untreated CSO may be discharged without any treatment and still a
dissolved oxygen concentration of 4.0 mg/1 will be maintained.

     As a result of these analyses and average summer Genesee River flow
vs. loading analysis it was concluded that the combined sewer overflow
treated discharge should be directed to the Rochester Embayment of Lake
Ontario via conveyance to the Frank E. Van Lare sewage treatment plant
for treatment utilizing existing and supplemental treatment facilities.
The question that remained involved the sizing of the required conveyance
and in-line storage facilities.

     The resulting benefit framework became the frequency of dissolved
oxygen violations in the Genesee River and the frequency of beach closing
days projected for Ontario Beach which is located on the Rochester
Embayment of Lake Ontario.

     Six structurally intensive alternatives and a series of source and
collection system management Best Management Practice abatement alternatives
were evaluated in the course of developing a Master Plan for the Rochester
combined sewer system.  These alternatives are outlined as follows:
                                    542

-------
                                                         Figure 9
                  CRITICAL   DO  SAG   PROJECTIONS





          Alternative1   Single  Treatment  Plant


          Flow  Regime: MA7CD/IO
   5.0 ••
                                             95 %  Removal
   4.0--
 o>

 E
   3.0--
O
a
cc
o
   2O- -
   I.O--
                                    60 % Removal
                      vNo  Treatment
                           5O                   10O



                                DISCHARGE  (MGD)
150
                                   543

-------
                                                          Figure 10
                  CRITICAL  DO  SAG  PROJECTIONS






            Alternative:   Single Treatment  Plant


            Flow Regime: Average  Summer
   6.0-r
   5.0- - —
   4.O--
                                                       95 %  Removal
Ill

I
o
a
cc
o
   3.O--
   2.0- -
   1.0- -
                           50                   100




                            DISCHARGE   RATE   (MOD)
I5O
                                  544

-------
Structurally Intensive Alternatives
     1.   Capture of first flush from the River overflow sites and
          treatment of all post first-flush flows with primary
          swirl devices.

     2.   Various storage and treatment options for capturing and
          treating the total overflow to the Genesee River with
          treatment facilities located on the Genesee River.

     3.   Various storage and treatment options for capturing and
          treating the  total overflow to the Genesee River with
          treatment facilities located at the present Van Lare STP
          and discharge to the Rochester Embayment of Lake Ontario.

     4.   Capture of first flush from the River overflow sites and
          discharge of untreated post first flush flows.

     5.   Use of primary swirl devices on each of the River overflow
          locations for  treatment of the entire overflow volume;
          the swirl effluents being discharged directly to the
          Genesee River.

     6.   Storage and treatment balance for intercepting the River
          overflows and conveyance of the overflows to the Cross-
          Irondequoit Tunnel for treatment at the Van Lare STP.

Best Management Practice Options

     1.   Interceptor improvements to improve the conveyance capacity
          of the existing St. Paul Blvd. Interceptor.

     2.   Regulator modifications to optimize the conveyance of
          wet-weather flows utilizing the upgraded interceptor
          capacity.

     3.   Selective regulation of high impacting overflows.

     4.   Implementation  of control structures to optimize the
          utilization of the available in-system storage capacity.

     5.   Optimal utilization of existing street sweeping and sewer
          cleaning operations.

     6.   Modification of the F.E. Van Lare existing treatment to
          allow split flow operation.  Under wet-weather conditions,
          flows in excess of the process capacity of the secondary
          facilities receive chemical assisted primary treatment
          and are blended with the secondary treated flow prior to
          discharge.
                              545

-------
     On  the  basis  of either  water  quality  limitations  or  a  cost-effectiveness
 analysis,  the  structurally intensive  alternatives  presented as  1,  2,  4,
 5 or 6 were  eliminated  from  further consideration.   Structurally  intensive
 alternative  3  was  selected for  further  consideration.  A  marginal  benefit/
 marginal cost  analysis  was conducted  on alternative  3.  The EPA developed
 SWMM II  model  and  the Dorsch Consult  model  were  utilized  to project the
 hydraulic  loading  from  the existing system and the storage  and  treatment
 balance  projected  for various storage capacities.  Figure 11 presents
 the number of  Genesee River  dissolved oxygen  contraventions as  a function
 of the effective in-line  storage capacity.

     In  the  development of the  Master Plan for the Rochester combined
 sewer system the pertinent Best Management Practice  alternatives have
 been identified for  implementation followed by the implementation  of
 the structurally intensive alternative as developed from the net a C/B D.O,contra-
vention  analysis presented in Figure  11. The  reduction in  the  number of
 beach closing  days at the Ontario  Beach and the  number of dissolved
 oxygen contraventions encountered  in  the Genesee River were plotted as a
 function of  program  implementation.   The  improvement  in  water  quality
 resulting  from implementation of the  Master Plan is  presented in the
 form of  Figures 12 and  13.

 Case Study
 Washington,  D.C.
     As part of the  Phase  I  Combined  Sewer Overflow  being  conducted on
the Potomac-Anacostia  River  System  for  the  Government  of  the  District
of Columbia, receiving water quality  investigations  were conducted to
provide a basis for  assessing the benefits of  any modifications of the
structural or operating characteristics  of the combined sewer  system
proposed for the purpose of  CSO pollution abatement.    The water quality
investigations in this study were focused on Rock Creek, Potomac River,
and Anacostia River  all of which receive significant CSO discharges from
the District.  Figure  14 illustrates  the study area,  the areas of significant
CSO discharges and the park  land or recreational space  adjoining the
water front.

     Reconnaissance  surveys  conducted during the Phase  I program included
both dry and wet-weather water quality sampling and  sediment analysis in
the Potomac and Anacostia  Rivers and  Rock Creek.  The water quality of
Rock Creek was sampled for dissolved  oxygen, fecal coliform, and suspended
solids thirteen times  during dry weather between July and  October 1978
and intensively once subsequent to a  rain event.  Rock  Creek sediments
were sampled once for  concentrations  of  selected heavy  metals  and chlorinated
pesticides, and once to assess the macroinvertebrate community.   The
Potomac and Anacostia  Rivers were surveyed once during  wet weather
conditions for dissolved oxygen, fecal coliform and  suspended  solids,
and also  for sediment oxygen demand  and sediment concentrations of
selected heavy metals, chlorinated pesticides,  and oil  and grease.
                                    546

-------
                              o
S4U9A3
          547

-------
I O

.U
   c
   O
 c  c
 Ol  O)

 o  *••

0-  3i

    c  o

£ •£  ^



 c £  S
 o ^  *^
•r*    ^^
    QJ
U




-------
 (A  V)

 >,  (U
 O.J


'
    OJ



    O)
JO

O




I  -

<8 <


0 H-


s °
C  C
O .2
 o  a>
| -a
a. Ji


£ |2
 o  v>

 o  o
 =>  Q.
TO  VI
 01  
Q: a:
o
H
PH
                                                  (VI
                       SAWQ    9NIS010    HDV3B

                          •IVIJLNBlOd   39VH3AV
                                     549

-------
                                            FIGURE 14

  RECEIVING  STREAM  STUDY AREA
                                     National  Zoological Park
Water Pollution
Control Plant
     Stream Bank Area With Combined Sewer Outfalls

^q  Contiguous Park  Land or Recreation  Space
'     Adjoining  the Water Front
                        550

-------
     A preliminary and simplified water quality model  was used to
confirm the observed water quality impacts of combined sewer overflows
and establish the potential significance of other levels of combined
sewer overflow loading.  Model calculations indicate the combined sewer
overflows can cause transient DO deficits of 3 to 5 mg/1 under a variety
of loading and environmental conditions.  A simple Lagrangian framework,
finite difference model of CBOD, NBOD, and dissolved oxygen was used by
Limno-Tech in this analysis (5).  The model inputs for physical data and
kinetic expressions were determined from analysis by the Annapolis Field
Office of EPA (6,7)._

Water Quality Impacts:  Potomac and Anacostia Rivers

     Analysis of historical data, reconnaissance data, and preliminary
model calculations were used to relate water quality problems to combined
sewer overflows.  Water quality problems in the Potomac and Anacostia
Rivers are similar and, many can be partially attributed to combined
sewer overflows.   Historical  monitoring data for the Potomac and Anacostia
Rivers revealed significant decreases in dissolved oxygen  and increases
in fecal coliform concentrations subsequent to rainfall events.   Algal
concentrations in the Potomac River were high and growth is potentially
being locally stimulated by nutrients from combined sewer overflows.
Estuary sediments were moderately polluted with respect to sediment
oxygen demand and selected heavy metals concentrations.

     Eight of twelve storm related data sets identified from the Potomac
River monitoring revealed significant reductions in dissolved oxygen
concentration subsequent to rainfall events, while seven of nine revealed
increases in fecal coliform concentrations..   In the Anacostia River
five of seven storm related data sets showed decreases in dissolved
oxygen and six of seven showed increases in fecal coliform.  The existence
of these impacts in water quality was further supported with reconnaissance
data and preliminary model calculations.  Overall, combined sewer overflows
were shown  to potentially decrease Potomac and Anacostia River dissolved
oxygen concentrations by as much as 3 to 5 mg/1  and increase fecal
coliform concentrations in excess of 200,000 organisms/100 ml.   Water
quality  in the Anacostia River was also persistently poor during dry
weather, even though the river receives no major continuous wastewater
discharges.  Anacostia River water quality was also observed to be
impacted by storm runoff in areas above the discharge of District combined
sewer overflows.

     Routine monitoring data showed high and variable concentrations of
chlorophyll a^ in the Potomac River, near Washington, D.C. often ranging
from 50 to 100  ug/1 while concentrations exceeded 200  g/1 in areas
downstream.  A comparison of the relative loadings and preliminary model
calculations showed that combined sewer overflow nutrients could cause
significant localized increases in Potomac River algal concentrations.
Combined sewer overflows from a once in two year storm during the summer
can increase nitrogen loading in the upper Potomac River by approximately
30% and phosphorus loading by 150%.  On a year averaged basis the combined
sewer overflow contribution was estimated to be approximately 1% for

                                   551

-------
 nitrogen and approximately 10 to 15% for phosphorus.   Preliminary model
 calculations indicate that combined sewer overflow loading  can  significantly
 increase local  river nutrient concentrations  and  that  these increases
 can  potentially increase algal  concentrations beyond 100  mg/1 chlorophyll-
      Measurements  of sediment  conditions  in  the  Potomac  and Anacostia
 Rivers  revealed moderately  polluted  conditions which may be partially
 caused  by  combined sewer overflows.  Measured sediment oxygen demands
 averaged 2.4  g/m2/day and were elevated over values measured in the
 Potomac River upstream of   the combined sewer overflows.  Measurements
 of selected heavy  metal conconcentrations  in sediments indicated moderately
 to heavily polluted  conditions according to  EPA  sediment pollution
 guidelines.   Concentrations of lead and zinc were particularly high.
 No significant pollution was observed with respect to sediment concentration
 of oil  and grease,  and chlorinated pesticides.

 Water Quality Impacts:  Rock Creek

     Analysis of historical and reconnaissance data from Rock Creek show
 a significant relationship  between storm events  and increased fecal
 coliform concentrations and a  possible relationship to decreases in
 dissolved  oxygen.  Six of eight storm related data sets  showed dramatic
 increases  in  fecal coliform subsequent to storm  events but only six of
 eleven data sets revealed decreases in dissolved oxygen.  Significant
 storm- related water  quality impacts were also observed in Rock Creek
 upstream of the District combined sewer overflows.

     Rock  Creek sediments were measured to be moderately  polluted although
 the extent of these  sediments was limited.   Again, as in the Potomac
 and Anacostia Rivers, lead and zinc concentrations were measured at
 polluted levels.  Sediment concentrations of oil  and grease and chlorinated
 pesticides were not  in the polluted range.   A characterization of  the
 Rock Creek benthic macroinvertebrate community indicates  a predominance
 of pollution  tolerant organisms.  No quantitative relationship could be
made in the preliminary study  between sediment pollution  and combined
 sewer overflows, although such a relationship may exist.  Concentrations
 of some heavy metals were observed to approximately double downstream of
 District combined sewer overflows.

 Preliminary Summary  of Impacts

     Figure 15 presents a summary of water quality issues as they relate
 to the discharge of  combined sewer overflows to  the Potomac River,
 Anacostia  River and  Rock  Creek.   Analysis  of historical and recent
water quality data as well as  preliminary calculations indicate the
 District combined sewer overflows significantly  impact dissolved oxygen
 and fecal  coliform in the Potomac and Anacostia  Rivers and potentially
 impact algal  and sediment concentrations.    District combined sewer
 overflows  also have  potentially significant  impacts on water quality in
 Rock Creek, although it is difficult at this time to differentiate these
 impacts from  other upstream storm related impacts.

                                  552

-------
                                                        FIGURE   15
      SUMMARY  OF  WATER  QUALITY  ISSUES
POTOMAC  RIVER
ANACOST1A  RIVER
ROCK CREEK
                                  Q


                                  O
                            LEGEND




        Major  Significance - Strong  Evidence of a  Significant Impact  From

                          Combined  Sewer  Overflows
_     _.   ...
Some  Significance -
                           More  Limited  Evidence  of  Significant  Impact
                                                     *

                           From  Combined  Sewer Overflows
        No  Significance- No  Evidence  of  a  Probtem  Area  Significantly


                       Impacted  by the  Combined Sewer  Overflows
                                553

-------
      It  has  been  recommended  that  a  program  of wet weather water  quality
monitoring be  conducted  on all  three receiving waters along with  those
special  field  studies which will be  used  to  develop mathematical  water
quality  models. The Phase II  program involves the development of  Rock
Creek fecal  coliform models as  well  as the adaptation of  the available
Dynamic  Estuary Model to predict the fecal coliform, dissolved oxygen
and nutrient levels within  the Anacostia and Potomac Rivers.  These
models will  provide a quantification of the  incremental improvements in
water quality  resulting from  various  combined sewer overflow control
alternatives.

Methodology of Water Quality  Utilization

     Now that  the application of water quality models has been presented
for the  Rochester, New York and Washington,  D.C. combined sewer overflow
planning studies, it would be useful  to summarize a methodology which
has evolved from conducting these  studies.   Figure 16 presents a  summary
of 17 steps which constitute  the process leading from the initial review
and compilation of the available water quality data base to the marginal
benefit/marginal cost analysis  and sizing of the most cost-effective
solution.
     In the initial phase of the analysis, considerable effort is required
in the definition of the baseline water quality limitations and the
selection of appropriate water quality parameters which can be utilized
to define the established limitations.   Based on this initial definition
a limited wet-weather impact survey should be conducted to confirm and
further define the urban runoff induced water quality limitations.  In
conjunction with this limited wet-weather impact survey the available
water quality models should be reviewed and impact-determining receiving
water characteristics (low flow, tidal influence, unique sources and
sinks, etc.) defined.

     After the initial definition of the system, decisions relative to
the impact analysis framework should be resolved.   If sufficient data
are available, a regression relationship may be developed which relates a
water quality parameter to several determining variables.  As an alternative
a deterministic rather stochastic model may be required.  A limited
data base may require the acquisition of water quality impact data;
the specific parameters and the extent of  the data base is largely
determined by the complexity of the receiving water system and type of
modeling (steady state, dynamic, etc.).  The collected data are required
for calibration and verification of the selected modeling technique.

     Following selection, calibration, and verification of the  water
quality analysis framework, simulation of both present and improved
water quality conditions may be conducted given the urban runoff loadings
projected for existing and improved conveyance and treatment systems.
A method of simulating the stormwater and/or combined sewer loadings is
required.  Available conveyance network simulators include Storm, the
Simplified Stormwater Model (SSM) and  the various versions of the EPA
                                   554

-------
                                   FIGURE  16

           METHODOLOGY OF WATER QUALITY MODEL UTILIZATION
           FOR CSO MARGINAL  BENEFIT-MARGINAL COST  ANALYSIS
                            Compile and  Review the
                       Available  Water Quality Data'Base
                                 Define  Present
                            Water Quality  Limitations
                        Select Water Quality  Parameters
                            Which Define Limitations
Review Available Water
Quality Models	
      Initiate Limited
Wet Weather  Impact  Survey
   Define Receiving Water
Determining Characteristics
                          Establish Model or Statistical
                         Framework for Receiving Water
                            Quality Impact Analysis   •
Develop  Reference
Storm  Hyetographs
                                   Determine the Benefits of
                                Single  Event Simulation
  Collect  Calibration  and
     Verification Data
Conduct Network Simula-
tionston Existing
,System and Alternatives
Simulation of Water Quality
 Response Associated with
 Alternative Implementation
     Establish Critical
      Receiving  Water
      Characteristics
                          Prepare Alternative Cost VS
                           Receiving Water Response
                         for Various Reference Storms
                                      Estimate Alternative     :
                                  Capital  and Operating  Costs |
                         Relate  Receiving Water Quality
                          Response to  Water Resource
                                  Enhancement
                           Conduct Marginal  Benefit-
                             Marginal Cost Anaysis
                                          555

-------
Stormwater Management Model (SWMM) which can be well adapted to specific
requirements to provide a dynamic or steady state loading.

     Coincident with the development of system loadings is the necessity
to establish the critical or representative water quality conditions
(water column temperature, receiving water flow, tidal cycle, etc.).  It
may be necessary to statistically analyze the probability of occurrence
of defined critical receiving water conditions along with the probability
of generating a rainfall event of given intensity and duration.  Establishment
of historical critical receiving water conditions as the basis for water
quality simulations can result in a significant overstatement of both
the problem and the facilities required to solve the problem.

     Simulation of the water quality response associated with alternative
implementation can be accomplished given the defined receiving water
conditions, urban runoff generated loadings and the calibrated water
quality simulator.  A progressive development of improved water quality
parameters along with the associated capital and operating costs allow
the construction of a cost curve vs water quality improvement for various
sized facilities.  From this cost vs water quality index curve the
marginal increase in receiving water benefit can be associated with an
increase in marginal cost.  Utilizing reasonable judgement, the point
where marginal benefits outstrip marginal costs can be assessed within a
band of accuracy established by the various simulators.
                               REFERENCES

1.  Drehwing, F.J., et al. s Combined Sewer Overflow Abatement Program.,
    Rochester, H.Y., Volume I5 EPA Grant No. Y005141, 1979.

2.  O'Brien & Gere Engineers, Inc., Genesee River Water Quality Investi-
    gations, Part I, Joint Venture, Rochester, New York, 1976.

3.  Canale, R.P., and Nachiappan, S., Steady State Modeling Program,
    Sea Grant Technical Report N. 27, University of Michigan, March, 1972.

4.  Bonham-Carter, G., Thomas, J.H. and Lockner, D., A Numerical Model of
    Steady Wind-driven Currents in Lake Ontario and the Rochester Embayment
    Based on Shallow Lake Theory, IFYGL Rochester Embayment Project Report
    #1, Department of Geological Sciences and Department of Mechanical and
    Aerospace Sciences, University of Rochester, Rochester, New York, 1973.
                                   556

-------
               METHODOLOGY FOR EVALUATING THE IMPACT
             AND ABATEMENT OF COMBINED SEWER OVERFLOWS
               A CASE STUDY  OF ONONDAGA LAKE,  NEW YORK

                                  and

                            Peter E. M6ffa
                              John C. Byron
                           Steven D. Freedman
             Stearns § Wheler, Civil and Sanitary Engineers
                       Cazenovia, New York  13035

                                  and

                            John M. Karanik
                               Randy Ott
                            Qnondaga County
                  Department of Drainage and Sanitation
                     North Syracuse, New York  13212
                              ABSTRACT
       A general methodology is presented for the evaluation of the impact
and abatement of combined sewer overflows on receiving waters.  It was
developed from experience with Onondaga Lake in Central New York which
receives combined sewer overflows from the  City of Syracuse.

       Field investigations of the combined sewer  system and the receiving
water must first be undertaken.  The field work includes flow measure-
ment and water quality sampling of the sewer overflows and the receiving
water during several different storms. Use of a computerized data bank
has been found virtually essential for the  storage and manipulation of the
large quantity of data resulting from the sampling and analysis.

       Mathematical modeling of the receiving water is undertaken to eval-
uate water quality as a function of pollutant load; the storm sewer system
is modeled to determine the quantities of pollutants discharged during vari-
ous storm  conditions.
                                  557

-------
       Abatement alternatives,  and their respective costs, for the reduc-
tion of pollutants from wet-weather sources, particularly combined sewer
overflows, are next investigated.  Using engineering judgment for the most
effective and economic abatement measures, a relationship is then devel-
oped between abatement cost and storm condition for each of several
water quality criteria or goals.   From the cost-benefit relationships thus
developed, a graphical determination can be made of the  "general optimum
solution11 (GOS) for reduction or treatment of combined sewer overflows.

       Abatement alternatives and their respective  costs are investigated
next for the reduction of pollutants from wet-weather sources,  particularly
combined sewer overflows.

       In the study for Onondaga County,  New York, 35 overflows from the
combined sewers of the City of Syracuse, which  serve an area of about
eight square miles,  were monitored for a period of one year.  Onondaga
Lake is approximately four and one-half square miles in  surface area; it
was sampled at ten surface locations for the period of influence of each of
six storms.  The U.  S. Environmental Protection Agency Storm Water
Management Model (SWMM) was applied to the City's sewer system.  A
27-segment, tnree-dimensional, dynamic water  quality model of the lake
was developed with capability of predicting enteric bacteria, dissolved
oxygen, nutrients,  and toxic materials.

       From the models, it was determined that the impact of combined
sewer  overflows (CSO) on dissolved oxygen concentrations in  Onondaga
Lake will not be critical after tertiary treatment facilities for dry-weather
wastewaters are placed in operation.  Combined sewer overflow contribu-
tions of phosphorus will be negligible in comparison to those from other
sources.

       In an average rainfall year,  38 violations of the fecal coliform
standard will occur in the area of the lake intended for contact recreation.
If abatement of CSO pollution were to follow the "general optimum solution"
of this methodology,  there would still be 13 annual violations,  ten of which
would occur from June through September.   Inasmuch as each violation
persists for about three days, more extensive CSO abatement will be
required if the projected recreational usage  of Onondaga  Lake is to  be
realized.

       The methodology was prepared in fulfillment of (U. S.  Environmen-
tal  Protection Agency) Contract No. R805096010 by Stearns & Wheler,
Civil and Sanitary Engineers,  and the Onondaga  County Department  of
Drainage and Sanitation.
                                  558

-------
                            INTRODUCTION
       The impetus provided by the Federal laws concerning water pollu-
tion (1, 2) has resulted in great  strides in wastewater treatment.  The pre-
vailing attitude leading to the 1972 .Amendments to the Water Pollution
Control Act (1) was that  there had been a general lack of progress in abat-
ing the country's pollution.   Consequently,  secondary treatment plants
throughout the country were built on the basis of the obvious benefit they
would  provide to receiving water quality.

       Since 1972, the required M20l" facilities  planning has progressively
expanded to include analyses and evaluations of separate sewer systems,
and most recently in some states, combined sewer systems. The combined
sewer  overflow  (CSO) investigations have had to  address an old but rela-
tively undefined component  of municipal wastewater; namely, stormwater.
The highly variable nature of this element and its double-edged impact on
receiving water of providing dilution as well as adding pollutants has
resulted in a more deliberate approach to cost-benefit relationships than
had previously been the case for facilities planning.

       The common "dry-weather" receiving water standard based on con-
secutive seven-day low flow conditions occurring once every ten years
(MA7CD10) for conventional treatment has gradually given way to the  con-
sideration of allowing deviation  from "dry-weather" for certain storm
events.  Recently,  attention has been devoted to  the concept of the "design
storm" as being that storm, for which abatement  facilities are planned (3).
However, the elusiveness of this approach has led to the consideration of
cumulative storm  effects over monthly or annual cycles.  A defined
approach is not  yet available, but actual ongoing experiences are now pro-
viding  the basis for practical approaches to the problem.

       The methodology outlined herein is based largely on,  but not limited
to,  the successful experience in identifying the discharges from a combined
sewer  system and their impact on the water quality of an urban lake in
Onondaga County just north of Syracuse, New York (4).

       The general tasks involved in a CSO study are shown in Figure 1.
The level of effort of each task will vary depending upon (1) the type of
                                  559

-------
       ..V,, H01VW
  CO
  O
  Ul


o o
CO CC
O Q-
I
I

I
ui
  *1|

I  "d
  S31
I  u s

I	I


_
i r


z
z
4
C
* CO
1*1
•
w CZ
2 2
O < <
(033
O O O
        I
        ! £ &j i
        i s|l
        II>|
        s
       LLTJ
           o

          (T 5

          pp

          *i

          (9
        s^
        g§
                         QL
                        y w
                         co £=
                         co _i

                         ui =>
                         cr or
                        ui
                        o to
                        H £

                        ri
                        fe <
                 Ul
                 w
S£

£3
  CO
  CO
  Ul

2 s
Ul <
VIR
                        £co
                        UJ Ul

                        2 >
                        2
                         cr
                         Ul
                                           to
                                           CD
                                           C
                                           o>


                                           •£
                                           "o
                                           XJ
                                           o


                                           o
                                           C/)
                                           o
                                           to
                                           o
                                           0)
                                          UJ
                                           I

                                           o»
                        jfc
                      ,,V,, HOiVW
                560

-------
assessment required; (2) the uses of the receiving waters; (3) the budgetary
limitations of the study; and (4) other factors particular to the study area.
             COMBINED SEWER OVERFLOW MONITORING

       The objectives of any wet-weather monitoring program are not to
capture the "elusive" critical storm conditions, but to sample a statisti-
cally significant number of storms to be able to develop relationships be-
tween rainfall-runoff-overflow and receiving water quality for the ultimate
purpose of projecting cause-effect relationships.  Short-term rainfall
events provide the "real time" data for  impact  assessment while long-term
or historic rainfall data provide the statistical  b'asis for making projections.

  .  .   CSO system monitoring and data gathering can be greatly assisted
by (1) interviews with knowledgeable municipal personnel; (2) previous
reports on the  system',  such as Infiltration/Inflow Analyses and Sewer
System Evaluation Surveys; and (3) record drawings on the interceptors
and diversion structures.

       The measurement of overflow quantity may involve the measure-
ment of the flow reaching the interceptor as well as the flow discharged to
the receiving water, depending upon the capacity of .the interceptor. • The
number of individual points that have to be measured should be minimized
to reduce equipment maintenance and data coordination.  It also is desir-
able to. install the flow metering device  at the same location as the auto-
matic sampler which measures quality  so that operation of the sampler  will
be triggered by ah actual overflow.occurrence.  Such time coordination
becomes critical for determining pollutant loads, which are calculated as
the product of flow quantity and pollutant concentration.   Thus,  the simplest
and most direct measurement is at the  overflow to a receiving water.
Under surcharge conditions, it is particularly necessary to estimate the
interceptor flow as well as the receiving water overflow in order to deter-
mine the true quantity of overflow originating from a tributary area.

       The frequency and duration of sampling should be based on an ade-
quate description  of the peak period of pollutant discharge, often called  the
first-flush period, and the  subsequent subsidence of this  period to rela-
tively insignificant quantities.  It has been found that in the period of five
minutes following the onset of runoff, significant concentrations of pollu-
tants can occur.   Thereafter, 7-to 15-minute intervals have proven to be
adequate to define an overflow discharge. In general, the minimum dura- .
tion of sampling should be ample to capture the runoff from a six-hour
storm.                                         •
                                  561

-------
quality.
       Table 1 shows some of the more important factors affecting CSO
ONONDAGA  COUNTY

       The combined sewer system of the City of Syracuse covers approxi-
mately eight square miles and has 87 active overflow sites, of which 35
were monitored for flow quantity and pollutant quality.  The monitored
sites encompassed 85 percent of the  total defined runoff area and hence
were representative of the overall system;  Six rain gages were located
throughout the City to monitor rainfall, each representing an area of
approximately two square miles.

       The main interceptor system has little or no capacity for wet-
weather flow and thus the quantity of flow discharged to the stream was
assumed to be the total quantity entering the  diversion structure.  This
assumption may not be applicable in  many cases and must be verified by
field data.

       Overflows were monitored through the use of ultrasonic flow meters
in combination with sampling units capable of taking 24 separate samples.
Samplers were initiated by a signal from the flow meter and were operated
at 7-1/2 minute intervals thereafter  (4).
              COMBINED SEWER OVERFLOW MODELING

       CSO system modeling is also an important task, and one which can
be troublesome if not approached cautiously.  As the complexity of the
urban stormwater model increases,  so does (1) the data needs for calibra-
tion and verification; (2) the associated costs for  obtaining the data; and
(3) the cost of operating the mathematical model.  A good rule to follow in
selecting stormwater models is to use the simplest one which will produce
the desired results.  The major types of data which  are required for most
models include (1) surface area and slope; (2) percent impervious;  (3) ma-
jor collection system components, such as sizes  and slopes; (4) overflow
hydraulics; and (5) rainfall.

       There are a number of mathematical models currently available for
CSO and stormwater analyses.  They range  from the "desk-top", hand-
calculated models, such as the simple "rational formula", to the sophisti-
cated EPA Storm Water Management Model  (SWMM) (6, 7).  Three  cate-
gories have been suggested for all stormwater assessment models,  based
on their applications, as shown in Table 2.
                                  562

-------
               TABLE 1.   FACTORS  AFFECTING CSO QUALITY
Antecedent'. Dry Weather
Affects amount of pollutants available on
surfaces and in sewers, and sets initial
infiltration capacities of previous
surfaces.
Land Use of Catchment Area
Pollutant Accumulation
and Washof f
Street Sweeping Practices
The type of land use can dramatically
affect the pollutant accumulation rates
in an area.

Accumulation is dependent on antecedent
dry weather, land use, street sweeping
practices and rainfall characteristics.
Pollutants are often modeled as fractions
of the amount of dust and dirt.

Affects the amount of solids on the
streets which is available for washoff
Soil Erodibility and Erosion
Control Practices
Erosion is a source of solids in combined
sewers.
Dry Weather Wastewater
Solids Deposition and Scour
in Sewer System
Determines the background quantity and
quality within an area.

Dependent on antecedent dry weather,
sewer system characteristics and rain-
fall characteristics.
Sewer System Characteristics
Slope, shape and size affect solids
transport and deposition.
                                     563

-------

          TABLE 2.  CATEGORIES OF STORMWATER MODELS
                                    Application
                               Problem Assessment

                               Planning

                               Event Analysis

       Table 3 further describes the three categories of stormwater
assessment in terms of model complexity and characteristics.

       Some models are capable of being used for more than one applica-
tion.  For example, the EPA SWMM model is considered  a single event, or
Level III model, but it can also be  operated continuously,  or as a Level II
model.  Those models which do not require the use of computers, such as
the EPA-Level I SWMM, the URS,  Inc.   procedure, and the EPA Areawide
Assessment Procedures Manual method, are considered Level I, or prob-
lem assessment models (8,  9, 10,  11).  In addition to the continuously
operated EPA SWMM,  other planning, or Level II,  models include the
U.  S. Army Corps of Engineers' STORM and EPA Simplified SWMM
(SSWMM) (11, 12).
                   RECEIVING WATER MONITORING

       Receiving water monitoring is one of the most important aspects of
the data gathering efforts.  While data may exist on both dry-weather con-
ditions and point source discharges,  there rarely is  available adequate
wet-weather data.  Dry-weather data, although providing an important
comparative basis, rarely describes wet-weather responses.

       A multitude of pollutants  can be introduced into a receiving water
from CSO's, however, the primary concerns can be  categorized into
(a) pathogens, because of public health,  and (b) dissolved oxygen, nutrients
and toxics,  all because of their direct effect on the aquatic system and indi-
rect effect on public health.  The fecal coliform and  fecal streptococcus
groups can be used as pathogenic indicators.  Dissolved oxygen, phosphorus,
and, in some cases, nitrogen,  are critical nutrients or indices to be
evaluated for their impact on aquatic life. Initially,  evaluation of toxics
should be limited to the heavy metals, unless specific information indicates
the need to investigate toxicity from organic chemicals,  e.g., pesticides,
PCB's. Organic analyses represent  a relatively high level of complexity.
Toxicity observed in the BOD test can be used as an  indicator for proceed-
ing with any additional analyses.  One exception to this  dictum would be the
case where bottom deposits that may be subject to toxic organic

                                  564

-------
TABLE 3.  STORMWATER ANALYSIS LEVEL
          AND MODEL COMPLEXITY
Analysis
Level
I





II













III











Model
Type
Desk-top





Continuous
simulation












Single
event
simulation









Model
Complexity
Low to
medium




Low to
medium












Medium
to high










Purpose
of Model
Problem assess-
ment, prelimi-
nary planning,
alternative
screening


Problem assess-
ment, planning,
preliminary
sizing of facili-
ties (particularly
storage) , alterna-
tive screening.
Assess long-
term impacts of
designs.



Analysis for
design, detailed
planning.









Model
Characteristics
No computers.
Equations, nomo-
graphs based on
statistical analyses
of many years of
records .

Program of few
hundred to few
thousand statements.
Uses many years of
rainfall records
with daily time
steps, or worst two
years with hourly
time steps. May
include flow routing
and continuous
receiving water
analysis .
Program to cover
10,000 statements.
Higher modeling
precision, from
rainfall through
sewers, possibly to
receiving waters.
Short-time steps
and simulation
times . Fewer
alternatives to be
evaluated.
                565

-------
contaminants are scoured frequently and play a significant role in the year-
round quality of overlying waters.

       Decisions as to what storms should trigger receiving water sam-
pling should be based on the anticipated  storm intensity.   Close coordination
should be maintained with those sampling the sewer system so as to maxi-
mize the number of overflows being monitored during receiving water
sampling.  Sampling locations and the frequency and duration  of sampling
are dependent upon the changes a particular parameter may undergo during
a storm period.  Thus, prior knowledge from other studies or background
data on the receiving water is helpful.  In  the absence of such  information,
the general guidelines outlined herein can  be used.

       The location of sampling points should reflect the intended usage
within  the body of the receiving water.   In general,  areas where contact
recreation is anticipated should be sampled more intensely than areas
where  fishing or boating are anticipated.  The circulation pattern of the
receiving water should be considered in determining surface locations and
depths of sampling. Surface locations should include tributaries as well as
in-stream locations reflecting the outer boundary of any such  tributary's
mixing zone.  In this way,  the impact of a particular tributary can be
estimated.  If a tributary receives combined sewer overflows, it should be
sampled so as to bracket these discharges,  either individually or collec-
tively.       .                                                       ^

ONONDAGA COUNTY                                                ;-

       Water quality data from dry-weather periods were available for
Onondaga Lake and were used to estimate  the pollutant levels  expected
during storms.  In actuality, these estimated levels were often exceeded.

       Anticipated storm intensity determined whether or not  receiving
water sampling was to be conducted for  any particular storm.   On each
positive occasion, the crews in charge of  sampling the sewer  system were
alerted.   The frequency and duration of sampling during a storm were
estimated at the onset of the project, but were later adjusted on the basis
of the first two storms.  The water  quality parameters generally measured
were bacteria, BOD, DO,  solids, nutrients,  and heavy metals.

       All major tributaries were sampled and measured at their points of
entrance into Onondaga Lake.  In addition, two of the tributaries were sam-
pled upstream of the CSO's so as to bracket its collective input.  Prelimi-
nary sampling was performed to determine the validity of sampling tech-
niques for each of the  tributary locations.  To be representative, tributaries
were sampled in cross section and in depth, depending upon their dimen-
sions and the degree of mixing in each.  The lake was sampled at ten

                                  566

-------
 surface stations; each station was sampled at two depths to reflect the
 epilimnetic (upper) and hypolimnetic (lower) waters.  One surface station
 was located at each of the lake's two deepest points, the remainder being
 located to reflect tributary inputs or shoreline  areas expected to be used
 for contact recreation.  Figure 2 illustrates the sampling locations.
                              DATA BANK

        Whenever CSO studies yield large amounts of data, it is important
that they be handled in a form that is easily retrieved and analyzed.  For
any kind of system, every sample must be adequately identified, e.g.,
storm number,  location,  date,  time,  and parameter.  A data file may take
the form of summary sheets or can be computerized, depending upon the
size of the project.   A computerized file will perm-it the rapid retrieval of
data in various  forms, which is particularly helpful in CSO projects where
dilution volumes of samples for certain parameters may have to be adjusted
following the first few storm samplings.  These and other adjustments,
such as monitoring locations and  possible elimination of certain stations/
recfuire that initial data be evaluated as soon as possible.  Provision should
be made to plot the concentration levels of the various parameters versus
time for a given CSO and, when combined with flow, calculate a plot  of
pounds of material versus time.  Other sub-routines could be developed to
determine the pounds of material discharged during an entire storm.

ONONDAGA COUNTY                                      :

        A computerized file system was established for the combined sewer
system and a  separate file was set up for the receiving water.   Since the
combined sewer system investigation and the receiving water investigations
were conducted by independent organizations,  separate data systems  were
developed; this  presented no problem since, with the exception of storm
sampling,  the data analyses were relatively independent of each other.   The
output requirements  and mathematical model ing for the sewer system were
markedly different from those of  the receiving water.
                    RECEIVING WATER MODELING

        Receiving water modeling must be tailored to the particular situa-
tion.  For example, simple steady-state biochemical oxygen demand-
dissolved oxygen (BOD-DO) models should first be used to assess impact
•rather than more complex dynamic models.  However,  in the long run more
detailed modeling may be required.  Simple fecal coliform die-off models,
such as Chick's Law,  can be used effectively to make preliminary assess-
ments.  When an estuary or lake is involved, the receiving water

                                   567

-------
abanos
NON.
                                             00
                                             a
                                             o»
                                             a
                                             •a
                                             c
                                             o
                                             V)
                                             8
                                             o>
                                             c

                                             15.


                                             a
                                             •a
                                             c
                                             a

                                             c
                                             _o

                                             Qu


                                             "5


                                             e
                                             CM




                                             3


                                             iZ
       568

-------
assessment may be slightly more complex than a flowing water body, but
the same "simple model first" rule should be followed.

       A receiving water model  should include certain fundamental func-
tions as well as those required to meet specific needs.  The model should
consider circulation as fundamental to the distribution of pollutants in the
receiving water. The circulation model should include advective and dis-
persive transport and,  in addition, mechanisms for settling and scour.

ONONDAGA COUNTY

       Both simplified and detailed modeling was performed for Onondaga
Lake.  Simplified modeling of a one-year  storm showed contravention of
bacterial standards throughout the recreational zones of Onondaga Lake.
In order to verify these results,  more detailed modeling had to be under-
taken.

       Simplified modeling was also attempted to determine the dissolved
oxygen relationships  in the lake.  However,  owing to the complexity and
intensity of algal growths and their significant influence on the DO level,
a more  detailed model was required.

       On the other hand,  a simplified approach was taken to determine the
significance of phosphorus loadings from storm runoff on  an annual basis.
This approximation was considered adequate to assess what turned out to
be the relatively insignificant contribution of stormwater to the phosphorus
content of the lake.

       Each CSO was identified as entering  one of two tributaries to the
lake. In the detailed receiving water model, the CSO's entering each
tributary were treated as single,  collective  inputs. The lake was divided
into 27 distinct segments for the purpose of  defining circulation patterns.
Twenty-one (21) of the segments were located in the epilimnetic (upper)
waters and six of the segments were located in the hypolimnetic (bottom)
waters, as illustrated by Figure 3.  The model utilized steady-state circu-
lation patterns and hydraulic loads in combination with time-variable pollu-
tant loads to simulate the lake's  water quality response to stormwater
discharges.

       The model described the  three-dimensional transient distribution of
fecal coliform,  total  phosphorus, BOD and dissolved oxygen.  The model
included advective and dispersive transport  and mechanisms for settling,
scour, and chemical  and biochemical reaction.

       The model is  based on the conservation of mass as expressed by the
following equation:

                                   569

-------
CD

 c
 o
 c
 0>


 o>
 
T3
 o
ro

 

 3
 C3>

iZ

-------
          Accumulation = Net Input of Mass + Net Input of Mass
             of Mass      from Advective     from Dispersive
                              Transport          Transport

                      t Sources and Sinks of Mass
or mathematically     ^~~= V. E V C -  V. UC + r
                       O t                                     .

where  C = concentration of pollutant (mg7l)

        t  = time (days)

        * = del operator

        E = disper-sion coefficient (l/m2/day)

        U = velocity (m/day)

        r = reaction sources or sinks (mg/l/day)

        Since a direct solution of this equation is usually not possible, the
solution requires  the use of approximate solutions.  The finite difference
solution,  which is equivalent to representing the continuous lake system  as
a collection of completely mixed cells that are interconnected by advective
and dispersive transport, was used in this model.  Figure 4 illustrates the
essential factors. While this approach sacrificed some spaeial accuracy,
the careful segmentation of the water body to closely represent the overall
circulation and mixing minimized these inaccuracies.  The non- steady con-
tinuity equation is developed elsewhere (13).                 ...

        The Onondaga Lake  model was calibrated and verified against four
distinct storms.   The resulting model was then used to predict two of the
same storms and,  in both cases,  resulted in good fits. In addition,  sensi-
tivity analyses were performed for the various rate coefficients.  In
general, 25 to 50  percent changes,  plus and minus,  in the coefficients ori-
ginally  employed were used to test the model's sensitivity to such values in
terms of water quality impact.

MODEL PROJECTIONS

        Prior to projecting abatement requirements,  the critical environ-
mental  conditions must be determined for use in the water quality model.
Input conditions must be carefully selected so as to be critical for the
parameter in question.   For example, the effect of fecal coliforms on water
quality  is most severe under conditions of minimal die -off (low tempera
tures) and rapid transport;  the dissolved oxygen parameter,  on the other
hand, is most adversely affected when biochemical rates  are maximum
(high temperatures) and reaeration is minimum. Bacterial die- off rates

                                   571

-------
I
                                                           I
                                                           4>


                                                           "o
                                                           0>

                                                           -§
                                                           O


                                                           "o

                                                           .o

                                                           c
                                                           0>
                                                           a

                                                           o

                                                           -a
                                                           a>
                                                           o>
                                                           o
                                                           CO
                                                           V)
                                                           to
                                                           o

                                                           E
                                                           o
                                                           o
                                                           a.
                                                           si

                                                           z*

                                                           1.1
                                                             2
                                                           O Q.
                                                           £ '-5

                                                           ^

                                                           0>
                                                           3
                                     572

-------
should be determined through laboratory-controlled tests on the receiving
water.  In predicting the dissolved oxygen content of productive lakes, light
conditions and the algal standing crop can represent the dominant effect on
the DO concentration.

       The  projection of pollutant loads requires considerable  judgment to
avoid a combination of events that may have a very infrequent reoccurrence
interval.  If,  for example, an MA7CD10 dry-weather stream flow is com-
bined with storm runoff having a recurrence interval of one year, the pro-
bability of both conditions occurring at the same time can exceed once in
30 years.  The selection of dry-weather flow conditions for a large lake is
less critical than for a stream because circulation is more influenced by
lake elevations and lunar effects than incoming tributaries.  In addition to
flow considerations, pollutant loadings in. a model must include dry-weather
components representing both point and background sources.  Point source
information often can be obtained through the NPDES or State permit  sys-
tem, but background source information may require field measurements
during dry periods.

ONONDAGA  COUNTY

       Current measurements taken on Onondaga Lake during the storm
sampling periods served as the basis for determining the in-lake circula-
tion patterns assumed in the mathematical model.  Background water
quality data was available from prior dry-weather  sampling.  Light condi-
tions were determined from regional data in conjunction with per cent-cloud-
cover measurements at the local weather station.

       Dry-weather data from prior years  served as the basis for  deter-
mining point source pollutants.   The data used were limited to  the months
of June through September and represented average values of two to six
years of data, depending upon whether or not the data reflected contempo-
rary operations in the tributary.  In some cases, modification  of the  sewer
system prevented the use of all six years of data.

       Projections of the water quality impact of storms with recurrence
intervals of one,  two,  and ten years were made.  It was found that  the 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.   It was  pre-
dicted that a storm of this magnitude would cause contravention of coliform
standards for a period in excess of three days.  Figure 5 shows the results
of an actual two-year  storm in terms of fecal coliform concentration.
                                  573

-------
      	A. \__
574

-------
       On the other hand, the dissolved oxygen demand from, even a ten-
year two-hour storm had only minimal effect on the dissolved oxygen bal-
ance of the lake; the maximum DO deficit for that storm, under the most
critical conditions, was determined to be 2.8 mg/1.  This is shown in
Figure 6.  It was determined that the  algal  standing crop and abrupt algal
die-off, following treatment of the domestic and industrial wastes, will
become the principal factor affecting the lake's dissolved oxygen.  Figure 7
illustrates this relationship. Assuming an  algal standing crop having a
chlorophyll a concentration of 86 ug/1, observed for relatively clear waters
in Onondaga Lake,  a die-off could result in a DO concentration approaching
2. 0 mg/1.  Although phosphorus is likely to be the limiting nutrient,  annual
storm-related contributions of total phosphorus are negligible when com-
pared to the projected dry-weather loads.
                  COST-EFFECTIVENESS ANALYSES

       The development of relationships between CSO discharges and other
wastewater sources relative to target water quality goals or standards is
an important first step in assessing impacts and determining if CSO abate-
ment is worth pursuing.   BHgure 8 presents an,example of present and
future  dry-weather and the various wet-weather BOD loads in relation to a
DO goal.  If some degree of CSO abatement will result in a significant
step toward a defined water quality goal, then the next task will be to per-
form a cost-effectiveness analysis.

       A common economic approach to stormwater treatment has been to
develop relationships between the impact of stormwater on receiving water
quality and the cost of treatment for various storm 'conditions. This cost-
benefit information has then been used by local  governments and regulatory
agencies to make reasonable economic decisions.

       Figure 9 illustrates the interrelationships between the various
phases of a CSO study and shows the development of a cost-effective treat-
ment solution.

       Figure 9A illustrates the realtionship between the pollutant loads
discharged from  a CSO system and various storm conditions described in
terms  of recurrence interval.  Recurrence interval can be used in this
manner only if other significant factors such  as pollutant residual and  ADW
are held constant.  Recurrence  interval is an appropriate variable from the
standpoint of the  general availab.ility of data and relationship to abatement
costs.

       Generally, several years of rainfall records are available whereby
frequency of overflow  can be determined on the basis of sewer system

                                  575

-------
«•
_J
eg
!§
8
CO
o .
§.
i i i i i 	 1 i 	 1 i i 	 1
CELL 1
A\





°1)j» 3353 ' 66.67 100.00 ' 133.33 ' 166.67 ' 200.00
TIME (HOURS)
8"'
CD
z .
CELL 1


^00 ' 33.33 ' 66.67 ' 100.00 ' 133.33 ' 1&.67 " 200.00
TIME (HOURS)
CD—
8
8
CELL 1
i •



TJ.OO        33.33      66.67      100.00      133.33      166.67      200.00
                           TIME  (HOURS)

  FIGURE  6.     ONONDflGfl LflKE  PROJECTION FOR  fl ONCE IN TEN
                YEflRS STORM:   LflKE RESPONCE NEflR PRIMflRY LOflD(!8)
                              576

-------
                                          I    I	1	H	-H	1
                  «     II    }    I - 1    <
3£0       7.00       10.50       14.00
               TIME (DflYS)
                                                            17.50       2100
                  »	1	1	1   •••<	1	I-	1	1    I
                      3.50
          1M       10.50       \4M
               TIME (DflYS)
17.51      21.00
                            •	1	1	1     i	1	1	1	1
                      3.50
          7.90       JWO      \4M
               TIME  (DflYS)
17.50       2UO
Figure  7.   Model  simulation  of  chlorophyll  dieoff.(IS)
                                    577

-------
r
                                        WET-WEATHER
                          UJ
                          O
                          Z
                          O
                          O
                               FUTURE
                                 DRY-
                              WEATHER^
                                      PRESENT DRY-WEATHER
                                          BOD  DISCHARGED (Ibs/doy)
               Figure 8. Determination of  required CSO abatement relative to water

                         quality  goal.(4)
                                                 578

-------
      RECURRENCE INTERVAL

   A. RAINFALL TYPE VS. POLLUTANT LOAD
  WATER QUALITY RESPONSE


B. POLLUTANT LOAD VS. WQ RESPONSE
        ^Allowoblt
   "AllowoW*

      RECURRENCE INTERVAL STORM (Rn) —••

    C. RECURRENCE INTERVAL VS. REQUIRED
      TREATMENT
   RECURRENCE INTERVW. STORM (Rn>      	P-
               or                  -«—
   NO. OF WATER QUALITY VIOLATIONS (WOV) PER YEAR

 D. RECURRENCE INTERVAL VS. COST
Figure  9.   Procedure for establishing the most cost-effective treatment
            level to meet  a  given water quality goal .(4)
                                     579

-------
capacity.  Such information, in combination with even limited quality infor-
mation,  can be used to determine the pollutants discharged from a system
over a period of years.  This is accomplished either through a series of
single-event calculations or continuous simulation.

       The significant factors in developing costs of abatement strategies
are flow rate and volume,  which are derived from the characteristics of
recurrence interval,  namely, rainfall intensity and duration.

       Figure 9B illustrates the relationship between pollutant load dis-
charged  from the sewer system and water quality response.  Based on a
selected water quality goal,  the allowable pollutant load determined from
Figure 9B can be related to an allowable recurrence interval from 9A.
Although the relationships illustrated by  Figures 9A and 9B can represent
independent study efforts,  they can be interfaced through the development of
the common relationship to pollutant load.

       Figure 9C illustrates the relationship between treatment required
and recurrence interval.  This relationship is based on a  series of points
selected from Figure 9B,  and can be expressed in terms of the pollutant
load in excess of the allowable (Wj^Wojio-vyabie).  The treatment require-
ments set the basis upon which the cost of a treatment facilities can be
determined.   The various  methods of source control and the various treat-
ment options are reported elsewhere (9).  The specific unit processes
available and the associated unit costs for both construction and operation
and maintenance are well documented (14,  15, 16).

       Figure 9D illustrates the cost of treatment as it relates to both
recurrence interval and the number of water quality violations.  It should
be noted that the costs represent the least expensive mix of control and
treatment options in each case.  These costs are developed on the basis of
engineering judgment and relationships developed through previous studies
(17,  18).
                     COST-EFFECTIVE SOLUTION

       Treatment costs can be related directly to either recurrence inter-
val or a measureable benefit, which in this case has been illustrated as the
number of water quality violations per year.  Since recurrence interval
(Rn) serves as the basis for determining the frequency of overflow in the
system,  the number of water quality violations can be directly equated to
recurrence interval.   Another  method to determine the number of water
quality violations is through the use of continuous simulation,  whereby the
number of violations per year for a particular treatment scheme  are
                                   580

-------
calculated.  Costs can then be plotted against the average annual benefit as
shown in Figure 9D.
       For the example selected, the knee of both curves shown occurs at
the same cross point.   The "knee" of the curves represents the point of
diminishing returns.  Below the knee, treatment may be as cost effective
but will result in less removal and subsequently more frequent violations of
the water quality goal.  At all points beyond the knee, the additional benefits
derived from incremental increases in expenditure decrease rapidly and are
commonly  referred to as being beyond the point of diminishing returns.
While it may be desirable to meet water quality standards beyond the "knee",
justification should be based on factors other  than economics (e.g., legisla-
tive decisions or "best usage" of the receiving water).

       A more comprehensive method of presenting abatement costs and
associated benefits involves performing the above procedure for various
water quality goals.  Figure 10 shows a plot of costs versus water quality
violations for various goals in order to illustrate the economic impact of
more or less stringent goals.  Such an approach can be used to determine
the most cost-effective water quality goal or "general optimum solution"
(GOS).  The term GOS is used herein to denote a  solution which encompasses
the essential factors necessary to arrive at a reasonable solution.
Figure 10 presents a summary of these factors from which the decision-
makers can select a reasonable water quality objective.     [

       The GOS can be determined by drawing a curve (A-IB in Figure 10)
through the knees of a series of cost curves similar to the one presented  in
Figure 9D.  The general optimum solution to  the  cost-benefit analysis
occurs at the knee of this curve.  This point not only defines the water-
quality standard and associated cost, but through the use of Figure 9D,
indicates the probable frequency  with which that standard will be contra-
vened.

       It should be noted that in some cases data will not yield a noticeable
break or knee at which the GOS occurs.   However, in any event, presenta-
tion of the  data as shown in Figure 10 will be  desirable  since it provides
the complete picture of the water-quality benefits which can be expected for
any given expenditure.  Even though a GOS exists, it may be desirable for a
variety of reasons (e.g., limited funds or inability to live with the cost-
effective water quality goal) to select another water quality goal.   The cost
data presented in Figure 10 enables the decision-makers to balance the
water quality goals against the total cost of achieving them.
                                   581

-------
=*&
 o
 o
 O
 IT

 1
 t-
 UJ
 ai
 8
                             •Cn- CONCENTRATION SELECTED FOR WATER
                                  QUALITY GOAL. GOAL IS INCREASING IN
                                  STRICTNESS FROM  Cj TO Cn.
                              On- OPTIMUM COST-BENEFIT FOR VARIOUS
                                  WATER QUALITY  GOALS.
                          GENERAL OPTIMUM SOLUTION
             FREQUENCY OF WATER QUALITY VIOLATIONS (NO./YR.)
Figure 10. Procedure for determining the cost effective water quality gpal.(4)
                                  582

-------
ONONDAGA COUNTY

       Curves relating various recurrence-interval  storms, CSO discharges
and water quality were developed for the Onondaga Lake study.  In that
case, the analyses showed fecal coliform to be the only parameter to show
significant water quality impact from the CSO discharges.  Therefore, only
examples of abatement solutions related to meeting the contact recreation
standard of 200 cells/100 ml established by the New  York State Department
of Environmental Conservation will be discussed in this section.

       Extensive statistical analyses of rainfall patterns were performed
on the historical records for Syracuse,  New York, and the results com-
pared to the findings  of both the CSO and the water quality programs.  This
procedure revealed that there are 170 rainfall events per year; based on
peak intensity, 65 of  these produce overflows.  On the  average,  the com-
bined sewer system is capable of accepting the runoff from a storm with
peak intensity of less than 0.05 inches/hour without causing any overflow.
Of these 65 overflow-producing events,  only 38 result in discharges of
sufficient magnitude to cause a violation of the fecal  coliform  standard in
the contact recreation zone.  Figure 11  presents the  number of violations
that occur annually as a result of treating  storms of  various intensities or
recurrence intervals for a fixed two-hour  duration.  Figure 12 presents a
similar plot in terms of days of violation.   Figure 13 shows the present
worth costs (1983) for constructing abatement facilities to treat these
	i	                                                    *
storms.
       Inspection of the curve presented in Figure 13 shows the  knee  of
the curve occurring in the vicinity of $12 million; even with that investment,
13 violations would occur per year.  Of the 13 annual violations,  over ten
will occur during the recreation season from June through  September as
illustrated by Figure 14.  This case is one where the knee-of-the-curve
solution is unacceptable.  Since ten violations would  essentially prohibit
use of the lake for contact recreation, more stringent levels of treatment
will ultimately be needed.  The selected goal,  for which treatment will
eventually be provided,  is to allow one violation of the  contact recreation
standard (200 cells/100 ml) per year.
                                  583

-------
                                                                      (A
                                                                      o
                       1—I	1—I—\
                                                               O>
                                                               CL

                                                               V>


                                                               O

                                                               +—
                                                               10

                                                               o
                                                                      0>
                                                                      *£.
                                                                      a
                                                                      o
                                                                     TJ

                                                                      o

                                                                     6
 O
 3


 C
 O

 0>
 o
 o
 o

 tO
                                                               B    .2 «>


                                                               O     >»~r
                                                               
(uq/uj)

                           ro  cvi     —
0)

•5
o>
                               584

-------
r~
                                                      §
                                                      i
                                                          CD
                                                                   O
                                                                   .3
                                                                   T3
                                                          OJ
                                                          ro
                                                          CM
                                                      tO
                                                      c
                                                      o

                                                      B

                                                      .2
                                                              o

                                                             a
                                                              o>
                                                              Q.
                                                             8
                                                       — CO
                                                          o
                                                                  «»- 3
                                                                   a>

                                                                   15

                                                                   -
                                                          .£ CA
                                                          J3 C



                                                           II
                                                                  •£

                                                                    E*^
                                                                   ^. o
                                                                   g cr
                                                                  CJ
(£86l)SJD||oa jo
                                  U
                                      585

-------
                                                                  o
o
(X!
                                                       CD
                                                       LO
                                                       O
                                                    .. CD
                                                    -- oo
-1 	
o
o
	 1 	 •
o
CD
	 « 	 •
O
CD
	 , 	
O
st
	 1 	
0 C
OJ
(£861)SJDJ|Oa JO
Ul S|SOQ
                                            |U9S9Jd
                                                            CD
                                                            
                                       O
                                       2g

                                       •S-2
                                       £ o
                                       £  >
                                       o  >>
                                       o ±:
                                                                  0)
                                       'i§
                                       -2^
                                       o c
                                       O O

                                        rr
                                       o a\
ro



 I
 g>
LL
                                586

-------
to
Ul
tr
o
x
tr
x
CD   O  CD
00   10  fO   —
O   O  O   O

                                                                        -S
                                                                          a>
                                                                          D
                                                                        o o
                                                                        C C
                                                                        a o
                                                                        o E
                                                                        M— k.
                                                                        — O
                                                                        o"S5
                                                                          ^^%
                                                                        o~
                                                                        o
                                                                        co (Q
                                                                        g
                                                                        »_ o>
                                                                        -S~
                                                                        p
                                                                        c  >
                                                                        a>
                                                                          o
                                                                          f>
                                                                          a
                                10         evi

                   HINOlN/SNOIlVnOIA 39VH3AV
                                                                        £

                                                                        I
                                                                        o>
                                                                        LL
                                    587

-------
r
                                         REFERENCES
              1.   Federal Water Pollution Control Act Amendments of 1972, Public Law
                  92-500.  92nd Cong.,  86 Stat.  816, October 18, 1978.

              2.   Clean Water Act of 1977,  Public Law 95-217.   95th Cong., 91 Stat.
                  1566,  October 27, 1977.

              3.   Shelley, P.E.,  and G.A.  Kirkpatrick.  Sewer Flow Measurement, A
                  State-Of-The-Art Assessment.  EPA-600/2-75-027, U.S. Environ-
                  mental Protection Agency, Cincinnati, Ohio, November 1975. 436 pp.

              4.   Moffa, P. E., et al.  Methodology for Evaluating The Impact and Abate-
                  ment of Combined Sewer Overflows, A Case Study of Onondaga Lake,
                  New York.  Grant No.  R805096, U.S. Environmental Protection
                  Agency, Cincinnati,  Ohio,  Final Draft Report,  June 1979. 120 pp.

              5.   O'Brien & Gere, Inc.  Combined Sewer Overflow Abatement Program.
                  Report Being Prepared for Onondaga County, New York.

              6.   Huber, W. C.,  et al.   Interim Documentation, November  1977 Release
                  of EPA SWMM,  Draft Report.  U.S. Environmental Protection Agency,
                  Cincinnati,  Ohio, November 1977.

              7.   Huber, W. C.,  et al.   Storm Water Management Model User's Manual
                  Version II.   EPA-670/2-75-017, U.S. Environmental Protection
                  Agency, Cincinnati,  Ohio,  March 1975.  367 pp.

              8.   Areawide Assessment Procedures Manual, Volumes I, II, and III.
                  EPA-600/9-76-014,  U.S. Environmental Protection Agency,
                  Cincinnati,  Ohio, July 1976.

              9.   Lager, J. A.,  et al.   Urban Stormwater Management and Technology:
                  Update and  Users' Guide.  EPA-600/8-77-014, U.S. Environmental
                  Protection Agency, Cincinnati,  Ohio, September 1977.  331 pp.

             10.   Amy,  G., et al.  Water Quality Management Planning for Urban
                  Runoff.  EPA-440/9-75-004, U.S.  Environmental Protection Agency,
                  Washington, D. C., December  1974.
                                               588

-------
11.   U.S. Army Corps of Engineers.  Storage-Treatment, Overflow,
     Runoff Model "STORM" User's Manual. Hydrologic Engineering
     Center, Davis, California, July 1976.

12.   Lager, J.A.,  et al.  Development and Application of a Simplified
     Stormwater Management Model.  EPA-600/2-76-218, for U. S.
     Environmental Protection Agency, Cincinnati, Ohio,  August 1976.
     153 pp.

13.   Limno-Tech,  Inc. Mathematical Modeling of the Impact of Storms on
     Water  Quality in Onondaga Lake. Prepared for Stearns & Wheler,
     Civil and Sanitary Engineers,  and Onondaga County,  New York.
     Ann Arbor, Michigan,  1978.   133 pp.

14.   Heaney, J. P., and S,. J. Nix.   Storm Water Management Model:
     Level I - Comparative Evaluation of Storage-Treatment and Other
     Management Practices.  EPA-600/2-77-083,  U.S. Environmental
     Protection Agency,  Cincinnati, Ohio,  April 1977.  105 pp.

15.   Reid,  G.K.  Ecology of Inland Waters and Estuaries.   Reinhold
     Publishing Corporation,  New York, New York,  1961.   375pp.

16.   Turner, E.G., etal.   1976 Needs Survey.  EPA-430/9-76-012,  U.S.
     Environmental Protection Agency, Washington, D. C.,  February
     1977.   400pp.

17.   Sonnen, Michael B.  Abatement of Deposition and Scour in Sewers.
     EPA-600/2-77-212, U.S. Environmental Protection Agency,
     Washington, D. C.,  November 1977. pp. 115.

18.   Stearns & Wheler, Civil and Sanitary Engineers.  Onondaga Lake
     Storms Impact Study.   Report Prepared for Onondaga County.
     Cazenovia, New  York, August 1979.
                                 589

-------
     A WATER QUALITY PLANNING METHODOLOGY FOR URBAN AREAS

                  Franklin W. (Skip) Ellis
                 CH2M HILL, Reston, Virginia
                      Ronald L.  Wycoff
               CH2M HILL, Gainesville, Florida
Abstract

    The Environmental Protection Agency's (EPA) 1978 Needs Sur-
vey reported that an estimated $36 billion is required to con-
struct necessary municipal wastewater treatment facilities.
Additionally, $26 billion is required to control pollution from
combined sewer overflow, and $46 billion is required to control
urban stormwater runoff.  In light of the current trent toward
reduced taxes, local funds for water quality projects will become
more limited.  Moreover, the public should demand that such pro-
jects demonstrate perceptible improvements in water quality and
are required for the protection of the intended beneficial use
of the receiving water.  Further, the most cost-effective pollu-
tion control alternatives should be used.  Without this assurance
required pollution abatement projects may well go unfunded.

Based upon information gained in the performance of the EPA's
1978 Needs Survey, a two-phase approach to water quality planning
is presented that determines the most cost-effective mix of con-
trol alternatives and their impacts on the receiving water.
This approach is general in nature but is oriented toward com-
bined sewer overflow (CSO) areas and the requirements of EPA's
Program Requirements Memorandum, 'PRM No. 75-34.

Phase I is an initial assessment designed to answer the following
questions:

    1.  What is the intended beneficial use of the receiving
        water?             •               •
    2.  What water quality goals or criteria are required to en-
        sure this use?
    3.  Based on the goals, is there a water quality problem?
    4.  Can CSO, urban runoff, or domestic wastewater abatement
        techniques solve the problem, and what degree of controls
        is required?
    5.  If the problem can be solved, what controls, in general,
        are the most cost effective, and what is the nature of
        the tradeoffs between cost, degree of control, and re-
        ceiving water quality?
    6.  Considereing these tradeoffs, are changes in the water
        quality goals, or in the desired beneficial use for the
        receiving water indicated?
                               590

-------
This two-phase approach is: designed to achieve water quality
goals established to protect a given beneficial use of the re-
ceiving water.  Rather than a fixed limit, such as a minumum
dissolved oxygen standard, statistical water quality responses
are used as the goals.  Examples from the EPA's 1978 Needs Survey
illustrate this point.

Phase I involves the use of a simple, continuous receiving water
quality model and the Heaney-Nix economic optimization procedure.
The model, the Continuous Stormwater Pollution Simulation.System
(CSPSS), incorporates the pertinent features.of an urban area and
its receiving water.  A very brief description of CSPSS is given.
The economic optimization is performed using estimated areawide
production functions for a first-cut analysis of the costs re-
quired for various levels of pollution abatement.  The water
quality model uses these relationships to produce cost-water
quality relationships that can be used to guide planners as to
the desirability of various pollution abatement projects.  Stud-
ies from areas with CSO's are presented,-

.If the results of the Phase I analysis indicate that there are
water quality problems that can be corrected with ,affordable
solutions, the detailed analysis of Phase II is required.  This
phase considers, in much greater detail, the area's hydrology,
combined sewer system hydraulics, nonpoint and point source
pollutants, and the receiving water quality response, thus re-
quiring more sophisticated data and models.

The product of the Phase II analysis is  a description of the
optimal mix of control alternatives, the total plan costs, and
the receiving water quality response due to the plan.  The de-
scription of the alternatives includes the level of effort re-
quired, the area to which the alternative applies, the expected
pollutant reduction due to the alternative, and the cost of the
alternative.

In summary, this two-phase approach can  result in substantial
monetary savings by obtaining economically optimal solutions.
Use of this methodo.logy also explicitly  evaluates the benefits
or improvements in water quality as a result of the project.
These features satisfy EPA planning requirements and should
assure the .public .that their tax dollars are wisely spent.

INTRODUCTION

The Environmental Protection Agency's (EPA) 1978 Needs Surveyl
estimates that $36 billion will be required to construct
necessary municipal wastewater treatment facilities, including
secondary treatment and AWT plants.  In  addition, $26 billion
will be required to control pollution from combined sewer
overflow  (CSO) and $62 billion will be required to control
stormwater runoff  (SWR) pollution from existing urbanized

                               591

-------
areas.  Considering the current nationwide trend toward  reduced
taxes, local funds for water quality projects will  become  more
limited.  The public will demand that pollution control  projects
competing for dwindling tax dollars demonstrate perceptible
improvements in water quality and use the most cost-effective
pollution control alternatives.  Without this assurance, bene-
ficial projects may go unfunded.

Program Requirements Memorandum PRM 75-34  (Program  Guidance
Memorandum PG-61), issued by the EPA, requires that projects
funded under the construction grants program for control of
pollution from combined sewer overflow be both cost effective
and necessary for the protection of the beneficial  use of  the
receiving water.  PRM 75-34 also provides several other  criteria
which should be met before potential CSO control projects  are
funded.  These are:

1.   Provision has already been made for funding secondary
     treatment of dry-weather flows.

2.   The pollution control technique proposed for CSO is a more
     cost-effective method of protecting the beneficial  use of
     the receiving water than other CSO control techniques or
     the addition of wastewater treatment higher than secondary
     for dry-weather municipal flows.

3.   The marginal costs are not substantial compared to  the
     marginal benefits.

In addition to the above requirements for CSO control, a more
recent (May 1979) Program Requirements Memorandum (PRM 79-7) has
established a case-by-case EPA-headquarters-level review proced-
ure for all significant projects providing municipal wastewater
treatment more stringent than secondary.  In order  to be approved
for funding, these projects must demonstrate significant water
quality and public health improvements.  Requirements such as
these have been developed to ensure cost-effective  and water-
quality-effective allocation of limited construction grants
funds.

A two-phase approach to urban water pollution control planning
which achieves the goals of PRM 75-34 and PRM 79-7  is presented.
The approach determines the degree of control required, the most
cost-effective mix of water quality management alternatives, and
the impacts of these alternatives on the receiving  water.  The
methodology can accommodate staged implementation,  allowing less
stringent water quality goals to be achieved first  while the
option of achieving more stringent goals in the future is
retained.  The methodology will also provide a clear understand-
                               592

-------
ing of the actual costs  incurred  in  meeting water quality goals
and/or standards/ "all of  the  time"  versus  the cost incurred
when an allowable frequency  of exceedance of the standard is
accepted.  This provides a feedback  system  for evaluating the
economic practicality of meeting  certain water quality goals.

PHASE I - INITIAL ASSESSMENT

The first phase of  the cost-effective  water quality planning
approach is designed as  an initial assessment of water quality
problems and potential solutions.  The methodology, based in
part upon information and  experience gained in the performance
of the Categories V and  VI portion of  the 1978 Needs Survey,2 is
intended to answer  the; following  questions:

1.   What is the intended  beneficial use of the receiving water?

2.   What water quality  goals  or  criteria are required to ensure
     this use?

3.   Based on these goals, is  there  a  water quality problem?

If the answer to question  3  is no, then the planning process is
complete and additional  water  quality  control facilities are not
required.  If the answer to  question 3 is yes, then the following
questions must be addressed:

4.   Can combined sewer  overflow, urban runoff, or additional
     point source pollution  abatement  techniques solve the
     problem; and what degree  of  control is required?

5.   If the problem can  be solved, which controls,  in general,
     are the most cost effective?

Responding to questions  4  and  5 will result in the  formulation
of a pollution control project as well  as a Preliminary Cost
Analysis.   If the project is  readily  affordable and the water
quality goals are fully  met, then the  analyses can  proceed to a
more detailed study, i.e., Phase II.   However,  if the costs  are
excessive or the water quality goals cannot be met,  the following
additional questions should be addressed:

6.   What is the nature  of the tradeoffs between cost,  degree of
     control, and receiving water quality?

7.   Considering these tradeoffs, are  changes  in the water
     quality goals  or in the desired beneficial use of the
     receiving water indicated?
                              593

-------
 If the answer to question 7 is yes, then new water quality goals
 and/or beneficial uses roust be established and the results of
 the Phase I planning process reevaluated until both acceptable
 goals  and affordable solutions are obtained.

 The initial Phase I assessment involves the determination of the
 beneficial uses of the receiving water and the establishment of
 appropriate water quality goals or criteria in order to answer
 questions 1 and 2.  Continuous water quality data and/or a
 continuous water quality computer simulation-model are used to
 determine the nature,  sources, and extent of water quality
 problems, a procedure  which will answer question 3.  The model
 is also used to obtain a preliminary estimate of the degree of
 pollution abatement that would be required to achieve the water '
 quality goals,  thus answering question 4.  An economic optimiza-
 tion procedure is used to determine the most cost-effective mix
 of control alternatives for various levels of pollution abatement,
 which  answers question 5 and 6.  Finally, a feedback process
 allows decisionmakers  a chance to evaluate the proposed project
 in light of costs and  receiving water response, which provides
 the necessary 'information to answer Question 7.

 RECEIVING WATER QUALITY GOALS

 Water  quality goals or criteria generally vary according to the
 intended beneficial use of the receiving water and act as an
 indicator for evaluating whether water quality problems exist.
 Water  quality criteria are also used in the determination of the
 level  of required pollution abatement if a water quality problem
 is identified.

 Water  quality -goals are often defined for a number of pollutants
 or water quality indicators.  However, for the purpose of this
 paper, only dissolved  oxygen (DO) will be considered.  The
methodology outlined here  is  general  and  may be applied  to
other parameters.

The intent of the  Clean Water Act is  to achieve fishable  and
swimmable waters.   The concentration  of dissolved oxygen  in a
receiving water  is  one of  the most important indicators  of the
ability of the water to support a viable  fishery.   Urban  areas
can supress the DO  concentration  by the introduction  of  oxygen-
demanding substances from  industrial  waste  discharges, municipal
wastewater treatment plant discharges, combined sewer overflow,
and stormwater runoff.  The combined  effects of these  sources
may be evaluated by considering the DO budget  of the  receiving
water.
                               594

-------
The dissolved oxygen criteria proposed by EPA3 for  freshwater
fish is:                            ,
                              \ •       "•?" .
     Freshwater aquatic life:  A minimum concentration of dis-
     solved oxygen to maintain good fish populations is  5.0 mg/1.
     The criterion for salmonid spawning beds is a  minimum of
     5.0 mg/1 in the interstitial water of the gravel.

Most states have established fishery standards that basically
agree with these criteria.  Many states differentiate between
cold-water fish (salmonids) and warm-water fish by  setting daily
average cold-water fish DO limits of 6.0 mg/1 and warm water
limits at 5.0 mg/1.  These criteria are the  lower limits of DO
concentrations that fish can tolerate for extended  periods of
time.  However, DO levels as low as 2.0 to 3.0 mg/1 can  be
occasionally tolerated for short periods without causing fish
kills.  Based on this fact, a minimum DO level and  its associated
allowable frequency of exceedance were developed as a basis for
estimating facilities needs required for control of combined
sewer overflow and urban stormwater runoff in the 1978 Needs
Survey.2

From available data on juvenile Brook Trqut4,5 a relationship
between lethal and safe DO concentrations and exposure times was
developed and is shown on Figure 1.  Since juvenile Brook Trout
are an extremely sensitive species,. Figure 1 should be conserva-
tive for most receiving waters.

The selected DO criteria used to evaluate combined  sewer overflow
(Category V) and urban stormwater runoff (Category  VI) facilities
required to provide a viable fishery are given as follows:      >

1.   The minimum receiving water dissolved oxygen concentration
     shall not average less than 2,0 mg/1 for more  than  4 consec-
     utive hours.
2.   The minimum receiving water dissolved oxygen concentration
     shall not average less than 3.0 mg/1 for more  than  72
     consecutive hours (3 days).

3.   The annual average receiving water dissolved oxygen concen-
     trations shall be greater than 5.0 mg/1 for all waters
     which support warm-water species and shall be  greater than
     6.0 mg/1 for all waters which support cold-water  (Salmonid)
     species.

Results .of the simulation studies developed  in the  1978  Needs
Survey indicate that if the 2.0 mg/1 for 4 consecutive hours
criterion is met, then in general all other  criteria are met.
That is, this criterion in most cases controls the  level of BOD
removal required.  The above criteria are used as the receiving
water quality goals in the case studies portion of  this  paper.
                               595

-------
 100.0 -,
  10.0-
w
tr

O
I
g


£
EC

Q
   1.0-
   0.1-
            72 Hours (3 Days)
                                     #/    #/  &/
SAFE
            10%-INDICATES

            MORTALITY

            PERCENTAGE
                                               i        i        r

             0.5     1.0      1.5      2.0       2.5     3.0      3.5

                        DISSOLVED OXYGEN, MG/L
  FIGURE 1.  Mortality of Juvenile Brook Trout due to low DO levels.
                              596

-------
These criteria are an example of identification of the intended
beneficial use of the receiving water  (i.e., establishment of a
viable fishery) and the development of a quantitative statement
of the water quality goals required to ensure this use.  Neces-
sarily, these are the first two steps of the water quality
planning process.

CONTINUOUS WATER QUALITY SIMULATION

Evaluation of the compliance or noncompliance of receiving water
quality with a desired water quality goal requires a long-term
record (several years) of water quality conditions at one or
more points on the receiving stream.  Such records are generally
unavailable; however, properly calibrated water quality computer
simulation models can be used for the problem analysis.  The
time variant nature of receiving water quality requires the use
of a continuous model.  In addition, the model needs to consider
point and nonpoint sources, upstream flow, upstream water
quality, and the dynamics of.the dissolved oxygen consumption
and reaeration process.  For the purposes of a preliminary
analysis, a relatively simple model that is inexpensive to
operate is desired.

Model Description

A model fulfilling these requirements was developed for the 1978
Needs Survey.  The Continuous Stormwater Pollution Simulation
System (CSPSS) is described in detail by Wycoff and Mara;6 only
a brief summary will be given here.  The purpose of CSPSS is to
simulate all major urban pollution sources and the receiving
water response in a simple yet rational manner.  Basic functions
which may be simulated on a continuous basis are  (1) local
rainfall, (2) local runoff, (3) pollutant washoff, (4) sewer
system infiltration,  (5) storage/ treatment systems, (6) dry-
weather wastewater flow, (7) receiving water streamflow, and  (8)
receiving water quality response.  These modules may be executed
in logical sequential order (shown on Figure 2) to produce the
desired simulation.

The model determines the water quality response of the receiving
stream immediately downstream from an urban area.  All waste
sources, including urban stormwater runoff, combined sewer
overflow, wastewater treatment plant effluent, and upstream
flow, are considered.  Oxygen demands included are ultimate
carbonaceous BOD, nitrogenous BOD, sediment demand, and back-
ground dissolved oxygen deficit.  The only oxygen source con-
sidered is atmospheric reaeration.
                               597

-------
            MAIN
          (CONTROL)
                                           RAINFALL
                                            RUNOFF
                                           WASHOFF
                                             SEWER
                                            SYSTEM
                                          INFILTRATION
  DRY-
WEATHER
  FLOW
STREAMFLOW
 (UPSTREAM)
 STORAGE/
TREATMENT
              RECEIVING WATER RESPONSE
                  (STREAM/ESTUARY)
          FIGURE 2. General flow chart for CSPSS.
                        598

-------
The model generates  the  data  required  to construct a cumulative
frequency distribution curve  for  each  water quality parameter
considered,  including minimum receiving water DO concentrations
for both existing and proposed conditions.

Water Quality  and Alternative Assessment

The results  of the water quality  model simulation or observed
data can be  used to  determine whether  the water quality goals
are met.  If the goals are  not met,  then the severity of the
water quality  problem can be  ascertained.  Figure 3 shows a
hypothetical example of  a cumulative frequency curve for existing
conditions and for a proposed improvement to current pollution
abatement controls.  For this case,  simulation of the existing
conditions shows that the 5.0-mg/l  DO  level is exceeded 30% of
the time and 2.0-mg/l DO level is exceeded 2% of the time.   If
the DO criteria previously  presented are applied in this case,
then a water quality problem  is identified.

The effects  of improved  point and nonpoint source.controls can
be evaluated by CSPSS to determine  whether the water quality
goals can be achieved and what degree  of control is required.
Simulation of  a proposed pollution  abatement control shows
(Figure 3) that the  5,,0-mg/l  limit  is  exceeded only 2% of the
time and the 2.0-mg/l limit is no longer violated,  indicating
substantial  improvement  in  water  quality.  CSPSS can be used to
determine the  minimum degree  of pollution abatement required to
meet the water quality goals, hence avoiding costly overdesign
of control alternatives.

If a model such as CSPSS has  been calibrated to an urban area
receiving water system,  then  the  effect of reducing each control-
lable pollution source on receiving water quality can be inves-
tigated.  This is accomplished with multiple simulations, each
representing a different pollutant  loading condition.  The
suggested procedure  is to run simulations with all major pollu-
tant sources held constant  except one, which is varied.  Once a
series of runs are obtained and the results are analyzed, the
effect on receiving  water quality of reductions in pollutants
from that source is  known.  This  process  is repeated  until
pollutant reduction/water quality response  relationships  are
established  for each controllable pollution source.   These
relationships,  plus pollution  control  alternative cost  data,
become input to the  economic  optimization portion of  the  water
quality planning process.
                              599

-------
r
                      100
                                                                                 Proposed Pollution
                                                                                 Abatement Controls
                                                                                                 10
                                                    Minimum DO (mg/l)
                      FIGURE 3. Example of cumulative frequency curves for minimum dissolved oxygen.
                                                          600

-------
ECONOMIC OPTIMIZATION

Optimization is the process of making a system as perfect,
effective, or functional as possible.  Economic optimization  is
therefore the process of maximizing a system's output or produc-
tion while minimizing its cost.  In pollution control/ economic
optimization is the process of minimizing the costs  required  to
achieve a certain water quality objective.  Thus, the purpose of
the economic optimization portion of the water quality planning
process is to identify the technology or mix of technologies
which will achieve the water quality objectives at the least
possible total cost.

Production theory and marginal cost analysis as described by
Heaney and Nix7 can be used to identify the least costly combina-
tion of point and nonpoint source pollution control  alternatives.
In pollution control, a production function is the mathematical
relationship between the amount of pollutant removed or the
improvement in receiving water quality and the level of effort
applied.  The definition of level of effort depends  on the
particular control measure.  For example, the level  of effort
for streetsweeping has been defined as the fraction  of days that
streets are swept, while for sewer flushing it has been defined
as the fraction of sewers flushed daily.  The shape  of the
production function is governed by the "law" of diminishing
returns, which states that as an input to a production process
is increased with all other inputs held constant, a  point will
be reached beyond which any additional input will yield diminish-
ing marginal output.

In economic decisionmaking, marginal analysis determines whether
an action results in a sufficient additional benefit to justify
the additional cost.  It indicates that more intensive use
should be made of control alternatives with lower marginal
costs.  As these activities are expanded, marginal costs increase
to the point where other options may become competitive.  If
this occurs, then the optimal (least-cost) solution  to the
pollution control problem will include a mix of controls, each
utilized in an economiccilly optimum manner.

Figure 4 summarizes the economic optimization procedure for a
watershed served by a combined sewer system.  Two nonstructural
(or low structural) control technologies, streetsweeping (SW)
and sewer flushing (SF), are considered along with a previously
optimized storage/treatment system.  The nonstructural controls,
                               601

-------

a
 ^- C ^ C O  m  O
 ocoo^ooScj
      O    O   LL
                                                                        LU
                                                                        a
                                                                        LU LL O
                                                                        -1 Q. 1—
                                                                                                                •a
                                                                                                                cu
                                                                                                                ca
                                                                                                                •a
                                                                                                                CD
                                                                                                                a
                                                                                                                E
                                                                                                                o
                                                                                                                o
                                                                                                                 in
                                                                                                                 01
                                                                                                                 CD
                                                                                                                 c

                                                                                                                 QJ
                                                                                                                j-J
                                                                                                                 CD

                                                                                                                O




                                                                                                                 8
                                                                                                                 c
                                                                                                                 o
                                                                                                                 Q.
                                                                                                                 O
                                                                                                                 o

                                                                                                                'E
                                                                                                                 o
                                                                                                                 c
                                                                                                                 o
                                                                                                                 o
                                                                                                                 CD
                                                                                                                 0)
                                                                                                                 CO


                                                                                                                 CD

                                                                                                                 O
                                                                                                                in
                                                                                                          LU
                                                                                                          cc
                                                                                                          Z)
                                                                                                          C3
                                                                                                          u_
                                                                                    O
                                                                                    O
            jA/q|
                                      jA/q|
                                                      602

-------
SW and SF, are combined to produce a nonstructural  fractional
total cost (FTC) curve, which is combined with  the  storage/
treatment option, using production theory,  to produce  an  optimum
total- cost versus pollutant removal curve for the combined sewer
watershed.

In order to evaluate point sources, separately  sewered (i.e.,
urban stormwater) areas, and combined sewered areas, total cost
curves for each of the, cibove are obtained individually and then
combined, as shown on Figure 5.  It is noted that the  total  cost
curves shown on Figure 5 are expressed in terms of  cost ($/year)
versus some measure of water quality improvement (a).   The units
of V should be in agreement with the water  quality  goals  for the
receiving water.

For example, if the goal is to reduce  (or eliminate) DO occur-
rences less than 2.0 mg/1, then a should equal  the  number of
hours per year that DO levels less than 2.0 are reduced.  Con-
sider a case where low DO levels (i.e., <2.0 mg/1)  occur  for 100
hours per year.  Alpha would then be defined on the interval 0
to 100.  An a value equal to 75 would represent elimination  of
low DO levels for 75 hours per year and /a equal to  100 would
represent total elimination of low DO levels.   The  transforma-
tion between pounds of pollutant, removed per year and  improved
water quality defined in terms of the parameter a is made by
applying the results of the wa'ter quality modeling.  Thus there
is a direct linkage between the water quality simulation  results,
the water quality goals, and the economic optimization.

Once the total composite marginal cost (TCMC) curve for the
urban area is established (see Figure 5), then  the  economically
optimum combination of point and nonpoint source pollution
controls for any desired level of pollution abatement  is  known.
This procedure determines whether improved  wastewater  treatment
plant (WWTP) effluent quality, urban runoff controls,  or  combined
sewer overflow abatement are desired separately or  in  combination
so that a given water quality goal can be achieved.  The  degree
of pollution abatement required will determine  the  degree of
each control to be used.  At this point in  the  planning process,
an economically optimal project or combination  of projects has
been identified.
PROJECT EVALUATION .

Using information obtained from the water quality  simulation  and
economic optimization, it is possible , to develop relationships
for both optimum cost versus the amount of pollutant, removed  and
.optimum cost versus receiving water quality  response.   Several
levels of pollution abatement need to be evaluated by both  the
water quality simulation and the economic analysis to adequately
define these relationships.

                               603

-------
      CO
           Combined Sewered Basin (CSO)

            TC                 MC
                         co
           Separate Sewered Basin (SWR)

             TC                MC
       CO
            Point Source Controls (WWTP)

              TC               MC
                         a
                         to
                                                                    Total Basin
                                                        a
                                                        «o
                                         TCMC
                                                       .>
                                                       CO
        TCTC
          LEGEND:

          TC
          MC
          TCMC
          TCTC
          cv
Total Cost Curve
Margin Cost Curve
Total Composite Marginal Cost Curve
Total Composite Total Cost Curve
Indicator of Receiving Water
  Quality Improvement
Parallel Combination
of Three Options
FIGURE 5. Schematic for the economic optimization of control alternatives for urban areas served
            by both combined and separate sewer systems, with point source controls.
                                            604

-------
The optimum cost versus water quality relationship  is  an essen-
tial key to the project evaluation because it allows decision
makers an opportunity to determine whether the costs of achieving
a water quality goal are worth the expected improvements in
water quality.  For example, the hypothetical optimum  cost
versus water quality curve shown on Figure 6 indicates that the
cost of meeting the goal of allowing only one 4-hour period per
year with a DO concentration below 2.0 mg/1 is four times the
cost of allowing five low-DO periods.

Ideally, this question should be addressed in light of expected
annual damage resulting from low-DO situations.  For example,
Figure 7 presents a hypothetical relationship between  expected
annual damage and annual duration of low DO levels.  Damages
should be measured in terms of dollars; however, other units,
such as number of fish kills, reduction in fish yield, or effects
on species diversity, would be useful.  Such a relationship may
very well show that five low-DO occurrences (Figure 6) are no
more damaging than one occurrence.  By changing the water quality
goal to allow five low DO periods, a significant monetary savings
could be realized while still protecting the receiving water's
fish and wildlife resource.

Unfortunately, the basic data required to construct a  damage
versus water quality curve are generally lacking, making this
type of relationship difficult to quantify at the present time.
Thus project evaluation currently involves a large  degree of
intuitive judgment, particularly in the area of damage or,
conversely, benefit assessment.  However, every effort should be
made to obtain or develop the basic data necessary  to  quantify
this relationship.

If, on the other hand, the optimum cost of meeting  a water
quality goal is too high and the goal cannot be modified and
still protect the beneficial use of the water, a change in the
water's intended use may be required.  The cost versus water
quality curves could be used to provide guidance to suitable
goals and uses that are monetarily feasible.  This  approach can
be used in staged implementation of pollution abatement projects.
Consider the case of a stream that once supported a cold-water
fishery, but due to oxygen-demanding substances, from  wastewater
treatment plant effluents, combined sewer overflow, and urban
runoff, extended periods of low DO concentrations had  rendered
the stream unusable even as a warm-water fishery.   The optimum
cost versus water quality analysis might show that  achieving
cold-water fishery goals is beyond present fiscal resources, yet
a warm-water fishery goal could be achieved.  Rather than aban-
doning the project altogether, facilities required  to  achieve
the warm-water fishery goal could be constructed immediately,
with facilities required to meet the cold-water fishery goal to
follow if and when funds become available.
                               605

-------
^
,a
o
      60 I
                          0 hr/yr < 2.0 mg/l
       0
                                                              100% Elimination of

                                                              Low DO Occurrence
                                                            175
200
   0      25      50     75     100    125    150



             a. Low DO Occurrence Eliminated, hr/yr



FIGURE 6. Hypothetical cost vs. water quality improvement curve.
     to

     Q
     C
                                                               Existing Conditions
25
1
50
1
75
i
100
1
125
150
175 20
          0



                         Low DO Occurrence-hr/yr (176-a)




           FIGURE 7.  Hypothetical damage vs. water quality curve.




                                      606

-------
PHASE II - DETAILED ANALYSIS

If the results of the Phase I analysis  indicate  that  there  are
water quality problems that can be corrected with affordable
solutions, the detailed analysis of Phase II is  required.   This
phase considers, in. much greater detail, the area's hydrology,
combined and separate sewer system hydraulics, nonpoint  and
point source pollutants, and the receiving water quality response.
This analysis may require more sophisticated data and  models,
depending upon the magnitude of the project.

The product of the Phase II analysis is a detailed description
of the optimal mix of control alternatives, the  total  plan
costs, and the receiving water quality  response  due to the  plan.
The description of selected nonstructural alternatives includes
the level of effort required, the area  to which  the alternative
applies, the expected pollutant reduction due to the  alternative,
and the cost of the alternative.  The description of  selected
structural alternatives includes a preliminary engineering
design of the facility^ the expected pollutant reduction due to
the facility, and the cost of the facility.  These are the  items
generally necessary to complete a Step  I  (201) facilities plan
for the study area.


PHASE I - CASE STUDIES

A brief description of two case studies developed as  part of the
combined sewer overflow and urban stormwater runoff portion of
the 1978 Needs Survey2 are presented to illustrate selected
steps in the water quality planning process.  The Philadelphia,
Pennsylvania, case study presents the simulated  response of the
Delaware River to improved WWTP effluent, and to combined sewer
overflow and urban stormwater runoff controls.   The Des  Moines,
Iowa, case study demonstrates the linkage between water  quality
simulation and the economic optimization procedure.

Philadelphia, Pennsylvania

The City of Philadelphia is located in  eastern Pennsylvania at
the confluence of the Delaware and Schuylkill Rivers.  The  study
area consists of the City of Philadelphia on the western bank of
the Delaware River and Pennsauken, Camden, and the neighboring
urban areas on the eastern bank of the  river.  This area is
heavily industrialized and contains many  industrial waste sources
as well as numerous municipal wastewater  treatment plants.
Approximately 45 percent of the study area is served  by  combined
sewers, while the remainder is either drained naturally  or  is
served by storm sewers.

Continuous observed DO data for the Delaware Estuary  at  Chester,
Pennsylvania, are available and were used to calibrate the
                               607

-------
receiving water response portion of CSPSS, as  shown on Figure
8a.  This figure shows good agreement between  simulated and
observed water quality, particularly in the range of 2.0 to 6.0
rag/1,  and further indicates that the Delaware Estuary is subject
to low DO (i.e., less than 2.0 mg/1) levels approximately 40
percent of the time.

The results of simulations with different pollution abatement
alternatives are presented in Figure 8b.  Based on the simula-
tion, improved treatment of domestic and industrial wastewater
can result in significant improvements in the  receiving water
quality, more so than storage and treatment of CSO and urban
runoff.  However, even with very high levels of municipal waste-
water treatment, DO occurrences at or below 2.0 mg/1 cannot be
completely eliminated.

A cost analysis-was not undertaken for Philadelphia, but it was
considered in the next case study.

Des Moines, Iowa

The Des Moines urban area is located in central Iowa at the
confluence of the Des Moines and Raccoon Rivers.  The study area
consisted of the Des Moines urban area tributary to the Des
Moines and Raccoon Rivers.  The extent of the  combined sewer
watershed is about 8% by area.  The remainder  is served by
separate storm sewers or by natural drainage.  The receiving
water consists of the Des Moines River below the confluence with
the Raccoon River, extending to Red Rock impoundment approxi-
mately 45 miles downstream.  Major BOD sources include upstream
flow (81%), municipal wastewater effluent (8%), combined sewer
overflow (1%), and urban stormwater runoff (10%).

Results of the simulation indicate that under  present conditions,
minimum DO levels in the Des Moines River are  less than 2.0 mg/1
134 hours per year and less than 5.0 mg/1 1,480 hours per year,
about 17% of the time.  These results indicate the existance of
a water quality problem by the criteria developed previously.
Furthermore, a typical stream quality "standard" would call for
minimum DO values to be greater than 5.0 mg/1  "at all times."
This standard would be violated 17% of the time under existing
conditions.

The simulation also indicates that a high level of control
(i.e., BAT) for all controllable BOD sources,  including WWTP
effluent, combined sewer overflow and urban stormwater runoff,
will eliminate all occurrences less than 2.0 mg/1, but DO levels
less than 5.0 mg/1 cannot be eliminated.  Even with BAT of all
controllable BOD sources, simulation results indicate that
minimum DO levels will be less than 5.0 mg/1,  approximately 900
hours per year, or 10% of the time.  Thus, a typical stream
standard requiring that DO levels always be greater than 5.0 mg/1

                               6SD8

-------
   100
    90
c
o
-o
c
c
to
§
    80-
    70-
    60-
7   50
o
o
4-»

~m

cr
LU



5   40
Q
o>
    30
    20-
    10
                              Simulated
                                                            10
12
—i

 14
                                    DO (mg/l)
  FIGURE 8a.  Receiving water dissolved oxygen calibration at Chester, Pennsylvania.
                                         609

-------
c
o
•*3

1
.i
_e

£
•M
c
ra
J=
UJ
.a
0
Q>
a.
      100  -i
      90  -
      80  -
      70  -
      60  -
      50  -
      40 -
      30 -
      20 -
      10 -
                                                                    Existing Conditions

                                                                    Secondary (30/30/10) Effluent,
                                                                    No CSO or Urban Runoff
                                                                    Controls

                                                                 — 20/20/9 Effluent No CSO
                                                                    or Urban Runoff Controls
                                                        minium
                                                                    (30/30/10 Effluent With
                                                                    High Level of CSO and
                                                                    Urban Runoff Controls

                                                                    5/5/2 Effluent No CSO
                                                                    or Urban Runoff Controls
                                \
                                4
                                                                 10
12
14
                                        Minimum DO (mg/l)

  FIGURE 8b. Simulated receiving water response to improved wastewater treatment plant
               effluent and CSO and urban runoff control.
                                            610

-------
cannot be met regardless of the degree of pollution control
provided.  Therefore, the economic optimization of control
alternatives presented here is based on minimizing the occur-
rence of DO levels less than 2.0 mg/1.

Figure 9 presents production functions for point and nonpoint
source control alternatives developed from multiple 5-year
computer simulation runs.  The output, or indicator of perform-
ance a, is defined as the number of hours per year DO levels
less than 2.0 mg/1 are reduced and the level of effort is defined
as the amount of BOD removed, in Ib/yr.  These production func-
tions indicate that additional WWTP (i.e., point source) controls
are largely ineffective in reducing low (<2.0 mg/1) DO levels,
whereas urban stormwater runoff (SWR) and CSO controls are
generally more effective.

This conclusion is valid for this study site, but these results
cannot be generalized to other areas since a case-by-case anal-
ysis is required.

Figure 10 presents estimated total cost curves for point and
nonpoint source control alternatives.  The wastewater treatment
cost curve is the estimated equivalent annual cost of pollutant
removal beyond the secondary level and the CSO and SWR cost
curve is the equivalent annual cost of pollutant removal for the
optimum mix of CSO and SWR controls.2

Marginal cost analysis reveals that for the entire range of
water quality improvement considered, CSO and SWR controls are
more cost effective  (i.e., lower marginal cost) than additional
point source control.  The optimum areawide total composite
marginal cost (TCMC) curve and the total composite total cost
(TCTC) curve are presented on Figure 11.  These curves provide
valuable information to the decisionmaker relative to the eco-
nomic tradeoffs of water quality improvements.

The marginal cost curve (Figure lla) illustrates the increasing
extra costs incurred when total or near total elimination of low
DO levels is the water quality goal and the total cost curve
(Figure lib) illustrates the increasing total project costs.
For example, in order to eliminate all DO levels less than 2.0
mg/1, we must be willing to pay up to $350,000 for the last low
DO hour eliminated and $6.2 million per year for the total
project.  Alternately, if we decide that DO levels less than 2.0
mg/1 are acceptable for a period not to exceed 4 hours per year
(i.e. a = 130), then we have also decided that we are willing to
pay up to $270,000 for the last DO hour eliminated and $5.2 mil-
lion per year for the total project.

One provision of PRM 75-34 states that the marginal cost of
control should not be substantial compared to the marginal
benefits.  For this case, if a is considered an adequate measure

                               611

-------
                             1             2            3

                                           V
                                BOD Removed—million pounds/yr
        FIGURE 9a. Production function for wastewater treatment beyond secondary,
                    Des Moines, Iowa.
  150-1
                         a = 134 (100% Reduction)
                              BOD Removed—million pounds/yr
FIGURE 9b. Production function for CSO and SWR control, Des Moines, Iowa.
                                  612

-------
                          12-,
                          ID-
                           6-
                       o
                       J.  4-
                       o
                       o
                          2-
Cost—Total Annual Cost of Treatment
     Less Annual Cost of Secondary
     Treatment.
                                     -" X BOD Removed =
                          Total Annual BOD
                          Removed Less BOD
                          Removed By Secondary
                          Treatment.
                            0       1       2       34

                              BO'D Removed—million pounds/yr

        FIGURE 10a. Estimated cost of wastewater treatment beyond secondary for 40-mgd plant,
                     DesMoines, Iowa.
                          7-1
                          6-
                          5-
                      I  3
                      I.
                          2.
                          1-
 Total cost curve is based on optimum
 mix of structural and nonstructural
 CSO and SWR control alternatives.
                                  BOD Removed-million pounds/yr

FIGURE lOb. Estimated cost for optimum CSO and SWR control alternatives for Des Moines, Iowa.

                                             613

-------
                        350
                       300-
                       250-
                   «    200
                   a
                                        $350,000/a
                                  25
              50
100   125
                                   a, Low DO Occurrence
                                     Eliminated, hr/yr
                                                                  4hr
                                                                  yr
FIGURE 1 la. Marginal cost curve for optimum control alternatives, Des Moines, Iowa.
                   C
                   O
                   E
7-


6-

5-


4-


3-


2-
                              Maj^Cost = $6.2(10^)/yrj9> oj= 134
                               BOD Removed = 4.5(10)" Ib/yr
                                    = $5.2(10h)/yr@a
                               BOD Removed = 4.3 (10)b Ib/yr
                             Cost= $2.8(10" )/yr@ a = 109
                             BOD Removed
                               3.1 (10)" Ib/yr
                                    a. Low DO Occurrence
                                      Eliminated, hr/yr

 FIGURE 11b. Total cost curve for optimum control alternatives, Des Moines, Iowa.

                                        614

-------
of benefits, the implied water quality goal is elimination  of
109 of the 134 hours per year with receiving water DO  levels
less than 2.0 mg/1.  Marginal cost in this case would  be  $60,000
for the last low DO hour eliminated and total cost would  be
$2.8 million per year.

It is obvious that a decision relative to acceptable receiving
water quality will have a significant impact on total  committment
of both Federal and local tax revenue.  It is also obvious  that
a relationship between the water quality indicator   and  actual
environmental damage (see Figure 7) would be of great  value to
the decisionmaker.

These case studies illustrate the site-specific nature of the
water quality planning process.  Continuous water quality simu-
lation of the Delaware River below Philadelphia indicates water
quality improvements due to AWT and little improvement due  to
SWR controls.  On the other hand, continuous water quality
simulation of the Des Moines River below Des Moines indicates
nominal improvement due to AWT compared to CSO and SWR controls.
Considering the number of physical, biological, and economic
variables involved, it is unlikely that the exact same mix  of
controls or receiving water quality goals will provide the
optimal solution for more than one situation.

SUMMARY AND CONCLUSIONS

Water quality planning in an urban area is a complex and  often
interative process.  Once the goal of secondary treatment of
municipal wastewater is met, identification of residual water
quality problems and their cost-effective solution must be
addressed on a case-by-case basis.  This paper has presented a
general framework for water quality and wastewater facilities
planning which consists of answering seven basic questions
relative to the urban area and its receiving water.  These  are:

1.   What is the intended beneficial use of the receiving water?

2.   What water quality goals or criteria are required to ensure
     this use?

3.   Based on these;goals, is there a water quality problem?

4.   Can combined sewer overflow, urban runoff, or additional
     point source pollution abatement techniques solve the
     problem, and what degree of control is required?

5.   If the problem can be solved, which controls, in  general,
     are the most cost-effective?

6.   What is the nature of the tradeoffs between cost, degree of
     control, and receiving water quality?
                               615

-------
7.   Considering these tradeoffs, are  changes  in  the  water
     quality goals or in the desired beneficial use of  the
     receiving water indicated?

These questions can be answered by application of tools currently
available to the water quality planner,  including (1) continuous
hydrologic/water quality simulation and  (2) economic  optimization
of control alternatives.

Urban water quality problems arise from  intermittent  nonpoint
sources of pollution, such as combined sewer overflow and urban
stormwater runoff, as well as from point sources,  such  as muni-
cipal and industrial wastewater effluents.  The intermittent and
extremely variable nature of nonpoint  source pollution.requires
consideration of receiving stream interactions on a continuous
basis.  Selection of "design events" for analysis will  not give
the water quality or facilities planner  a true picture  of the
behavior of the receiving water or of  the impact  of proposed
pollution control facilities.

Economic optimization procedures, such as production  theory and
marginal cost analysis, are available  and are  applicable to the
urban water quality planning process.  Considering the  availa-
bility of practical economic optimization tools,  there  is no
reason why suboptimal solutions should be considered.   It is
always possible for a given set of economic circumstances to
determine the optimum mix of feasible  technologies which will
achieve a given level of pollutant removal from one or  more
sources.  In many cases, it is also possible to link  the results
of continuous receiving water quality  simulation  to the economic
optimization and thus establish the optimum mix of control
alternatives required to achieve a given water quality  goal or
an array of alternative water quality  goals.   Given these eco-
nomic/water quality tradeoffs, the most  appropriate level of
water quality improvement and/or project cost  can be  chosen.
Knowledge of actual environmental damage or benefits  versus
water quality indicators would be invaluable in this  decision-
making process.  Unfortunately, such information  is generally
not available.

ACKNOWLEDGMENTS

Credits

The development of continuous stormwater pollution simulation
system (CSPSS) and the case studies were funded as part of the
1978 Needs Survey for control of pollution from combined sewer
overflow (Category V) and urban stormwater runoff (Category VI),
contract No. 68-01-3993.  Philip H. Graham, P.E.,  was the project
officer.  The authors would like to thank their colleagues at
CH2M HILL for their extensive support  and encouragement in the
preparation of this paper.

                               616

-------
REFERENCES
1.
2.
3.


4.



5.
7.
U.S. Environmental Protection Agency.  "1978 Needs
Survey—Cost Estimates for Construction of Publicly
Owned Wastewater Treatment Facilities."  EPA 430/9-79-001
(FRD-1).  (February 10, 1979).

Wycoff, R. L., J. E. Scholl, and S. Kissoon.  "1978
Meeds Survey—Cost Methodology for Control of Combined
Sewer Overflow and Stormwater Discharge."  EPA 430/9-79-003
(FRD-3).  (February 10, 1979).
U.S. Environmental Protection Agency.
for Water."  (July 1976).

Gehm, H. W., and J. I. Bregman  (ed.).
Water Resources and Pollution Control.
Reinhold Company.  (1976).
"Quality Criteria
Handbook of
 Van Nostrand
Shepard, M. P., "Resistance and Tolerance of Young
Speckled Trout (Salvelinus Fontinalis) to Oxygen Lack,
with Special Reference to Low Oxygen Acclimation,"
Journal of the Fisheries Research Board of Canada 12,
387-446.   (1955).

Wycoff, R. L., and M. J.Mara, "1978 Needs Survey—Continuous
Stormwater Pollution Simulation System—Users Manual."
EPA 430/9-79-004.  (FRD-4).  (February 10, 1979).

Heaney, J. P. and S. J. Nix.  "Stormwater Management
Model:  Level I—Comparative Evaluation of Storage-Treatment
and Other Management Practices."  EPA 600/2-77-083.
(April 1977).
                               617

-------
                                     10
          WORKSHOP ON PRACTICAL APPLICATIONS AND RESEARCH NEEDS FOR
               RECEIVING WATER RESPONSES TO URBAN STORMWATER

                                 SESSION I

           Moderators:   Francis J. Condon, EPA
                        James P. Heaney, University of Florida
         Speakers:  1) James P. Heaney, University of Florida
                    2) Raymond P. Canale, University of Michigan
                    3) Thomas L. Meinholz, EcolSciences, Inc.
                    4) Robert E. Pitt, Woodward-Clyde Consultants
                    5) Miguel Medina, Jr., Duke University
Mr. Condon

      Five minute presentations will be given by Messrs. Heaney, Canale,  Mein-
holz, Pitt and Medina to summarize what they said yesterday and give you  time
to put down some questions.  After they all have finished then we'll open it
up for discussion.  You may have questions related to special  project findings,
problems encountered and associated solutions, receiving water impact defi-
ciency and research needs, short term and long term effects, and assessment
of the adequacy and reliability of the data base for receiving water studies.
Can you identify the most crucial gaps in deficiencies in the  data base?
What is the role of the indicator parameters?  How can we decipher urban  wet
weather impacts from dry weather impacts?  Is it worth having  a complex model
or would we be better off with more of the black box models?  The last word I
would like to leave  with you,  the price you have to pay for  this discussion
is to identify for us some of the real world research needs, which will then
also give us justification for trying to get the funds to accomplish them.
So it works two ways.

1) Dr. Heaney

      Good morning.  I think the message is pretty clear, we've only just be-
gun to look at receiving water impacts and I would like to suggest we start
fresh.  The Sutron work in particular, looking at the continuous dissolved
oxygen over a year, two years, three years, suggests a lot of  different ideas
that have been traditionally used for looking at the water quality standards.
So, I think that's one definite area.  Associated with that is the whole
question of reliability, and we found numerous cases of incidences of spills,
                                     618

-------
and we've got the seven day, 10 year minimum flow still  in the back of our
rnind.  What about the reliability of sewage treatment plants?  How frequently
do they spill?  Industrial  waste discharges?  How well  are your ordinances
being enforced against discharges?  Would you really know of some of these
discharges if it wasn't that bad?  All of these are going to give you a
system reliability that's going to determine what your limiting ability is to
control pollutants.  We need to look at the percentage of the time we can
expect to achieve certain quality water conditions. .We've.also been doing
work on a data base for EPA for four or five years trying to collect what
information is available.  It's very costly and expensive to do that.  As
far as research needs on that, I don't see anything too glamourous that needs
to be done except there is  a lot of hard work in checking and verifying the
data.  I think there has,been kind of a loss of "back to basics".  There is
not good quality assurance  in data collection, reporting and dissemination
and we found even the best  data still has a lot of garbage in it.

      We thought it was going to be an easy project - we get the data on a
tape and then you just put  that on a larger tape and then make that available
to users.  But, if you really get in there and look at that data, there is a
lot of bad information in there and we don't have good checking procedures.
We need to establish some quality assurance in data handling that we don't
have now.  There are a lot  of ways for processing data retrieval storage
systems.  The capability is out there, but I still think we have to do a
better job of putting it in clean so it will come out clean.

      As far as modeling goes, during the last four or five years, you couldn't
mention the work "model" and hope to get any money for research.  Instead,
you have to talk about model application.  Maybe we have over-modeled, but
there is still a need there to' do more fundamental work.  Maybe we have enough
of the differential equations models.  We don't have standard methods yet for
models, we're just beginning to think about this, and I think this is a fruit-
ful area for some research.     .       ......   .

      The other problem we  dealt with is the question of economic efficiency,
cost effectiveness, and you have heard a Tot about that, pro and con.  There
is a lot we can do selectively to find those receiving waters whose benefi-
cial uses are most seriously impaired.  Economic .optimization should be
based on selective control  of the most serious problems and beneficial uses
of receiving water.  When you go to selective abatement, the burden of proof
lies on the regulator to prove that, in fact, this is the best way to go.  So
the analysis part of the whole pollution control problem becomes a much larger
cost.                                                             .           '

2) Dr. Canale

      I would like to share some problems that I have encountered in the appli-
cation of receiving water quality models for evaluating the impact of CSO's.
I want to talk about developments calibration, verification of the model  and
some problems encountered in trying to abuse these calibrated models for water
quality projections.

      It is possible to develop a standard type differential equation mass
balance model to simulate the event of a combined sewer overflow problem.

                                      619

-------
The response of Onondaga Lake to a storm in September, 1976 shows an increase
followed by the die-off of the organisms.  Also, a similar kind of response
was noticed for oxygen in various parts of the lake.  These responses can be
modeled but we need various kinds of physical data, more biology, circulation,
dispersion, material loadings from point sources, agricultural sources, urban
runoff, etc., reaction rates, and then finally we need to input various envi-
ronmental conditions such as temperature for the case of dissolved oxygen.
Many of the models are very sensitive to the activities of algae, so we need
to know something about light, light extinction, incident light, the level of
algae, chlorophyll and nutrients.  These have to be specified prior to making
calculations, so during the calibration or verification stage, you may be able
to measure these factors and simply use them as input parameters in your
model.

      The problem comes up, what do you do for, let's say, chlorophyll-a level
when you are doing a projection?  You have this model, you've calibrated it,
you want to make a projection, you need to assume something regarding the
chlorophyll, nutrients, pre-storm conditions, light and temperature.  Well, if
you are in a situation where the assumption you make regarding the chloro-
phyll level can be an overwhelming factor and be even more important than how
you may manipulate material loadings.  In that case, the problem turns out to
be pretty arbitrary depending on what you assume for chlorophyll.

      During this period, June to September, 1976 in Onondaga Lake, you can see
that the surface dissolved oxygen was varied from ten milligrams per liter
all the way down to less than 2 milligrams per liter.  This is an overwhelm-
ing factor that overshadows the impact of any combined sewer overflow.
At the same time, the chlorophyll is doing the same kind of variation.  These
kinds of variations, independent of the combined sewer overflow, are very
serious problems in applying models.  If we are dealing with different para-
meters for control, a different time scale for the different kinds of model-
ing approaches may be required.

      In this particular case in Onondaga Lake, we went through and developed
a model for chlorophyll and BOD and TKN and DO which were daily average values
over a season which could then be input and used as initial conditions or
baseline conditions for any model attempting to respond to a short time event
such as a CSO.  We also would recognize that this time scale problem would
be an important consideration in other kinds of non-point source control,
evaluation, and of course point source control evaluation as well.

      The Genesee River in New York showed the critical DO sag as a function
of the volume of discharge of CSO for various levels of treatment.  This was
with zero treatment, and a small amount of discharge to increase the amount
of discharge the critical DO increases.  This whole graph may be sliding up
and down by several milligrams per liter depending on what was assumed for
the incident level.  This is a very important consideration because you can
see that dependent on what you assume, you may be above or below some stan-
dard.

      The final point would be that in order to evaluate the cost benefit type
of considerations in this kind of problem you must have more capability than
                                     620

-------
simply being able to project or calculate an incremental effect of the com-
bined sewer overflow or urban runoff.  You must have the ability to project
the overall value or overall concentration.

3) Mr. Meinholz

      Some of the unique problems that occurred in Milwaukee related to sedi-
ment.  We were not able to identify in our early stages of the project, the
source of the mechanism that was causing the drastic DO problems in the
Milwaukee River.  Through long term field investigations we were able to iden-
tify that sediments were there and causing the problem;  that it was more of
a physical problem, a scouring action directly related to overflow volumes
and rainfall intensities.

      In identifying sediment sources in a municipal area such as Milwaukee5
where the downstream is nothing more than Lake Michigan, we tried to quantify
the upstream sources, the sediments in the storm sewer areas and what they
were contributing, as well as the combined sewer areas.  What about instream
velocities?  Are they critical enough to scour sediments and how does it
impact other receiving streams that are not like Milwaukee?  How does the
depletion exert itself as a function of sediment scour?

      The long term buildup capabilities of different areas, different streams
are also another research need that we have identified.  The Milwaukee sewer
district is undertaking a program to identify what would happen if sediments
are removed.  If you have to remove sediments and buildup will reoccur again
with time, it is awfully difficult to justify the expenditure in order to do
something like that.

      What are the effects of various control strategies?  If Milwaukee goes
to deep tunnels or any municipality goes to sewer separation or some type of
satellite treatment can we, in fact, through a limited program since construc-
tion is quite a few years away in most instances, can we verify some of our
predictions of what the improvement in water quality will be?  If in fact,
we can predict the improvement, can we then relate it to benefits?  That is
probably the biggest need right now.

      The whole area of toxics we have touched on in a number of topics in
the last two days.  In Milwaukee we are trying to relate this to sediment
problems.  Years and years ago when the industries, in and around Milwaukee,
especially the plating and canning industries, were discharging to the rivers,
some of the sediment core samples showed evidence of very high levels of
cadmium and other metals two to three feet below the water-sedimeiiit interface.
These toxics are directly related to some of those industries in the lower
reaches and we will be paying for these mistakes that were made by industrial
as well as municipal discharges in years to come.  If sediment scouring
or dredging occurs, do all these toxics come up to the surface?  The impact
there may become much greater than if we simply went with abatement techni-
ques without dredging.
                                     621

-------
     In Milwaukee the. difference between wet and dry weather impacts is ex-
tremely noticeable.  In other municipalities it may not be as dramatic.  We
identified our wet weather impact strictly on a physical basis.

    And finally the whole portion of complex versus simple models.  Complex
models were sold as being the answer, yet in the final analysis we're still
some way from the final answer.  It's the basic understanding of the problem
and the ability to predict the basic response to the problem that's most
important.  It is required to identify in a simple model what the problem
is, what the source is, and what the mechanism is.  After we have identified
the source and mechanism using simple models it may become possible to use
the complex model for more verification.

4) Mr. Pitt

     Basically, most of the research effort that has to be expended in the
next several years should be directed to eliminate beneficial use degradation.
Right now I think we really have to examine site specific biological degrada-
tion and relate that to the chemical conditions in the receiving water.  A
major biological concern would be relating the loss of the habitat where the
reduction of a component in the aquatic food chain would reduce the number
and type of a specific organism in the habitat which is the ultimate com-
ponent of concern.  If we are concerned about fish, that doesn't necessarily
mean that we monitor fish, but critical components for the food chain.  Now,
there are many reasons why they also monitor fish, but fish are easy to catch
and relatively easy to identify, but because they are mobile they are very
difficult to relate to a specific location in the receiving water.  If your
receiving water is not homogeneous, which is most likely the case, a lot of
their food materials are somewhat less mobile and therefore somewhat easier
to relate to specific problems within their receiving water.  The beneficial
use impairment should be based on short and long term conditions.  Short term
is a lot easier and that is basically what we have always been concerned
with historically, relating to water quality conditions.

     We have also found that sediment conditions can exert longer term bio-
logical degradation in receiving water and we don't really have a very good
handle on how to address that as far as the need for control.  You need to
monitor sampling locations affected by various levels of degradation.  Control
test areas and intermediate situations should be monitored in order to try
to find or establish an acceptable biological degradation situation and then
relate that to the chemical conditions found at that location and affecting
that from upstream areas.  You have to sample similar habitats or a family
of habitats at all your stations.  You have to be careful about locating
specific sampling points near local points of interference such as areas
where there is an excessive amount of bank erosion, or ground water infiltra-
tion.  You have to worry about locating extremely close to something like an
outfall where the specific effects from one outfall might overly influence
smoother gradation of conditions that you are attempting to look at and of
course the micro analysis of the degradation from a single outfall as we
heard from Seattle is an extremely important consideration also.  But, you
                                     622

-------
have to separate the physically short-term versus, the longer term types of
problems that you are observing.  The biological "'sample, has to include
representatives of various levels of the food chain including invertebrates,
algae, fish, etc.  As I mentioned, fish are relatively easy to sample and
identify, but quite difficult to relate to a specific chemical condition at
a location.                                 ,      .

5) Dr. Medina

      My comments are basically directed to the fundamental research needs and
problems from the complex interactions and interrelationships of urban storm-
water controlled systems and receiving water responses.  When we look at the
large number of variables I think they can be grouped broadly into three main
categories:  a natural science perspective in order to get a better under-
standing of the chemical, biological and physical  phenomena; we can look at
them from an engineering perspective to achieve some type of useful purpose,
for example deterministic or statistical prediction of transfer processes, .
storage treatment, the design of conveyance and containment systems and so
forth; we could also look at it from a management planning perspective to
achieve some degree of control over the state of the system within the con-
text of a rational decision-making process.  The basic engineering problem
has always been to obtain amicable relationships between the inflow and the
outflow processes, and the variables describing the state of the system.

      For example, such variables as density, volume, temperature, and so
forth can be considered state variables.  Decision variables can be how much
storage and treatment you are going to implement.   Regardless of whether you
group these variables in any one of -these categories, we have to converge.
We have to use a multi-disciplinary approach and converge towards a unified
purpose.  I describe that as acceptable levels of receiving water quality.
Among the generally recognized needs, I'm going to quickly go through some
of these; geophysical, hydrodynamic characterization of natural receiving
water systems.                                                         .

      We still don't know enough about the geophysical hydrodynamic problems.
Especially, identification of the transition.  We have fresh water, tidal
rivers and estuarine systems.  What happens in the transition and what physi-
cal-chemical biological reactions are unique to these?  Specification of
heterogeneous kinetic interactions of all the sequentially reacting substances
- we don't know enough about that.  We could model better if we had more
basic research in this area, such as boundary effects on various substances
at the air-water interface.

      We also need to establish better analytical  relationships for land
eroded suspended matter, and I think that one of the real problems is how
to get a good handle on the up-land erosion.  I think once we have the sedi-
ment in a particular channel, we can transport it, but relationships like the
universal soil loss equation are too highly .empirical and they are not storm
event dependent.  In other words, there is a lot to be done in that area. ,A
fundamental understanding of the accumulation and final disposition of syn-
thetic pollutants produced by our chemical and industrial firms through the
natural food chains that exist in the wastewater treatment systems, as well
as in the receiving water quality systems, is needed.

                                      623

-------
      We need to have a better handle on the dynamics and control of wastewater
treatment plants and stormwater storage treatment systems.  We need a better
integrated receiving water response to time-variable point and distributed
waste inputs.  We also need to continue to develop refined methodologies
within an analytical framework that will evaluate cost effectiveness for
these urban wastewater control strategies in terms of ultimate receiving
water quality benefits.

      Those are general comments.  Now I want to sort of get into some speci-
fics;  problems that I feel are critical in receiving water quality modeling
for urban stormwater management planning.

      Getting back to kinetics formulation, we have eight or nine equations
for reaeration rate for streams, but we need to have more fundamental re-
search on the separate and combined effects of surface active agents at the
air-water interface, reduction of turbulent mixing by high suspended sediment
loads and dissolved chemicals.  All of these tend to reduce the oxygen trans-
fer rate.  Some work by Covar, for example, is of some interest whereby he
presented a general approach to selection of the appropriate reaeration
equation for streams, now a similar approach is needed for lakes and estuaries
and hasn't been done.  There really hasn't been enough done on the relative
magnitude or importance of all the parameters in these equations either.

      If we get into the deoxygenation rate, the k-j factor, as we like to
call it, we still use a first order rate model, which does not include speci-
fic growth rate which makes it really a second order reaction.  I think more
basic research is needed there.  The whole traditional BOD model is under
challenge.  We also do not have a good handle on the time dependency of all
these rates during high river runoff conditions.  Very little work has been
done in this area.

      Now, going to assessment tools,  two basic assessment tools you need are
instrumentation and mathematical models of some sort.  I'm not as qualified
to talk about instrumentation as somebody like Phil Shelley who is in the
audience and has done quite a lot of work in this area, but everybody knows
we need better instrumentation.  I think one of the critical needs is to
optimally design the networks, both the water quantity, rain gauge and
stream flow, as well as the water quality networks using a better methodology.
In other words, analyzing the character of the hydrologic signals and then
determining the response time that is required for the instruments and so
forth.

      Going into mathematical models, I will only briefly say that we need
to simplify first and then use the complex models.  General guidelines for
model selection, I think, is the key issue.  Most of you do feel like you
need some general guidelines for selecting your particular model for a par-
ticular situation.  Now model development needs again span such things as.
simplification of routines for storage treatment, for dual purpose operation
control systems, and complex mechanisms including transport.  I want to
get down to the planning and management techniques and very briefly say that
in a full scale testing of all of these management techniques, such as
detention, storage and so forth, we have to quantify benefits and unless we
                                     624

-------
relate water quality pollution strategies to the benefits derived, we are
really going to fall on our faces.

Moderator - Dr. Heaney

      Would anyone like to make a comment?  Ask a question?  Besides Peter
Moffa.  I'm going to make a. few comments, and I hope they will  be responded
to.  I have to agree that one of the areas of the toxic picture, and a very
important part at that, is to evaluate the toxics in terms of the specific
chemistry of that receiving water and the synergistic effects.   But, what I
really want to get to, and it relates somewhat to the toxics, is that we've
got to begin now to collect the data.  You know we can do stormwater sampling
and wet weather impact work as part of the CSO element of facilities 201
planning.  This is fundable, it's been accepted, somewhat hesitantly.  You
can go out and sample before a storm, during and after a storm, but you can't
spend very much money collecting that background data.  So, I think, we have
got to work through our EPA representatives here as well as back home in the
regional offices to "get them to recognize that we've got to establish a fund-
ing vehicle to begin to collect this data now to go buy a data base and not
wait until you go in with a CSO or proposed CSO program.  I think it's all
related to what you've said and I'd like to see some responses.

Question:  Is there anybody who knows EPA allocations for data collection?

Answer:  I haven't been able to trace down the figures, but the percentage
of the total EPA dollar that goes into the hardware is quite high and the
amount that goes into the data collection of studies has definitely gone down
over the past years because of this need to move ahead and implement more
construction grants.  So, hopefully, we're seeing a shift a little bit away
from that, and a little bit more towards analysis and forethought before you
jump into these things.

Mr. Greg Welter, O'Brien & Gere Engineers, Washington, D.C.

      I have a comment regarding the question of biological benefits.  We
have had some people that have approached this by doing natural biological
monitoring longitudinally down the stream and they have been able to establish
a variation as we go further downstream.  But, as far as a practitioner or
somebody working for a client and trying to recommend facilities and demon-
strate a retrievable benefit, we haven't seen those studies carried to the
point of being able to say, "Well, if we implement this alternative we can
see this kind of change, we'll see this difference."

      Where we've got a degraded habitat or something like that, we'll see a
different biological regime.  Now, in terms of the analytical receiving stream
models where it's attempted to project those kind of differences, we've had to
take one step back in terms of going to the chemistry of it.  For instance,
going to dissolved oxygen when your real benefit that you are looking for is
fish.  I don't really care whether it's fish or mayflys or whatever, it appears
to me that we are able to do reasonable dissolved oxygen models, that we know
a fair amount about reaeration and things like this, but in the absolute case
                                     625

-------
we've got red book values that tell us what the criteria are for fish toxicity
in terms of low dissolved oxygen.  When you actually get out to a particular
case, in this particular river, I can demonstrate that there aren't fish
there that used to be there before man was there.  People do like to fish so
there is a benefit.  I can demonstrate that there is periodic dissolved
oxygen contraventions so that's a possible cause, but I defy you ever to find
somebody that will definitely say that dissolved oxygen, or for that matter,
anything else, is the cause.  You will never find a biologist that wJll say,
"This is the reason that we don't have the fish here."

      Do biologists want to come up with a definitive statement?  I think you'd
say the eastern United States are experienced in trying to get data.  People
are just used to no fish.

      If you look at most of the developed urban areas in our lifetime, it has
always been that way.  So, in these areas the Illinois Historical Society is
doing a very worthwhile thing.  You have to go back to the turn of the century
to when the receiving waters were still in fairly good shape and they have
got some data on the fish populations at that point that can show the long
term degradation, the loss of fish, and then the recovery.  But, if you are
in a developed area and you talk about, "Gee, we're going to make this swimm-
able and fishable," most people will laugh at you because they teach their
children at a very young age not to go near those waters and they accept that.
So, there is a whole education process developed there.  Now, in the western
United States in the newer areas, there is much more sensitivity because you
have a very good hope of anti-degradation, but in the east and a lot of the
older areas this is not the case.  We found there are no fish kills in many
of these areas because the fish are long gone.  So, I don't know what you
conclude from that.

Dr. Al Zanoni, Marquette University, Milwaukee, WI

      I'm not a biologist, just a form of request.  We at Marquette University
are starting on the project where we are looking at some of the sludges and
residue that will be generated from storm treatment devices, physical, biolo-
gical and chemical characteristics, so if anyone is aware of any device any-
where in the United States or elsewhere, that might be available for treating
stormwater, we'd be very happy to hear about that and possibly get some m
material for the study we are doing.

Mr. Mike Lopez, U.S. Geological Survey, Brandon, FL

      We are starting a program in the USGS, Tampa, and one of our concerns
is the high cost of data collection.  Practically over half the cost will go
into analysis of the samples.  Now these prediction models are planning models
using the data that we put in and we estimate the pollution loads which
we will try and verify by collecting data.  The amount of data collected would
affect the certainty with which we would be able to estimate model proficiency.

      A modeling conference sponsored by EPA was held in West Point about six
months ago to discuss various model applications, model accuracy and the
confidence people have in models.  It is very easy to make small changes in
model coefficients and find large changes in model projections.  If the

                                      626

-------
level of change in the model projection is large compared to potential  changes
we might get from a management alternative, then we are in trouble in order
to be able to justify such things.  We may need to develop a family of cri-
teria to show response due to different concentrations and duration.  Also,
in the case of a river model with let's say, fifty segments, what generally
is done is to pick the lowest of the fifty and try to meet the standard in
that one segment.  We should realize that only one segment out of fifty is in
a strain and that we are trying to protect fourty-nine segments that are doing
alright and maybe the fish are avoiding the problem.

Dr._Marty Wa n i e1i s ta, University of Central Florida, Orlando, FL

      We've had some experience on the selection of the storage areas within
a watershed and essentially it is an economic trade-off that you're dealing
with, especially if you have the treatment capacity at the end of the pipes.
The major problem that we ran into was this:, if you get into a downtown
urban area in which you have a limited right-of-way, the municipality has a
responsibility of either getting an easement or fitting something into a right-
of-way, and there is where your problem develops.  It is fitting something
into a right-of-way  because most of the streets, if you look under them
as in downtown Orlando, there is a six foot box that's filled with electri-
cians wiring and to relocate some of these would be very expensive.  The
comment is that it is difficult to find the space.

Moderator - Dr. Heaney

      We may have frequency curves depending on the average time of hourly
data, weekly data, and monthly data, but we do not have a response curve to
calculate our benefits.  We certainly don't have it for a series of events.
If you have data indicating 20 percent kill of the fish, what happens to the
remaining 80 percent?  Are they unstressed, partially stressed?  If you talk
to biologists they just laugh at you.  I don't think that method is going to
take us too far, because I think you just have to go back case by case again.
For example, the Rhode Island Lab of EPA has spent a couple million dollars
already trying to do comparisons qn when you close beaches.  They got some
data from somewhere around Egypt where there are grossly polluted beaches and
people are swimming in them, versus trying to discern for the U.S. beach.
They go out and talk to the bather and say, "Are you. a swimmer?  How do you
swim?  Did you actually stick your head under water?  How do you feel?"   It's
very expensive to collect information, but they are doing these studies.  I
think they said many are working in coastal zones now and they are going to
move to fresh water areas.  It's very expensive but there are some numbers
that are beginning to come.

Mr. Dick Bain, Brown & Caldwell, Seattle, WA

      I wanted to make a non-definitive statement about qualification of bene-
fits.  I wanted to go back to the days before anti-degradation, which was
10 or 12 years ago, and mention that at least in the Federal agencies at that
time there was a challenge that many of us had to develop dollar-benefits
for either benefits or declarements to Federal type projects.  These were
                                     627

-------
co-engineering declarement-type projects but we were dealing with water
quality effects and we had to work closely with agencies like the Bureau
of Outdoor Recreation and the Sports Fisheries and so. forth.  I was working
on the west coast with what was then the Federal Water Quality Administration,
and we did develop curves which would go to Cincinatti and pay homage to the
water quality biologist there.  We tried through knowledge of what the fishery
resource was, what the sport fishing and the commercial fishing experiences
were, and the like to come to some sort of systematic approach for arriving
at a dollar benefit or detriment to a particular project.  We diid some lakes,
we did some rivers, and we did some estuaries in the Bay area of San Francis-
co.  There was an attempt working with things that existed like Senate Docu-
ments and other things like that and there is some relationship between
quality and the resource.  In this way we tried to come up with dollar figures
and they were used.  I wish there were some better way to get these "would
you believe curves" so that people would believe them.

Moderator - Dr. Heaney

      I think your comments are certainly borne out by our literature search.
There is a gap.  There is some very good work that was done in the 50's and
60's and you will see that in literature and then it kind of dies out in the
last ten years.  This is because it is no longer part of the game.

Ms. Madeline Snow, Mass. Dept. of Env. Quality Engr., Brighton, MA

      We have two NURP projects which are going on, and our needs are for very
practical information.  Short of sponsoring 10 or 12 graduate students to do
detailed biological studies in our streams, I need information about how to
get a handle on the impact of stormwater on the biological community.  We
are interested in very simple information,  if it is at all possible to get.
Second, I would like to emphasize to any of the researchers here today, do
not underestimate the social, political  constraints out there in the real
world.  The models may be beautiful, the curves may fit, and the projects
will still not fly.  So, just be aware of those constraints because they are
very real and they are the deciding factor.

Mr. Mike Terstriep, Illinois Water Survey, Urbana, IL

      I have some mixed feelings and I'll try to put some of them together.
I've seen some of these earlier meetings where we were all hydrologists and
we were talking about the urban problems and so forth.  Then we had chemists
come aboard and now we have biologists aboard.  I think we finally may have
the right group of people to talk about these problems, and we certainly
have to learn to communicate.  I was impressed with Dr. Lee's work and some
of his comments concern me.  I think he said something to the effect "I'm
wasting my time to put together a table of chemical values and doing more
quality work."  I can certainly see after this meeting that this is true to
a certain extent.  However, I have been working on storm sewers so there is not
much of anything to count down there, so I'm going to stick with water chemistry.

      I also think that Dr. Lee has to soften his view a little bit and look
at chemical constituents when he is doing his bioassay and his hazard assess-
ments, and he has to learn to relate to those chemical constituents.  Arbi-

                                      628

-------
trary standards are here to stay, I think they are a reality.   There are too
many regulatory agencies and we have to have something that they can get
their hands on.  Something so they can go to the lab and get quick answers.
We've seen laboratory instruments expanded and become fantastic tools that
will give us these larger array of constituents.  I think that biologists
must also relate to these things that are in the real world.  I would like
to give you one little story from a lady in Massachusetts on the public and
the fact that they are becoming more sophisticated.  The other day, we have  a
NURP project going on also, I was installing a rain gauge.  School was out
on that day and there was a nine year old boy who came over to me and asked
what I was doing.  I showed him the rain gauge and said we were putting
these rain gauges out.  He said, "Well that only tells you that it rains right
here in this one spot.  What about on the other side of the street over there?"
I thought, "My God!  I know professional hydrologists who haven't come to that
conclusion yet."  So, we have some pretty tough customers coming up in the
future.

Dr. Jim Falco, U.S. EPA, Athens, GA

      I work primarily in rural areas rather than urban, but I thought I would
share some of the experiences we have had, particularly with relationship to
these frequency-duration calculations and the statistics that might be involved
in more realistic standards or more quality goals.  In relation to the evalua-
tion of pesticides which is our major concern in the rural area, we do, in
fact, use frequency duration curves in the Office of Pesticide Programs that
estimate the impact of used pesticides in receiving water bodies.  There are
some limitations that we have found.

      Basically, if you work in the areas where there is not a research project
where there is only USGS or core data, the highest frequency that we can get
is essentially daily.  We can't go to four hours or eight hours and I suspect
that in urban areas you run into the same problem.  In relation to another
comment that was made in referring to the statistical approach mentioned, the
statistics that we used started with DiToro's work and added in the duration
parameter.  Our limited experience in checking the density of the rain
gauges and so forth to adequately characterize an area is that you can do
fairly extensive areas even when you are concerned with the summer-intense
small thunderstorms.  We did some checking between 10,000 square miles and
essentially one rain gauge every 1,000 square miles and found that the pollu-
tographs that we generated showed very little difference, which is encourag-
ing.  However, I would temper that with saying that, that is one basin in the
southeast, so experiences in other countries may be different.  We found that
the major determinant in variation in area dealt more with the distribution
of land uses than variations in rainfall pattern.

      The last comment I have regards sediment scour. That of course is a
major problem in the areas we work in.  There are a number of models which
are either just completed or will be completed in the next six months and
it might be beneficial for people in the urban area who are concerned about
scouring and movement of material after deposition in previous time to search
that literature for information.
                                     629

-------
Mr. Skip Ellis. CH2M Hill, Reston, VA

     In my presentation yesterday, I mentioned economic optimization.  I
have one sort of problem in that the methodology we use comes out with a
global optimum.  This is the cheapest way, overall, to solve our problem.
But it is not necessarily the cheapest way for a municipality to solve their
problem because we may be using some control measures that are not funded
by EPA - street sweeping, management controls, things of that nature - and,
if we are going to make a push towards using economic optimum solutions, I
think some thought has to be given, or maybe it has been given at EPA, to
funding items that have not been funded in the past.  Otherwise, if I was  a
municipality director I would say, "That's fine but you're going to bankrupt
me, so let's put in the storage treatment because I can get a 75:25 split  and
I can afford to do something."

     We have a report coming out that addresses that question specifically.
The dilemma for the Federal Government, as I see it, is that they recognize
that it is a problem but they have no mechanism for funding 0 & M.  They can-
not make a long term commitment for 0 & M.  They can make commitments for
construction grants.  So, I think there is a recognition of the problem, but
that may be a topic for other discussion.

Mr. Robert Molzahn, Harza Engineering Company, Chicago, IL

     A comment was made about 30 billion dollars of the Needs Survey earmarked
for coliform control and in a lot of the big cities we have a lot of storage
in people's basements for combined sewer overflow and if it is a public
health problem on beaches, is it then also a public health problem in the
basements?  What is EPA's posi.tion on this?  I can respond to that one be-
cause I have cleaned out my basement in Skokie, Illinois five times a year
for four years.  Government accounting offices just looked at the public
health benefit.  In fact, I may be involved in some of the studies.  They
went around and surveyed people.  "Did anyone get sick after you cleaned out
the basement of all of the combined sewage in it?"  They couldn't find
anyone that really got sick from that.  They thought maybe someone got a
skin infection or something.  Chicago is going through a whole reassessment
trying to quantify benefits.  But, this is one that they are looking into.  •
It's kind of a unique problem.

Mr. Dennis Athayde, U.S. EPA, Washington, D.C.

     From my perspective on controlling urban runoff and combined sewer over-
flows, wastewater treatment plants has been the easy way out for the initial
cleanup.  We've done the easy part and the hard part is remaining.  Control
of combined sewer, overflow is probably where the next emphasis is going to be
on.  We talked to the construction grants people.  They talk about having  26
billion dollars earmarked for combined sewer overflow control.  Is it a
federal problem or is it a local problem?  We are taking the position that it
1s a federal problem and we are putting it in the construction grants program.
The money is going to be available for combined sewer overflow control so
there i.s no real question.  But, that's the answer to your question.  The
money is going to be there.
                                    630

-------
      We made the decision that we are going to put money into the combined
sewer overflows.  The next decision that we have to make is, are we going to
put money into separate storm sewer discharge cleanup?  The act was changed
and the funding for the treatment of storm sewer discharges was not allowed
any more for a five year period.  We need information on whether or not
separate storm sewer discharges are a problem.  What we need is a good defi-
nition of what our problem is, and Madeline tried to point that out.  We have
to be able to describe what our problem is.  We have to have something to
hang our hats on.

Mr. Steve Tindale, City of Tampa - DPW, Tampa, FL

      We've made a decision that the analysis should be case and site specific.
We've run around in circles in terms of how to analyze problems and we've
said we have to go to the sites and look at them.  Maybe, when we talk about
beneficial use we need to talk about site specifics.  Maybe we need to get
away from thinking about some of these general indicators.  We need to go to
the local public and a'sk them what the beneficial use is.  Maybe the federal
role is looking at major streams, national impact, major recreational areas,
or something of major impact on major communities, not small communities.
Maybe what we're talking about is trying to come up with a process that says
"Let the people decide the options, all the benefits, and the cost and come
up with the beneficial use."

      When we put the burden on the local government, maybe what we can
talk about from the beneficial use is some kind of procedure to be developed
by the EPA, that states, etc., so the local people can take a local problem
and describe beneficial uses and the cost that they are willing to pay to
obtain those beneficial uses.                                 "!
                                                              -i
Dr. Fred Lee, Colorado State University, Fort Collins, CO

      I feel very strongly on that approach and in the City of Fort Collins.
where we're working now, we are developing procedures to specifically
answer that kind of question, how you take chemistry and biology information
and put that together to assess what the public will get from the expenditure
of so much money.  You lay it out to the public and say, "This is what you'll
get."

.  . . Coffee Break followed by continuation of the workshop ...
                                  SESSION II

              Moderators:  John L. Manicini, Manhattan College
                           Richard Field, EPA

      Speakers: 1) Richard D. Tomlinson, Municipality of Metro Seattle
                2) Raul S. McQuivey, The Sutron Corporation
                3) John L. Mancini, Manhattan College
                4) Martin P. Wanielista, University of Central Florida
                5) Peter E. Moffa, Stearns and Wheler
                                     631

-------
Mr. Field

      Good morning - I think the earlier workshop session went very well.  I
got a lot out of it.  There were a lot of interesting remarks made.  I'm sure
these comments, many times, were just general and vague but they can start to
make you think.  Now, if anyone has any comments we are going to have more
time after the second set of speakers.  We are going to make an attempt in the
formal proceedings to have this workshop included, but I say it is an attempt
because these things could be a problem.  I don't know how the tape recorder
is working and things like that.

      We were talking about having Phil Shelley make some comments and I no-
ticed before the coffee break we missed him because all of a sudden the
discussion picked up again.  I would like to give Phil that opportunity now.
I'm sure he'll have something important to say.

Dr. Philip Shelley, EG&G Washington Analytical Services, Inc., Washington D.C.

      Some of us have expectations of R & D  that perhaps aren't really appro-
priate.  To me, it is not the function of the research dollar to solve
immediate problems.  To solve today's problems today.  Rather, the research
dollar should be used to anticipate tomorrows problems and to try to come up
with the combined bag of tricks that we need to effectively deal with those
problems.  Now we have concerns for certain things and emphasis shift to
another set of concerns.  The pendulum swings back and forth and I think it
will always be that way.  The R & D community has responsibility to maintain
the technology base so that the proper perspective can be applied as best we
can at all times.  What are our real R & D needs as I perceive them - problem
identification, we've talked a lot about that today but so much work needs to
be done there and although problems are locally specific - the problem exists
in the eyes of the beholder.

      There are some things that are a little more universal that we can
address so that people in the public participation process can be given the
opportunity of choice.  The R & D program owes the public alternatives and
explanations of these alternatives in clear enough terms so that they can be
understood and the people can, in fact, choose.  We talk about the technolo-
gies that are involved, we talk more about the economics that are involved
because the notion of affordability is really beginning to come home.  We do
not have an unlimited supply of bucks to pour into any given public concern
arena, environmental or otherwise.

      If we look at the AWT program and we see some of the localities that now
have a beautiful plant and say, "Oh my God, I don't know how to run it - I
can't afford to run it, I'm running it ineffectively and I don't know what
to do - you know, maybe we need to come up with better designs that can be
run by the mayor's son-in-law."  Like Madeline indicated, there are institu-
tional problems that we have to take into account.  It is inappropriate to
expect EPA to fund research efforts that don't in some fashion meaningfully
relate to its essential mission as a regulatory agency.  We talked to some
extent about the need for standards, criteria and research that will support
regulations and legislation.
                                     632

-------
      When I look at this horrible thing called water quality, it's more than
we can say grace over.  Also, we sometimes tend to over apply some of our
particular pet little route skills that we know a whole lot about, and we
lose sight of the bigger picture - we overly complicate because we know what
to do.  I look at modeling of stormwater as it started 10 years ago, domi-
nated by the plumbers, because if there was one thing we know how to do it
was to route water around.  We understood the hydraulics of this situation
so the models were very good in hydraulics.  They were lousy in water chemis-
try because we didn't know much about it and besides the people working on
the models weren't by and large chemists.  When we combine the physical-
chemical characteristics, overlay these with the ecological considerations,
that the little critters we are concerned about, we have a very complex system
indeed.  We do, however, have a responsibility to look at new tools as they
come along and to continue to step back and to try to maintain, regain or
establish fresh insights.  DiToro's talk yesterday was an example, one case
in point, Miguel Medina mentioned common filtering, another very strong tool
that we have got to apply in water quality work.

      Time series analysis and life cycle cost analysis are other areas where
we have a great application.  We focus on the capital cost portion and some-
times you give lip service to 0 & M and you say "OK that will be about 10% of
my invested capital" and you wing it at them.  And I submit, it is very poor,
if you are not examining the manpower requirements, the training require-
ments of those people, you may find that the nation's labor pool is unable to
support some of the solutions that the R & D comes up with.

      Finally, I think that we can develop uniform methodology despite the
local specific nature of the problems.  We can use the same approaches.  It
doesn't mean that the answers, as I grind through this uniform methodology,
would be the same for Kansas City as they would be for Washington D.C. or
Los Angeles, but the approach can be the same.

      Lastly, when we talk about benefits - we've been talking a lot about
benefits because our awareness is being opened more and more in the area of
Proposition 13 - about who benefits.  Because we have to determine not only
who benefits but who pays, and that is an area not for the technician but
for the politician.  But, insofar as we can establish no way that mankind
benefits from the presence of these particular critters in this given creek,
it also doesn't mean that we know everything.  We need to continually main-
tain a bit of a humble posture because we may learn something next year
through a well-applied R & D dollar that will point out to us why critter X
did happen to be important and he has a way sometimes, if he vanishes, of
making us pay in ways we can't even foresee now.

1) Mr. Tom!inson

      I'm not a modeler but I do have some veryldeep interest in the dis-
cussions that have been transpired this morning.  We do have the situation
where we have existing very visible fisheries, noteably our salmon fisheries,
and the people in the Northwest, specifically Seattle, are very much plugged
into the mystique of this whole thing.  It is their way of life and it's
very, very important to them.  Because of this, aquatic scientists in the
Northwest have even stronger urging by their public to meet these questions

                                      633

-------

in a practical manner and attempt to relate to the fisheries.  I myself,
personally feel kind of sorry for the bugs, the crawlers in the sediment,
too; but I have to face the fact, that people in general do not relate well
to bugs in sediments.  What they do relate well to are things that they
encounter in their fishing experiences and their other water recreational
activities.  So that hopefully, if we can approach our responsibilities by
attempting to preserve fisheries, the bugs and the sediment will go along for
the ride.  Now, with this sort of thing in mind - I guess one thing more that
I should mention - is that one of the things that we have done in the Seattle
area, this is specifically in metropolitan Seattle, or METRO the organization
that I represent, is the setup that we call the salmon enhancement program.
We have taken three streams that are partially degraded at this point but are
still capable of sustaining salmon fry eggs going through development to
fry, and then hopefully, will support salmon fisheries with the idea that
we plant salmon there now and allow these fry to go to the ocean to return
in approximately four years time.  We have four years in which to clean up
those streams and make them a viable habitat for those returning salmon.
Now chances are, that if we don't do a good job, the returning salmon may be
unsuccessful in their spawning and we will see a very gross failure of our
efforts.  Either way, it is a kind of a no-lose situation.  People in that
area are very critically interested in this sort of thing.  If they clean up
the streams and they have a success to look at, they are going to be very
proud of that.  If the streams are not cleaned up - or' for one reason or
another not able to produce a success - that is going to be very visible.
That's something that you can point out to the public by saying, "Look, those
streams are not fit for those fish that you put so much credence in.  It won't
support them."

      So, here is the emphasis - for us to do the job as far as showing bene-
fits through our research work.  I will say, now, the information I am about
to present here, in my way of thinking it does not represent a state of the
art effort.  What I am going to talk about is a brief assessment of water
quality impacts on the ecology in the fresh water environment and this plugs
into the talk that I gave on Monday.  The reason that it does not represent
state of the art is because I consider that right now for this type of
assessment, state of the art probably gets into multi-varied analysis and that
sort of thing, and this is a much more simplified approach; but beyond that,
thinking of Dr. Lee's comments, this does not even begin to really think in
practical terms about what it means to fisheries.  And the reason that we
didn't get into that, quite frankly, is that we didn't know how to do it.
Dr. Lee said that he has some techniques that are becoming very successful
in this direction and I'm certainly looking forward to talking to him about
those in some detail.

      Our general approach in a very simplified manner of presentation was to
try to determine the relationships between numbers of organisms with distance
from the outfall using regression analysis.  Assumptions being made, of
course, since you don't have the outfall at a control site and you don't
have a distance from outfall effect.  You consider that zero, so anything
significant in terms of distance from outfall effects shows an impact from
that outfall.  Now we essentially summarized our data in three ways and I'm
going to show a couple of brief slides here, and I'll discuss them briefly.
                                      634

-------
COUNTS  WEIGHT (mgl
          NON-SIGNIFICANT CORRELATIONS

           COUNTS: NONE
           WEIGHT: Ch
                  (Co. N. P NOT TESTED)
                                               COUNTS  WEIGHT Img)
                                                           NON-SIGNIFICANT CORRELATIONS

                                                             COUNTS: 0
                                                             WEIGHT: Ch. O. T
                                                                   [Co. N. P NOT TESTED)
         ~—10	_.20__   30    40
                 DISTANCE (ml	.


          NON-SIGNIFICANT CORRELATIONS
           COUNTS: Co. N. P. T
   -10-L-1.0  WEIGHT: Ch. O. T
                  (Co. N. P NOT TESTEDI
                                                         NON-SIGNIFICANT CORRELATIONS
                                                           COUNTS: T
                                                           WEIGHT: T
                                                     L.3J)         (Co, N, P NOT TESTED)
    Figure 1. Net change in numbers of organisms and  biomass per core, as a function of
              distance from the combined sewer and storm drain outfalls in February and
              September, 1978.  All relationships determined by regression analysis and
              significant (at the 95% confidence level) unless otherwise noted.
              Ch: chironomids,  O: oligochaetes, Co: copepads, N:  nematodes,
              P: pelecypods,  (Pisidium spp),  T:  total for all organisms.
                                             635

-------
                   DISTANCE (m)
     NON-SIGNIFICANT CORRELATIONS
••2-°    COUNTS: Ch
       WEIGHT: O (Co, N. P NOT TESTED)
                                                            20   DISTANCE (m)
           NON-SIGNIFICANT CORRELATIONS
              COUNTS: O
              WEIGHT: Ch, O (Co, N, F NOT TESTED)
                                               WEIGHT (mo)
     NON-SIGNIFICANT CORRELATIONS
       COUNTS: NONE
       WEIGHT: O (Co, N. P NOT TESTED)
          NON-SIGNIFICANT CORRELATIONS
            COUNTS: O, P
-10-1--1.0     WEIGHT: O, T (Co, N, P NOT TESTED)
                         Figure 1. (Continued)
                                   636

-------
We also calculate coefficient of determination, which is the square of the
regression correlation coefficient.  If I could have the first slide, please.
OK, what I'm showing here are on the top from left to right is represented
a single outfall, in the bottom left to right a second outfall.   They are
both combined sewers.  On the left hand side of the figure we have one season,
February, which is typically part of the rainy season where we have a lot of
overflows and on the. right side we have the drier season.  I'm showing here
relationships of the changed number of organisms and the change in biomass
per core in relative units, with distance from the outfall.

     There was an enhancement as you move toward the outfall of oligochaetes.
and this is for a polluted type condition and it has been mentioned several
times during this conference - that's a fair indicator.   So, move on to the
next slide please.  For the storm drains we ran into some extra problems and
basically these consisted of two things:  it is difficult to get a proper
super imposition or sampling grids, and I think we missed a significant part
of the effluent plume.  Secondly, we saw  in these conditions  significant
scouring right by the" outfall which depleted the number of organisms immedi-
ately next to the outfall.  This had a tendency to strengthen the positive
correlations and weaken the negative correlations, so we had plots like this
and we were not able to use a straightforward interpretation.  Indeed, in-
spection of the data showed that this occurred for oligochaetes only because
of the scouring next to the outfall.  In other words, what I'm trying to
say is, from the storm drains we saw enhancement of oligochaetes from the
outfall.

     To encourage discussion a bit further I will mention that we did make a
general attempt to get a ball park figure of what the depletion or enhance-
ment of these organisms might mean in terms of fish tissue.  Now, this invol-
ved a number of assumptions which are very gross, but it was a matter of
essentially trying to get an idea of what the order of magnitude might be.
Taking the most extreme condition, where we had found that enhancement of the
oligochaetes, which in this case were responsible for about or better than
90% of the number of organisms in the sediment of that outfall.   Essentially,
what we did was, first determine the area of impact by relating what we saw
at this outfall to what we saw at control sites at similar depth and then
using that area as a basis we integrated the numbers of organisms that we
saw increased within this area which we were able to calculate using a couple
of other assumptions.  One was a general figure of dry weight to fresh weight
conversion for oligochaetes that we pulled out of another paper and the other
was to use a 10% conversion factor for the amount of fish tissue that you
could expect to be produced by 100% of your given value.  We came up with
some interesting figures that might give a ball park figure as to what these
impacts might be on fish.  This depends on the assumption that all of these
organisms are available for consumption and the fish are successful in finding
all of them.  That would mean, in terms of fish tissue,  the biomass that we
calculated for the oligochaetes total enrichment was something like 47.3
kilograms around this single outfall and this outfall represents a substantial
CSO, one of our more frequently overflowing CSO's in the Seattle area, and
that translates to about 4.7 kilograms fresh weight of fish.  If you assume
one generation of organisms per year, this would say that perhaps you are
                                     637

-------
making ten additional pounds of fish available, or you are increasing the
total fish biomass in the lake by 10 pounds for the years time.  It's easy
to shoot this full of holes, but it gives a potential, it's sort of a maximum.
We are thinking only in terms of food for fish and not considering very
important additional factors such as reproductive impacts, which we did not
even attempt to assess, and I'suspect they are probably more important than
what we are talking about here.  But to me, 10 pounds of fish was sort of a
surprising find, in that it wasn't nearly as much as I expected it might be
from a given outfall.  If someone were talking to me about the enhancement
value of an -outfall as opposed to the toxic effects that we might have seen,
this was one of the most extreme effects.  It was not really as high as I
had expected.  Maybe I had my sights set a little bit too low.

     I would like to leave off there and would welcome someone shooting this
full of holes, simply because it would stimulate some discussion in this
realm.  We very desperately need to find a good approach to this kind of
evaluation.  If there is someone using, say multi-varied analysis, to try to
coordinate their observations of biotic and abiotic factors at community
distributions versus, say, presence or absence of taxicants of various types,
I would like to know if there has been some success with this, because we're
very heavily into this sort of thing.  We are getting very strongly into the
measurement of impact, the presence or absence of organic toxicants and we
are going to be hopefully extending the types of studies that I'm talking
about here to that realm.  We desperately need this sort of information.

2) Mr. McQuivey

     I'm going to briefly discuss some of the results that we have obtained
from a sediment modeling on the Cuyahoga River which was one of the areas
that we found a potential problem.  Many of the problems were sediment re-
lated and, as such, we really wanted to define the fate and the effect of
sediments using mathematical models.  Our study was conducted in conjunction
with Colorado State University and the objective was really to determine the
feasibility of determining the fate of CSO materials using models.  We select-
ed the Cuyahoga River at Old Portage down to Lake Erie primarily because there
was a DO problem that we had recognized in an earlier study, plus the U.S.
Geological Survey had collected many years of sediment data at two locations.
It was one of the areas that we felt had enough good historical data that we
could make a fairly good attempt at using the model.

     The principles of the sediment modeling was a two step process.  We
used an unsteady flow model that Dr. Keefer had developed while he was with
the U.S. Geological Survey and we coupled that with a sediment transport
model developed at Colorado State University.  It was a watershed model and
we modified and coupled that with the flow model which was one dimensional.
The advantage of this type of model is that it allows you to look at what
happens in space and time and there is a great deal of things that we could
talk about in terms of the flow model and sediment model and the data needs.
                                     638

-------
Our reach is just below Akron, Ohio down to Lake Erie, about 40 miles.  This
gives you an idea of some of the things that;,we>found in terms of the DO
study as well as some of the observed stages and discharges as well as the
modeled results there.  The backwater profile at a given flow condition is
a useful technique that we developed to calibrate our flow model and the
sediment model.  We're interested in the deposition and scour at different
locations along the river.  This model was set up to take particles in the
sand, sizes  .06 millimeters on up, and the material was either scoured or de-
posited at different places along the river in the longitudinal direction.
In a stormwater runoff bypass below the Akron sewage treatment plant there
are two little sediment traps in the reach of the river.  Little impoundments
trap much of the material as you route it downstream, and you can change the
bed elevation, Delta Z.  It is depositing and then when the flow comes through
down there, within minutes the materal is scoured.

     We also did some other things,where we were concerned with heavy metals
dealing with the fact that there is a specific gravity variation in materials
that are deposited.  What happens to heavy metals as they are input into
receiving waters?  Where do they deposit and how are they scoured?  I .think we
used a five year storm event.  .Most..of the sand sizes were scoured and deposi-
ted down in Lake Erie.  The Corps of Engineers dredges there every year to
keep the channel open.  Most of the heavy metals settled out in these two
sediment traps.  Even with the five year event they did not go through the
sy s tern.

     As you probably know there is not a lot of sediment data collected.  I
think the U.S. Geological Survey has about 1,000 sediment stations.  You can
get daily data from about half of those.  Some of them are weekly, monthly,
semi-annual grab type samples.  Sediment data is expensive to collect and
therefore in terms of using mathematical .models in fate and effect of sedi-
ments, it is probably not likely that the location is going to be exactly
where you are going to have good data.  There are only a few of those places
around the country.  Data is probably one of the most critical issues in
modeling sediment transport.  Perhaps we need to collect a good set"of data
below a CSO and verify the model a little better than we did in this .parti-
cular case.  There were some shortcomings.  We have no data from the storm-
water bypass below the Akron sewage treatment plant.


      I have great faith in the model.  We have been applying this for the
 Corps of Engineers and various other people, in terms of determining where
 to dredge.  It is something you are going to have to do annually, and we've
 been, more or less applying it to situations like that, except in terms of
 CSO materials.  I think  that without data  we are only going to get a quali-
 tative feeling for what is happening.  With good data  I think we can put
 some quantitative numbers on the results.  In terms of the research require-
 ments  we need to better verify the sediment models in terms of CSO materials.
 Even though we feel comfortable because they have been applied for years in
 different circumstances, we need to collect a good set of data and verify
 these models.  Most sediment models have been developed primarily for the
 sand sizes, and in CSO's we have materials that aren't necessarily spherical
 in size.  There is a broad range of materials that most of the present sedi-
                                      639

-------
ment models do not take into account, so one of the things we need to do
is to extend these models to the fines, the cohesives as well as to other
particulate shapes, fall velocities  and what not.

3) Mr. Mancini
                              • ,9
      I  would  like  to  divide my remarks  into  two  portions.   The  first  part  of
my comments will address what  I  see  as  general research needs,  and  the  second
part will  concentrate on more  detail  in -the  mathematical modeling area.  The
perspective that I make these-comments  from  is as follows.   I see that  the
urban runoff  and CSO  program,  as proposed, will  cost the taxpayers  of this
country from  40 to 80 billion  dollars in  capital  cost.   I would gather  that
over the lifetime  of  these projects  that  cost might  be  tripled  in terms  of
both capital  and operating costs.  These  expenditures are being looked  at  in
an environment and a  period where  heady concern  for  water quality has abated.
The  environmental  concerns are beginning  to  face the realities  of competition
in an open society with health,  defense,  education,  energy  and  the  economy,
and  it  seems  to me that our profession  is facing a crisis where the survival
of the  remnants of the environment movement  are  really  at stake in  the  long
run. I think, in  addition to  these  factors, there is an assidious  force that
may  be  driving our profession.   Toxics  has become the word  of the decade.
Let  me  relate to you  the experience  I had as a professional.

      Ten years ago the word of the  decade was thermal  pollution, and in many
places, thermal pollution was  a very real  environmental  threat.  Hundreds
of millions of dollars were spent  on studies of  thermal  pollution.   Billions
of dollars were spent on the abatement  of thermal  pollution.  Thermal pollu-
tion has not  been  mentioned in an  environmental  conference  that I have  attend-
ed in five years because it has  turned  out to be a very site specific problem:
a real  concern at  special locations, but  a non-problem  across the nation.  I
would submit  to you that toxics  may  inherently have  the same characteristics.

      At the present time toxics  are  the concern  of the  decade,  and  they  are
clearly of vital concern with  certain types  of materials and in certain
environmental settings.  Clearly,a dump of keptone in the James River that
contaminates  the river, the fisheries and has health implications is  a  signi-
ficant  environmental  concern with  respect to toxics.  The same  thing  is  true
with RGB's in the  Hudson and'Escambia Bay,,   But, do toxics  really  pervade
all  of  our environmental concerns?  I submit to  you  that the jury is  still
out, but there is  a real possibility that toxics may be overblown as  to  im-
portance in our kind  of thinking and our  research people ought  to be  very
aware of this in terms of their guidance  of  our  research efforts.   Therefore,
I suggest that our research efforts  should maintain  a balance.   They  should
clearly continue to address the toxics  problem which has a  lot  of money  being
poured  in from other  areas in  addition  to the municipal  branch.  The  municipal1
branch  ought  to kick  in their  share  of  the pie,  but  I think  it  would  be  a
grave mistake for  the research and development efforts  in the municipal  area,
and  in  particular  in  the CSO and runoff area, to concentrate too large a per-
centage of their efforts in the toxic area to the exclusion  and ignorance  of
the  conventional pollutants.

      Getting  back  to  the first point that I  made that the climate in  which we
                                     640

-------
are operating as environmentalists is changing.   There are two very signifi-
cant examples that highlight this.  In the first instance, EPA is beginning
to require and request information in the CSO program, which is a funded con-
struction grants program by Congress.for a benefit analysis of CSO before they
spend capital.  So, the times are changing.  We are no longer in the era
where we are willing to afford expenditures for environmental control without
regard to benefits.  The agency, being a representative of the administrative
branch of government, is, in fact, reflecting the public attitude and be-
ginning to require a benefits analysis with CSO.  Even a closer reflection
of this public attitude is Congress1 attitude  as indicated by Dennis Athayde.
They essentially said, "Hey, is urban runoff a national problem; and,will you
define for me, Congress, the problem in terms of impairment of beneficial
uses?"  So, the handwriting is on the wall.  We are now competing with all
sorts of public programs for money.  In addition, CSO and urban runoff  are
problems of national significance, or if they are problems of local signifi-
cance, there are some very specific questions that we are being forced to
answer now.  The pressure to give more definitive answers to these questions
will  increase with  time.  As an example of the  questions:  Does this type of
waste, CSO, urban runoff and so forth have to be removed?  jf SOs a-f- what
locations?  Every single pipe, every other pipe, one  in ten  pipes, a hundred
pipes?  How large a facility or action do  I have to take to  remove it?
Should I remove 10% of the total  annual load?   Should  I remove  all  of the
small frequent events and take a  beating once every year?  Five times a year?
Once  every ten years?  I think these are very significant questions because
they  change the order of the 20 to  60 to 80 billion dollars.  If you design
for the frequent small events you will change your order of  expenditures.

      Finally, what  type of trade-off exists?  I've got to say I'll give 10
pounds of fish for  4 million dollars.  You know those are the kind  of deci-
sions, if this profession is going  to survive,  that I  think  we are going to
have  to make.  Our  research efforts  have to be  directed towards providing
those answers for both conventional  pollutants  as well as the toxicants.

      Now, let me be a little more specific in terms of the general research
needs.  It seems to me the impairment of beneficial uses  is  a real area of
opportunity in terms of a research  program.   In particular,  it seems there
is a  need to define what is the beneficial use, and to define a beneficial
use at several levels.  So it's not just enough to say fisheries are a bene-
ficial use.  We have got to define  that at a  number of levels so that we can
begin to consider tradeoffs,,  We  can do that  in terms  of dollars,  in terms of
fish  population, or we can do it  implicitely  in terms  of  standards,  but we
have  to come  to grips with that problem if we are going to be viable as a
profession.   Finally, we have to  develop methods of assessing benefits where
there are competing resources.  As  an illustration - many urban runoff pro^
blems and CSO problems are confined to very small streams, this is particu-
larly true with the urban runoff, very small  streams  within  cities.  Surround-
ing these cities may be very luxurious alternate resources.  Beautiful lakes
with  large fisheries and so forth.   What is it  worth  to clean up the little
stream when the population has an alternate resource  which is much more de-
sirable?  These are very tough judgements  and we as technologists  have to
have  input from the  information standpoint to the public  and the elected
officials, otherwise they are going to make the decisions without  the know-
                                      641

-------
ledge and we are going to cry, one way or another.  If we solve all  of these
problems in terms of beneficial uses, we have to translate those into method-
ologies that can be applied without spending 100 million dollars on  each site
to determine what we should do on that one site.  So, the methodologies have
to be condensed and simplified so that they can be applied on a site specific
basis and answer some of these real problems with respect to effects.

     It seems to me that there is a need to look at effects for combined over-
flows and urban runoff, acute and chronic effects.  We might want to consider
looking at these biological effects in two ways.  The first is by the indivi-
dual components, where we can go back to the literature and pick off much of
the historical data base;  that is, cadmium, LC 50's, measurements of this,
that, and the other organism, and for fish and the lake.  On the other hand,
it might be worthwhile to sponsor an effort which would look at "urban runoff
and CSO's in their totality" to see what the effects of the total combination
would be.  Some remarks have been made here about cynogistic effects, antago-
nistic effects, and they may need to be second order, but are not there.  It
seems to me that some screening research money to look at effects, both
chronic and acute, from urban runoff and CSO's are needed in this area because
we  have a very unique problem in the stormwater area.  We had 88 days of rain
in  Austin, Texas.  That's almost a desert for pete's sake.  So there is a lot
of  little shots of either toxic effects or low DO's and so on.  The chronic
impacts may be significant.  On the other hand, there are some very large
events in which the acute effects may be very significant.  We as a society,
have to consider devoting some resources to looking at both of those aspects.

      If we do determine some areas where there are effects, we have to be:able
to  synthesize this fact in terms of estimating populations.  When we think
of events that are essentially a telegraph wave of various side events that
occur in time; we get a rain and it's dry, we get a rain, and it's dry.  Then
factors like recruitment and regrowth of the population are very significant
so that if we get a 10% shot each time it rains in terms of mortality or
a destruction of the population, but in the three to five days between rain
storms the population can revitalize itself back to equal zero or 100%,
then we really don't have a water quality problem.  So, there are whole areas
that have to be addressed in terms of coalescing the laboratory work with
theories that will allow us to understand recruitment issues, regrowth issues
and population issues, areas which will dovetail with our ability to look at
benefits of biological effects.

      Another significant point is the fate of the materials from the CSO's.
Let me point out something that Jim Heaney called to my attention several
years ago and I really grabbed onto this as a nice analogy,1..,  If you stop
and think about it, there are two types of effects from CSO's and urban
runoff discharges.  There are the primary, or immediate effects so that you
take an urban runoff or a CSO discharge and there is some little critter in
there, and he gets hit with a low DO or copper cadmium concentration that's
high and he responds - he dies, or lives, or gets sick - something like that,
but it is an immediate effect.  The storm has passed, the critter begins to
revive and return to the original state depending on what the effect was.
That is the'primary effect of the CSO discharge.  The primary effects are
applicable for conventional pollutants as well as for toxic materials, or non-
conventional pollutants.
                                     642

-------
      Then  there  is  a whole  series  of  secondary  effects which  are associated
with  the bottom  deposits  and  things of, that-nature.   You  hit  a  body of water
with  a  CSO or  urban runoff  and you get deposition  of  material.  Again, it's
appropriate for  both the  conventional  and,non-conventional  pollutants.   I'll
give  you two specific  examples:  dissolved oxygen  bottom  demand, very impor-
tant  in many situations,  often created by  CSO's and urban runoff,  and often
created by algae settling upstream or runoff  from  agricultural  areas.  Many
causes, but that is a  long  term  assidious  effect that could be  associated
with'the two factors that we  are dealing with.   Also, the fate  of  toxics,
particularly cadmium,  is  unknown.  It may  become part of  a  lignent system ab-
sorbed  on  the  particulate or  is  chelated by organics  and  eventually drops to
the bottom.  What happens to  it  in time?   Anaerobic  conditions  may occur,
the redox  potential may change,  and so we  may get  a resoluabilization of the
heavy metal  or the  organic  material and there could be a  long term effect of
the same load.   So, there are two  aspects  of  the CSO  and  urban  runoff, the
primary effect,  and then  this long term subtle  impact.   It seems that we
should  consider  committing  vital research  money towards defining the fate of
these contaminants-, both  toxic and conventional pollutants.  When  we start
putting in all of these wonderful  devices, we've.-got  to monitor compliance
with  them.  So,  some effort is necessary to begin  to  decide how we're going
to monitor and whether we're  making it or  not.   Do we have to increase the
amount  of  capital and  operating  costs we are  putting  in?   Do  we have an  oppor-
tunity  to  reduce it?   We  need monitoring on a wholly  different  time and  space
scale than we  do for the  continuous discharges.


     Given this list of effects  and benefits,  how successful are we in  our
treatment devices in terms of really handling  all of these tox'ics  and  conven-
tional pollutants?  Can we make  it in  terms of this benefit analysis with the
existing technology for the treatment  devices?.  All of these research efforts
require either lab studies,  field studies,  or  model calibrations.   This  pro-
fession  is in a changing climate  for environmental  control and we  are  being  re-
quired to  justify and  demonstrate the  benefits from expenditures that we  re-
commend.  We need to organize our research  program to provide  packets of  infor-
mation that fit into an'overall  mosaic that allows  us to answer  some of these
questions.   Looking backwards, CSO and urban runoff programs have  been  very
successful.  They are  one  of the  shining examples of EPA  research which  led
our profession in the  treatment  and load assessment area from zero, knowing
nothahg  about effectiveness  of treatment and a real estimate of  loads  from
CSO's  and  urban runoff, to the point where  we  can consider a whole  array  of
treatment devices and  make some  pretty subtle  comments on the  kinds of  load
variations that are occurring.  Now this program is facing a new challenge
to meet todays needs,  different  needs.  They are in terms of benefits,  effects,
and fate,  and if these  guys  can  respond to  that they will  have done the agency,
the public, and our profession a  service.

4) Dr.  Wanielista

     I would like to amplify on  John's comments and the  underlying  theme  of
the conference which essentially is receiving  water quality responses  and
expand on it to indicate receiving critter  responses  and biological responses
in addition to the water quality parameters that we measure.  Yesterday  and
                                      643

-------
the day before  we heard some talks on lakes and lake responses.   I  want to
specifically limit my comments to a project which is ongoing in this local
area.  One that Harvey Harper has responded and talked on,  Lake Eola.  Essen-
tially, we are dealing with the storm sewer system and land-locked  lake.
The lake itself is about 27 acres and the watershed is only about 131  acres.
We are in a fortunate position because it appeared that the problem was  very
well defined to start out with;   that is, from the Mayor's office right  down
through public works, citizen groups and everything.  There was a problem
with the odor out of the lake and the aesthetic condition of the  lake, not  to
mention anything about water.quality.  There was certainly a political move to
do something in order to keep the politicians in office, and they were very
responsive to the needs of the people.  Joggers complained.  There were  numer-
ous cultural events, bandshell concerts    and craft events.  Over 400 craft
booths spread around the lake in which people actually paid to put a craft
booth there and sell their wares during the show.  There are two  of these,
called Fiesta in the Park, which go on at certain parts of the year and  then
 are extended to other times of the year depending on the demand.   So, there
 was indeed an economic benefit as the city saw it.   There were.ialso children's
 benefit programs, fishathons and what not, that were being conducted  and
 essentially the city was very interested in;protecting this and  doing some-
 thing further.

      In 1972 we did a brief benefit cost analysis which had nothing to  do "
 with the craft shows and everything else which is going on, but  had to  do
 with the land values around the lake and what would happen to the land  values.
 Or, how can you compare this urban setting to another urban setting where you
 had a difference in the water quality for similar activities.  I  think  you'll
 recognize the value of land in the downtown area and essentially you  could
 add up the dollars and find out that it would be significant.  It was really
 an unscientific study which was based on calls to realtors to find  out  what
 their opinion was.  We came up with a figure which  was over one mi11 ion dollars
 for the land value, if indeed,  the lake went to an  odorous condition.

      There are a  lot of other ways  of doing  benefit  cost studies.  My own ex-
 perience  in recreational,  values   say that  it's  very difficult, and you  have
 to come  up with some very  well  designed  programs.  For example, on the Lake
 Eola thing, there are somewhere in  the neighborhood  of 2,000  people that go
 there over 30  times  a year to visit a concert.   There is a  concert which is
 held fortwo hours.  It is supported  by the  local  newspaper  and the city of
 Orlando.   There is  a dollar value  attached  to  this.  People actually expend
 energy,  time,  money  and everything  else  to  get down  there  and listen to  school
 bands, Navy band,  University  band,  art performances  and what  not.  Some  things
 are quantifiable  and I'm saying this  because  I think there was justification
 for going  into and  looking at a restoration  of this  particular lake.

     The program we designed had a lot of detail  in it and we used models.
 One of the earliest models we used was a time series model.  We  tried to re-
 late the responses of phosphate in the lake to discharges of stormwater.  We
 actually came up with some numbers and some relationships and I  would say
 that it was called guessing numbers and fitting the data at that time.   But,
 at least it gave us some kind of indication that we might have a problem here
 and maybe we could amplify and get some more data.   Our own studies then

                                      644

-------
went into the bioassays which were presented to you.  We also did the eutro-
phication modeling, the Vollenweider modelsiDillion models, Larson, Shannon
Brezonik Index and all of these kinds of things which indicated that the lake
was eutrophic to a certain degree.  This gave us an indication that possibly
we should remove some phosphorus from the lake.  The bioassay studies gave
us an additional indication that indeed, the biomass responses seemed to be.
stimulated by the additions of stormwater.  We noted some toxic effect and
the visual,observation of the situation gave additional  information after a
storm event.  If  you went out and just visually looked at the system, you
would probably find out that there was some response - that the algal blooms
tended to be associated with the rainy season and they tended to be more
frequent during that time period.

     One thing we did that was interesting was to go in and examine the run-
off.  Over 50% of the nitrogen and phosphorus was dissolved in the actual
runoff collected going into the lake.:  This would limit your technology for
best management practice.  If you're talking about technology, which essen-
tially is only dealing with sedimentation or physical removal, you may limit
your choice of the particular management practice; also, the particular com-
pound or form that you're looking at.  Does this particular form actually
produce a response in the algae blooms that we were getting on the lake?  So
comparing the existing models, the studies that we did, and everything else
led us to a management scheme where we are diverting the first quarter to a
half inch of runoff.  Probably from about 0.9 to 1 inch of rainfall in that
particular area.  This is based on a mass removal that we have done over a
long period of time.  We actually went out and measured these diversion sys-
tems, found out what the diversion volumes were, the quality jn the diversion
votumes, and then went to a stochastic simulation using stochastic rainfall
traces over a 20 year period and tried to estimate what the mass diversion
would be and then come up with frequency distributions on the mass discharging
into receiving bodies of water..  This all resulted in an implementation plan
which we're currently undergoing.

     This project wasn't too hard because we had a lot of help all the way
from U.S. EPA Storm and Combined Sewer Program, 314 Clean Lake Program,
Region 4, Legislature, Tallahassee DER, down into the local level and every-
body got together and recognized the problem.  The funding level for the
actual implementation actually came from the State Legislature who was
friendly toward the project.  They came up with a big chunk of money, $190,000
from the legislative pot.  The city and the university also provided a small
matching amount and took that money;and went to a 314 Program.  To try to
get in you must get everything together and try to do something about the
runoff into the lake system.  You should look at the receiving water body and
possibly make some adjustments within the receiving water body itself.  Lake
Eola didn't have very much rooted vegetation.  It didn't look like there was
an "equal system" set up to assimilate and recycle the material that will
have to be changed in some ways to at least provide more of a natural system
for similation.  This is also acceptable to the people of the area.  They are
the driving force.  They would like to see all of the natural area returned
to the downtown area of the city.  I believe benefit studies can be done.
They're expensive and they can be done in a lot of detail or in a little less
detail.  I believe that the driving force on many of our projects is not so
                                     645

-------
much what the water quality is but what the people want, and that has been
brought up quite a few times.  As engineers and scientists we should be
responsive to that need.

5) Mr. Moffa

      I want to touch on one element that was brought up twice and that is
public participation.  We're talking benefit.  I don't think there is any
question that we need the numbers to define the benefit, but most important
appears to be justifying what we're going to be doing since we surpassed the
secondary treatment stage.  That was really mandated by the 1972 Amendments.
We have finally come to the point where we are riding out that crest and
facing a new challenge.  We environmentalists had better show that we are
really doing some good before we go any further because we have some very real
competitors.  Energy is probably the one that is going to be very much the
lead in some of our priority issues.  We've got to talk to some of the people
in the community that are going to be receiving the benefit of the expendi-
tures we're proposing.  This is a whole new ball game and frankly we engineers
are totally inept in this area.  We've got to bring in the communications
experts.  We've got to make it simple enough so it's understood.  We can no
longer say, "this is mandated, we've got to do it."  Even when we could say
that, it was a challenge.  We are reaching the point where the regulatory
agencies are saying that we need justification and I am glad to see it
having been involved in a number of projects where we studied the impact a'fter
the facilities were designed.  It really isn't the way to go.  So, I think,
we've got to begin to bring together the forces needed to heighten our public
participation and set our objectives before we go any further.  In a way it
might be a plug for 208, although that seems to be running out of funds.

Mr. Bob Foxen, U.S. EPA, Arlington, VA

      I agree with a lot of the things that John said about the need to con-
sider a cost benefit in deciding what kind of treatment we want to provide.
A problem is the way the law is structured now.  It precludes looking at
weighing cost versus benefits in deciding what kind of treatment is required,
whether it's AWT treatment or CSO.  The law requires that we achieve fishable,
swimmable waters where attainable and as I see it, there is not room legally
for us to decide whether it's worth 10,000,000 dollars, whether 10 pounds
of fish are worth $10,000,000, so it seems that the answer is that there is
some kind of legal change that is required before we can adequately look and
weigh the costs and benefits in deciding what level of treatment is needed.
Karen, perhaps you could describe your program.

Ms. Karen Klima, U.S. EPA, Washington, D.C.

      Bob and I worked on the same program.  I'm working on the AWT reviews
and the issue there is similar to CSO control because it's a situation where
secondary treatment may not be adequate to meet the water quality goals or to
achieve the beneficial uses.  But, in the AWT review, we're not deciding
whether the cost of advanced treatment is worth the investment.  We are
                                     646

-------
not deciding Whetherit is worth the 10,000,000 dollars to add on additional
filters and whether the benefit in terms of a fishing use is worth that invest-
ment.  We are assumming that the fishable, swimmable use, is a goal  we're
shooting for.  I don't necessarily agree that it is the way to go, but I know
for a fact that we've had arguments inside the agency to the effect that we
cannot weigh cost versus benefits in deciding, whether this 10 million dollar
investment is justified.

    I have no other interpretation, I can just say that we are talking about
the subtle distinction between cost benefit analysis as opposed to cost
effectiveness analysis that we have been doing in our AWT review.  Bob and I
feel like we don't have the legislative reform to really be looking at cost
versus benefit, but we can look at cost effectiveness now.  As to how that
distinction applies to the CSO benefits that we are looking at, let me suggest
thi.s as one possible answer.  If that is the case and the agency does decide
it is a mandate from Congress, you are very fortunate because that gives your
research program time to anticipate that the next level of question iVgoing
to be, "What am I getting for my dollar," rather than, "How effectively am I
spending my dollar?"  Maybe you can use that time .effectively to get those
answers.  I don't know whether the agency position is right or wrong.  I think
it's a subtle distinction that really isn't in the public's best interest.

Dr. Fred Lee, Colorado State University, Fort Collins, CO

    There is a real undercurrent to shoot down inappropriate standards in a
very practical way.  It is not just us academic and technical guys that are
talking to each other.  I know some guys that are even thinking about putting
up some money to alter standards.                             •

Mr. Phil Graham, U.S. EPA, Washington, D.C.

    P661 was a first effort to look at the benefits because we thought CSO's
were going to be so expensive.  Most of the AWT reviews are not CSO projects,
although a few of them were,.  Some of them involved sewer separation where
it had gone so far that it was too late to even worry about it.  I think if
we clarify it for the treatment plants, that's probably the agency policy and
PG61 is also the agency policy on CSO's.  Its handbook coming out in the
future will probably say that the way to bring urban stormwater into it, you
have to consider the effects of it in your planning for both point sources
and CSO to make up for what you can't do.  That's not eligible in urban runoff
so  I don't know what will happen with that.

Dr. Mike Sonnen,

    The people from EPA are not quite correct.  They do have the legislative
mandate to do benefit analysis.  It was in the 1977 act and some before that
but it hasn't changed lately.  The numbers are Section 302B2, if you would
like to look it up, and it has to do with the administrator being warned in
advance by Congress that, if certain dischargers can show that the benefits  of
putting in advanced treatment are not as great as the cost that the admini-
strator will have to take, then take a second look at the discharge require-
ments for that discharger.  This is an area that I have worked in since 1967,
so  I know that benefit analysis has been done and EPA has- had the legislative

                                     647

-------
wherewithal! to do it.  They have also run into problems with the economy
lately and I have had people from EPA say, "Yes, but we're just not going to
enforce that section."  Well, that's a different problem.  To say that they
don't have the mandate is not true.

Mr. Dennis Athayde, U.S. EPA, Washington, DC

      I would like to stress two points that John Mancini made.  One of them
was kind of hidden and the other one was obvious.  The one that was obvious
to me was the need for us to define the relationship between combined sewer
overflow and separate storm sewer discharges.  If we are going to enter into
a controlled program for combined sewer overflows, we can't ignore the separ-
ate storm sewered areas in these cities.  We are going to have to have some
optimum type program.  That is one of our needs.

      The other thing I would like to quickly touch on is the use of models.
How are they ultimately used?  How was their output used?  The decision-makers
use models as a tool to try and help them make better decisions.  Ultimately,
the decisions are not made from the output of the computer, they are made
politically.  We understand that and we are not going to worry about the level
of detail or the complexity of our models.  We are going to go for the simpl-
est one that will give us an answer and allow us to make a good decision.

Mr. Ken Wei don, Florida Department of Transportation, Tallahassee, FL

      By virtue of that, I have a slightly different outlook than some of you.
I came to hear some definitives regarding various methodologies for handling
some of the problems due to the impact of receiving waters.  We are faced
with a situation of being able to obtain permits through DER, EPA, and so
forth.  So what happens in our case?  We run into a problem, not only with
terminologyj but in a general overall sense of a lack of understanding in
certain areas of hydraulics and hydrology.  For instance, I think we need
a definitive set of standards for non-CSO's as has already been mentioned.
Secondly, we need to assess the actual condition or standards for the location.
There are many areas in which we are invloved where we are accused of creating
an impact in which the existing area is undeveloped and by virtue of its
characteristics is worse off at times than some of the urban areas.  We need
to establish a norm and I think it should be done on a point basis because
too often we're spending too much of the taxpayers money in areas in which we
don't need to.  Again, we get back to cost benefits.

      In using these models for determination, such as was this a stream, lake,
or what-have-you,  I think we need to give more study to the idea of the
hydrology of the situation.  For instance, I know several times the frequency
analysis was mentioned and I'm quite amazed at some of the input that is used.
I mentioned that to Dr. Lee, and I realize that often times he doesn't con-
sider the frequencies, but one of the things mentioned this morning was a
five year storm used to determine a certain impact.  You may start out with a
five year storm but you don't know if you are going to end up with a 25
year event at the particular site or location.  We need to be able to use more
control points in these modeling situations if we are going to ascertain
anything at all.  TF]ere has to be sufficient data before it is ever attempted
                                     648

-------
to establish what the norm is for a specific location and I think the norm of
the location should be used primarily for the calibration of these models.

    In regards to research, development should be made for establishing meth-
odologies for input.  I'm involved quite often in trying to design outfalls,
or storm sewer systems which naturally enter into lakes.  There is no way you
are going to prevent it.  They are existing situations and we're going to
fiave them until the end of time.  We're going to have to develop some kind of
terminology, some design, whereby certain things are described, again method-
ology due to velocities.  Sediment transport seems to be one of the major
problems.  They are stirring up the bottom deposit and again we get right
back to what is the norm.  I want to encourage someone to come up with some
definitive classification for types and bodies of waters.  For types, I don't
care whether you want to consider a developed area around a swamp or streams
just remember - they are not the same and shouldn't be considered the same.
It seems li.ke everything we have is on an idealistic basis.  I just don't
thing it applies.

Dr. Dennis Lai, Clinton Bogert Assoc., E. Brunswick, NJ

    I want to explain Dennis1 comment that we should not separate combined
sewer overflow abatement from storm sewer discharge.  The annual average dis-
charge from the combined sewer area is very much comparable with the separate
sewer area.  We feel that it doesn't make sense to abate combined sewer over-
flow without abatement of stormwater discharge from the separate sewer area,
so we are proceeding with the project and hope to make a case to EPA to fund
separate sewer area abatement.  I hope that in other parts of the country
where they have separate sewer systems, they find the same situations and
we can communicate with them.  The point is to encourage seeking out the
relationship on site specific basis between CSO loads and their impacts in
urban runoff.
                                      649

-------
                      URBAN STORMWATER AND COMBINED SEWER
                   OVERFLOW-IMPACT ON RECEIVING WATER BODIES

                            EPA NATIONAL CONFERENCE
                                ORLANDO, FLORIDA
                               NOV. 26-28, 1979

                                  Summary of

             Workshop on Practical Applications and Research Needs
               for Receiving Water Responses to Urban Stormwater
                                   Session I

                                      by

                         James P. Heaney, Ph.D., P.E.
                  Dept. of Environmental Engineering Sciences
                             University of Florida
                          Gainesville, Florida  32611
     The opening session was divided into two major activities:  opening
statements by panelists, and discussion from the floor.  A summary of the
comments, not necessarily in order of importance, is presented below:

1)   Documented evidence of receiving water impacts from urban stormwater is
     scarce.

2)   Existing stream standards are inappropriate for incorporating stormwater
     impacts.  An overall response in terms of a frequency distribution would
     be an improvement.

3)   Receiving water quality "problems" need to be related to impaired bene-
     ficial uses and whether installing controls will enhance benefits.  While
     a case-by-case approach is preferred to uniform standards, it remains to
     be seen whether such an approach is practical from an administrative
     viewpoint.

4)   More formal procedures for calibrating and verifying simulation models
     are needed.  Included in such procedures would be guidelines for better
     matching the model to the data base.

5)   The importance of benthal deposits in analyzing wet-weather problems was
     stressed repeatedly.
                                     650

-------
6)   More effort is needed in using simple biological indicators to examine
     receiving waters for planning purposes.

7)   Fundamental research is needed on deoxygenation and reaeration kinetics.

8)   Information on health effects of stormwater needs to be gathered even
     though it is expensive.

     Overall, there seemed to be a consensus that we are returning to the
earlier (pre-EPA) philosophy of environmental quality management wherein the
orientation was towards public health and .the burden of proof is partially on
the regulator to justify that the proposed control program is cost-effective.


                                  Session II

                                      by,

                                John L. Mancini
                               Manhattan College
                                Bronx, New York


Receiving Water Impact Research Needs:

1)   There is a need for research and development programs to maintain a
balance between projects which address conventional and/or classical pollu-
tant variables and those projects that consider some of the issues associated
with toxics and heavy metals.  This requirement is particularly significant
in the area of impacts from intermittent urban discharges such as CSO and
urban runoff.  Programs for construction grants for facilities 'are entering
the heavy expenditure stages for CSO and are under evaluation in the urban
runoff area.  Analysis of the water quality impacts of these discharges and
identifications of the effects on water usage are currently being required
and will become more critical in the future as the large size of expenditures
and tenuous nature of impacts becomes more evident.

2)   A specific research and development need is found associated with relat-
ing the size of urban areas and receiving water bodies to distinguish situa-
tions in which there is a reasonable probability of encountering a water qua-
lity problem from those where the existence of a water quality problem is a
more remote possibility.  The- definition of what constitutes a water quality
problem is beginning to focus on water use interference and therefore research
and development projects can be specifically directed towards this definition
of water quality problems.  The assessment of water quality problems, using
interference with water usage as an operative definition, will also provide a
framework in which estimates of benefits associated with .treatment of CSO and
urban runoff can begin to be developed.

3)   EPA research and development programs should, in the opinion of the
writer, begin to address the areas of water quality problem identification
and benefits assessment.  Expenditures for environmental control have begun to
                                     651

-------
encounter increasing'competition on the federal, state, and local levels from
other programs.  In order to retain the momentum of the national drive for
environmental quality it will be necessary to identify tangible benefits to
society.  Research efforts in this respect should consider projects in each
of the following areas:

     a)  Studies of the fate and transformation of pollutants in natural
         environmental systems.  Analysis should be related to the types
         of receiving water systems, such as streams, lakes, estuaries
         and oceans and must also consider the unique characteristics of
         urban discharges.

     b)  Particular attention needs to be directed to the solids associated
         with urban discharges.  The transport and interactions of particu-
         lates and associated pollutants are key elements in evaluating
         impacts or urban discharges.

     c)  Investigation of effects of contaminants associated with urban
         runoff represents fruitful areas of research activity which can
         be carried out conjunctively with other EPA programs in toxics,
         monitoring and standards.  Particular emphasis should be directed
         towards using information from these other programs to build a
         base of knowledge which can be employed to address the unique
         aspects of urban discharges associated with the frequent yet
         intermittent nature of these impacts.
                                     652

-------
                URBAN STORMWATER AND COMBINED SEWER OVERFLOW
                      Impact on Receiving Water Bodies

                            LIST OF PARTICIPANTS
Alvarado, Luis C.
MDC
59 Selden Hill Dr.
West Hartford, CT  06107

Ammon, Douglas C.
U.S. EPA
Edison, NJ  08817
Anderson, Craig P.
Jefferson Parish Env.
3600 Jefferson Highway
Jefferson, LA  70121

App, Charles W.
U.S. EPA
716 Raynham Rd.
Collegeville, PA  19426

Athayde, Dennis
U.S.  EPA
Washington, DC  20460
Badalamenti, Salvatore
U.S. EPA
26 Federal Plaza
New York, NY  10007

Bain, Richard C.
Brown & Caldwell
100 West Harrison
Seattle, WA  98119

Barfield, Larry D.
Fla. Dept. of Transportation
2630-A Old Bainbridge Rd.
Tallahassee, FL  32303
Bel anger, Thomas V.
Florida Institute of Technology
112 W. Peekskill PI
Melbourne, FL  32907

Bell, John M.
Purdue University
3 Concord Place
Lafayette, IN  47905

Bennett, Francis P.
Reti red
14103 Rutland
Detroit, MI  48227

Berger, Arthur W.
City of Painesville
7616 Mountain Park Dr.
Mentor, OH  44060

Birkitt, Beverly F.
Dept. of Environmental Regulation
1907 Sherwood Dr.
Tallahassee, FL  32303

Blancher, El don C.
University of Florida
Gainesville, FL  32601
Boyd, Gail B.
Woodward-Clyde Consultants
3 Embarcadero Center Suite 700
San Francisco, CA  94111

Bozeman, Martin
Woodward-Clyde Consultants
3 Embarcadero Center Suite 700
San Francisco, CA  94111
                                     653

-------
Branch, James C.
Clark, Dietz Engineers, Inc.
500 W. Fulton St.
Sanford, FL  32711

Brill, Edward L.
Floyd Browne Associ. Ltd.
750 Sable Court
Youngstown, OH  44512

Brown, E. Ryland
and Community Development
N.C. Dept. of Natural Resources
Raleigh, NC  27611

Brown!ey, Dennis D.
U.S. Air Force
USAF Clinic
Randolph AFB, TX  78148

Brunner, Carl
U.S. EPA
Washington, DC  20460
Burnett, William C.
Florida State University
Dept. of Oceanography
Tallahassee, FL  32306

Byrne, Christian J.
FSU-Oceanography
1900 Nininger St. #7
Tallahassee, FL  32304

Calabrese, Mark
Reynolds, Smith, & Hills, Inc.
7120 Lake Ellenor Dr.
Orlando, FL  32856

Canale, Raymond P.
University of Michigan
Ann Arbor, MI  48109
Celikkol, Barbaros
University of New Hampshire
P.O. Box 361
Durham, NH  03824

Chojnowski, Kathy
Church, Paula H.
Dade County Dept. Env. Res. Mgmt.
909 SE 1st Ave
Miami, FL  33131

Clarke, Bill
Proctor & Redfern, Toronto
9 Cedar Ave.
Toronto, Ontario, M4E1K1

Clarkson, Robert A.
Calocerinos & Spina
1020 Seventh North St.
Liverpool, NY  13088

Cole, Michael S.
City of Syracuse Water Division
141 Hickok Ave.
Syracuse, NY  13206

Condon, Francis J.
U.S. EPA
Washington, DC  20460
Courtney, Charles M.
The Del ton Corp.
990 N. Barfield Dr.
Marco Island, FL  33937

DeZolt, Jim
Cavicchi, Paul
NH WS & PCC
3 Penwood
Penacook, NH  03303
Diniz, Elvidio V.
Espey Huston & Assoc.
2500 Louisiana NE #310
Albuquerque, NM  87110
                                     654

-------
 Diorio, Gary 0.
 Floyd Browne Assoc. Ltd.
 4631 New  England Blvd.
 Youngstown, OH  44512

 DiToro, Dominic M.
 Environmental Engr. & Sci. Prog.
•Manhattan College
 Bronx, NY 10471

 Donahue,  Charles R.
 Dept. of  Oceanography-FSU
 1305 Pull en Rd.
 Tallahassee, FL  32306

 Duba, George A.
 Western Michigan University
 220 S. Kendall #0
 Kalamazoo, MI  49007

 Duvall, Leland R.
 U.S. EPA
 6021 W. 99th
 Overland  Park, KS
                          ^y
 Ellis, Franklin W.
 CH2M Hill
 Reston, VA 22090
Field, Richard
U.S. EPA
Edison, NJ  08817
Fisher, Anthony P.
U.S. EPA
80 Hayward St.
Yonkers, NY   10701

Flood, John J.
E.T. Killam Assoc., Inc.
R.D. #1 Roxitn'icus Rd.
Far Hills, NJ  07076

Foerster, Klaus E.
Metro Waste Control Comm.
8274 Rhode Island
Minneapolis, MN  55438

Foran, Frank
George F. Young, Inc.
6797 21st Wag. S.
St. Petersburg, FL  33712

Forger, Dan              :
 El son,  Wayne  D.
 U.S.  EPA
 1514  Hill  Ave. #2
 Wheaton, IL   60187

 Evink,  Gary L.
 Fla.  Dept. of Transportation
 3419  Thresher Dr.
 Tallahassee,  FL  32303

 Falco,  James  W.
 U.S.  EPA
 225 Hampton Ct.
 Athens, GA 30605

 Fancher, Dick
 Dept. of Environmental  Regulation
 6336  HarvaHd  Ct.
 Pensacola, FL 32504
Foxen, Robert J.
U.S. EPA
4300 Old Dominion Rd.
Arlington, VA  22207

Frahn, Kurt S.
Post, Buckley, Schuh & Jernigari
1351 Yulee Dr.
Clearwater, FL  33515

Franklin, Marvin A.
U.S. Geological Survey
3044 Godfrey PI.
Tallahassee, FL  32308

Garie, Henry L.
Rutgers University
17 Myrtle Ave.
Dover, NJ  07801
                                      655

-------
Garlick, Dan
Dept. of Env. Regulation
515 Peck Ave.
Ft. Myers, FL  33907

Garrity, Rick
City of Tampa-DPW
Tampa Municipal Bldg.
Tampa, FL  33602

Gerard, David
Illinois EPA
2200 Churchill Rd.
Springfield, IL  62706

Gietz, R.J.
Regional Municipality of Ottawa
655 Shefford Rd.
Ottawa, Ontario, KIJ8G8

Glandon,- Robert P.
Michigan State University
6227 Balfour
Lansing, MI  48910

Godlewski, Victor J.
Stottler Stagg and Ags.
7878 Shoals Dr., Apt. B
Orlando, FL  32807

Golding, Bernard L.
Howard, Needles et al.
8982 Islesworth Ct.
Orlando, FL  32811

Gouin, Denyse
Govn. Quebec
C.P. 852 St. Redempteur
P. Quebec, GOS3PW

Graham, Donald S.
University of Florida
Department of Civil Engineering
Gainesville, FL  32611

Green, Raymond F.
Martin Marietta Corp.
229 Lk. Seminary Cir.
Maitland, FL  32751
Green, Robert I.
Agri-Leis Corp.
P.O. Box 6232
Lakeland, FL  32803

Greenstein, Bary P.
Collier Cty,. Bd. Cty. Comm.
1270 Cooper Dr.
Naples, FL  33940

Griffin, Thomas L.
U.S. EPA
9003 Forest Glen
Palos Park, IL  60464

Grimshaw, Herbert J.
Oklahoma Water Resource Board
2833 SW 86th St.
Oklahoma City, OK  73159

Grizzard, Thomas J.
Virginia Tech
Box 733
Manassas, VA  22110

Gruber, David A.
Halbert, Bruce E.
James F. MacLaren Ltd.
151 L'Amoreaux Dr. TH111
Baincourt Ontario, Canada

Hampson, Paul S.
U.S. Geological Survey
2600 Art Museum Dr.
Jacksonville, FL  32229

Harper, Harvey H.
University of Central  Florida
CEES Department
Orlando, FL  32816

Hawkins, Robert A.
TenEch Environmental  Const.
515 Park Ave.
Louisville, KY  40208
                                     656

-------
Heaney, James P.
University of Florida
Gainesville, FL  32601
Hentschel, Greg E.
Housel & Assoc., Inc.
202 W Bearss Ave., Suite 250
Tampa, FL  33612

Hermanson, George H.
CH2M Hill
P.O. Box 3317
Naples, FL  33939

Hever, Robert F.
Floyd Browne Assoc. Ltd.
5855 Lawnview St. NW
Canton, OH  44718

Hewitt, James L.
Espeg, Huston & Assoc.
2500 Louisiana NE #310
Albuquerque, NM  87110

Hoffman, Eva J.
University of Rhode Island
Oceanography-URI
Kingston, RI  02881

Holtkamp, Michael L.
SW Fla. Water Mgmt. Dist.
13404 Monte Carlo Ct. #48
Tampa, FL  33612

Hoyt, Dennis K.
Lamp, Rynearson & Assoc.
9290 W. Dodge Rd.
Omaha, NE  68114

Huber, Wayne C.
University of Florida
Gainesville, FL  32601
Hvitved-Jacobsen, Thorkild
Institute of Civil Engineering
Aalborg, DENMARK
James, William
McMaster University
127 Dalewood
Hamilton, Ontario,. L8J4B8

Jenness, Frederick M.
Minnesota Pollution Control
1935 West Co. Rd. B2
Roseville, MN  55113

Joiner, Thomas E.
Joiner Engineering, Inc.
113 W. Savidge St.
Spring Lake, MI  49456

Jones, Ann R.
Colorado State University
Fort Collins, CO  80521
Jouseau, Marcel RG
Metropolitan Council
300 Metro Square
St. Paul, MN  55104

Jowett, James R.
U.S. EPA
5622 Eastbourne Dr.
Springfield, VA  22151

Kaufman, Herbert L.
Clinton Bogert Assoc.
188 W. Ramapo Ave.
Mahwah, NJ  07430

Keefer, Thomas N.
Sutron, Inc.
1709 Prelude Dr.
Vienna, VA  22180

Keller, David B.
Keller & Kirkpatrick, P.A.
10 Ridgedale Ave.
Florham Park, NJ  07932

Kersten, Robert D.
College of Engineering
University of Central Florida
Orlando, FL  32816
                                     657

-------
 Klima, Karen S.
HQRS EPA
401 M St., SW
Washington, DC

Klingelhoefer, Ginger D.
Anne Arundel County
906 Primrose Rd.
Annapolis, MD  21403

Klos, Kenneth C.
Dept. of Environmental Regulation
3319 Maguire Blvd.
Orlando, FL  32803

Knauer, Gregory W.
Booker Assoc., Inc.
1822 Marriott Lane
Barnhart, .MO  63012

Knoff, Robert M.
Burlington Cty. Planning Board
716 High St.
Burlington, NJ  08016

Koval, Peter R.
O'Brien & Gere Engineers Inc.
648 Beacon St.
Boston, MA  02215

Kreutzberger, William A.
Rexmord Environmental Research
2504 So. 14th
Milwaukee, WI  53215

Kutash, Bill
Dept. of Env. Regulation
3915 McKay Ave.
Tampa, FL  33609

Kynast, Frank
Am. Soc. of Civil Engr.
Walt Disney World
Orlando, FL

Lai, Dennis F.
Clinton Bogert Assoc.
18 Manton Ave.
E. Brunswick, NJ  08816
Landon, John C.
Post Buckley Schuh & Jernigan
400 E. Merritt Ave.
Merritt Island, FL  32952

Lartigue, Date
Proctor & Gamble Co.
Ivorydale Tech. Center
Cincinnati, OH  45217

Latch, Mark M.
Dept. of Environmental Regulation
1907 Sherwood Dr.
Tallahassee, FL  32303

LaZenby, Mack
City of Sanford
P.O. Box 1778
Sanford, FL  32771

Le, Tung T.
City of Tampa
3314 W. Louisiana Ave.
Tampa, FL  33614  .

Lee, G. Fred
Colorado State University
Fort Collins, CO  80521
Librach, Austan S.
Washington Council of Govts.
1875 Eye St., NW
Washington, DC  20006

Loijens, Harry S.
222 Queen St., 10th Floor
Ottawa, Ontario KIF5V9
Lombardo, Robert J.
Civil Engineering Consultant
825 4th St. West
Palmetto, FL  33561

Lopez, Miguel A.
U.S. Geological Survey
912 Tangelo Place
Brandon, FL  33511
                                     658

-------
Malueg, Kenneth W.
CERL-U.S. EPA
200 SW 35th St.
Con/all is, OR  97330

Mancini, John L.
Manhattan College
New York, NY
Medina, Miguel A.
Duke University
Durham, NC  27706
Meinholz, Thomas L.
EcolSciences, Inc.
Milwaukee, WI
Mangan, Bob
Manz, Pete E.
Henningson, Durham & Richardson
8404 Indian Hills Dr.
Omaha, NE  68114

Markunas, Raymond
The Metropolitan District
51 Virginia Lane
Toll and, CT  06084

Markus, Ed
Marshall, Frank E.
Russell  & Axon, Inc.
1803 Peninusla Ave.
New Smyrna Beach, FL  32069

McGinn,  Joseph
MA Dept. of Env. Quality Engr.
70 W. Boylston St.
Worcester, MA  01606

McGrew,  Thomas W.
Pinellas County 7607 Palmbrook Dr.
Tampa, FL  33615

McLellon, Waldron M.
CEES Department
University of Central Florida
Orlando, FL  32816
Merrick, S. Preston
Reedy Creek Imp. Dist.
6 W. Preston St.
Orlando, FL  32804

Messenger, Allen
Texas A & M University
College Station, TX  77840
Miertschin, James D.
Espey, Huston & Assoc., Inc.
3010 South Lamar
Austin, TX  78704

Mikalsen, Karsten T.
Georgia Env. Protection Div.
3326 Wheeler Dr.
Atlanta, GA  30340

Mil gram, Otto
E.T. Kellam Associates, Inc.
2375 Monica PI.
Scotch Plains, NJ  07076

Miller, Lynn
SW Fla. Water Mgmt. Dist.
5060 U.S. Hwy. 41 South
Brooksville, FL  33512

Miller, Robert A.
U.S. Geological Survey
109 Bagberry Rd.
Longwood, FL  32750

Moffa, Peter E.
Sterns & Whele'r
Cazenovia,  NY  13035
                                     ,659

-------
Molzahn, Robert E.
Harza Engineering Company
150 S. Wacker Dr.
Chicago, IL  60606

Moreda, Lydia S.
City of Tampa-DPW
4604 John Moore Rd.
Brandon, FL  33511

Moray, Lawrence J.
Watson and Company
2811 E. South St.
Orlando, FL  32803

Motchkavitz, Robert J.
The Del tons Corp.
3250 SW 3rd Ave.
Miami, FL  33129

Mueller, Thomas W.
Post, Buckley, Schuh & Jernigan
5001 Lacasa Ct. #494
Tampa, FL  33617

Murali, R.S.
Dept. of Environmental Regulation
2600 Blairstone Rd.
Tallahassee, FL  32301

Musgrove, Richard J.
NW Fla. Water Mgmt. Dist.
Rt. 1, Box 3100
Havana, FL  32333

Nagy, JanetrA.
Joiner Engineering, Inc.
113 W. Savidge St.
Spring Lake, MI  49456

Noss, Richard R.
M.I.T.
106 Centre St.
Milton, MA  02186
                          /
O'Toole, Michael C.
U.S. EPA
570 Frederick Lane
Hoffman Estates, IL  60195
Ott, Randy R.
Onondaga County
125 Elwood Davis Rd.
N. Syracuse, NY  13212

Perez, Armando I.
Post, Buckley, Schuh & Jernigan
7500 NW 52nd St.
Miami, FL  33166

Peterson, Ray S.
U.S. EPA
23924 Florence Acres
Monroe, WA  98727

Pew, Kenneth A.
Watermation, Inc.
8321 Celianna Dr.
Stringsville, OH  44136

Pierro, Robert H.
Civil Engineering Consultant
825 4th St. West
Palmetto, FL  33561

Pitt, Robert
Woodward-Clyde Consultants
San Francisco, CA
Porc.ella, Donald B.
Tetra Tech Inc.
3746 Mt. Diablo Blvd.
Lafayette, CA  94549

Poruczniak, Thomas
Pratt, Benjamin C.
Lee County-County Commissioners
P.O. Box 398
Ft. Myers, FL  33902

Ragan, Robert E.
Calocerinos & Spina
1020 Seventh North St.
Liverpool, NY  13088
                                     660

-------
Ramsey, Frederick V.
Environmental Science & Engr.
P.O. Box 13454
Gainesville, FL  32604

Ray, Larry T.
Reynolds, Tom D.
Texas A & M University
College Station, TX  77840
Risley, Clifford
U.S. EPA
536 S. Clark St., 10th Floor
Chicago, IL  60605

Saini, Mohinder K.
Dept. of the Army
MAEN-A Bldg. 677-A
West Point, NY  10996'

Sartor, James D.            ;
Woodward-Clyde Consultants
3 Embarcadero Center Suite 700
San Francisco, CA  94111

Schenk, John E.
ENCOTEC
3983 Research Park Dr.
Ann Arbor, MI  48105

Schmidt, Edward J.
University of New Hampshire
46 Madbury Rd.
Durham, NH  03824

Schwartz, Larry A.
Waccaman Reg. Plan. Council
134 Clemson Rd.
Conway, SC  29526

Sell en, Joyce
Seminole County
1 North Park Ave.
Sanford, FL  32771
Shubinsky, Robert
Water Resources Engineers
Springfield, VA
Simons, Allyson A.
Johnson Engineering Inc.
P.O. Box 1550
Ft. Myers, FL  33902

Singer, Sam N.
Ontario Ministry of the Environment
3 Massey Sq. Apt. 1601
Toronto, Ontario, Canada

Sloane, Joanna
University of Washington
2020 E. Howe St.
Seattle, WA  98112

Smedile, Joseph A.
N.E. Illinois Ping. Com.
2929 Indian Wood Rd.
Wilmette, IL  60091

Smith, John P.
Florida Institute of Technology
255 Monaco Rd., West
Melbourne, FL  32901

Smith, William W.
FMC Corp.
44 Lookover Lane
Yardley, PA  19067

Smolenski, Frank
Smolenyaic, Kevin J.
The Deltona Corp.
3250 SW 3rd Ave.
Miami, FL  33129

Snow, Madeline
Mass. Dept. of Env. Quality Engr.
88 Kenrick St.
Brighton, MA  02135
                                     661

-------
Snyder, Bruce
University of Central Florida
CEES Department
Orlando, FL  32816

Solanki, Himat T.
County of Sarasota
3012 Bucida Dr.
Sarasota, FL  33582

Sonnen, Michael B.
The M.B. Sonnen Co., Inc.
2855 Mitchell Dr. #215
Walnut Creek, CA  94598

Spence, Jeff
Stalker, Richard D.
Area Planning Board
2300 Palm Beach Lake Blvd.
West Palm Beach, FL  33409

Taylor, James S.
University of Central Florida
CEES Department
Orlando, FL  32816

Terstriep, Mike
Illinois Water Survey
Box 232
Urbana, IL  61874

Thora, Ronald M.
University of Washington
13729 27th Ave. NE
Seattle, WA  98125

Thomas, Dan
Timmons, Wilson R.
Brevard Co. Water Res. Dept.
Rt. 1, Box 195
Cocoa, FL  32922

Tindale, Steven A.
City of Tampa-DPW
Tampa Municipal Bldg.
Tampa, FL  33602

Todd, William L.
Post, Buckley, Schuh & Jernigan
6326 Presidential Ct.
Ft. Myers, FL  33907
To!pa, Robert D.
U.S. EPA
1925 -Oakton St.,
Des Plaines, IL
                                                       Apt.  2-C
                                                       60018
Tomasello, Richard S.
Airan Consultants
900 Overbrook PI.
W. Palm Beach, FL  33406

Tomlinson, Richard D.
Municipality of Metro. Seattle
821 2nd Ave.
Seattle, WA  98104

Tompkins, Marian L.
Iowa Dept. Env. Quality
3211 30th St.
Des Moines, IA  50310

Tully, William P.
SUNY-CESF
109 Emann Dr.
Canillas, NY  13031

Turkeltaub, Robert
U.S. EPA
Edison, NJ  08817
Tiemens, Myron
U.S. EPA
Washington, DC  20460
Turner, Ralph R.
Union Carbide
Oak Ridge Nat'l  Lab
Oak Rtdge, TN  37830
                                     662

-------
Unger, Heinz K.
Reid, Crowther & Partners Ltd.
c/o R.C.P.L. #5 Oxford Park Ave.
Kingston 5, Jamaica

Valetutto, Angelo J.
Metcalf & Eddy, Inc.
524 Linden Ave.
Woodbridge, NJ  07095

VanNote, Kirby
Professional Services. Group
835 No. County Rd. 18
Minneapolis, NM  55427

Wagener, Fritz
U.S. EPA
345 Courland St.
Decatur, GA  30033

Walesky, Richard
Dept. of Env. Regulation
1141 Fernlea Dr.
West Palm Beach, FL  33409

Wanielista, Martin P.
University of Central Florida
CEES Department
Orlando, FL  32816

Ward, Jeffery J.
U.S. Army Corps of Engr.
625 Shelter Creek Lane
San Bruno, CA  94066

Watkins, Frank
Dept. of Env. Regulation
9645 Bay Meadows Rd. #713
Jacksonville, FL  32216

Wei don, Ken
Welter, Gregory J.
O'Brien & Gore Engineers
1432 Lawrence St., NE
Washington, DC  20017
Wiegand, Cameron
Metro Washington Council/Gov'ts
1875 Eye St., N.W.
Washington, DC  20006

Wiemhoff, John R.
U.S. EPA
109 Norwood Ct.
Rolling Meadows, 60008

Williamson, James A.
Finkbeiner, Pettis & Strout
3709 Harley
Toledo, OH  43613

Woodruff, Gary I.
Tulsa City-Cty. Health Dept.
4616 E. 15th St.
Tulsa, OK  74112

Wycoff, Ronald L.
CH2M Hill
2521 NW 63rd Terr.
Gainesville, FL  32601

Yaap, Warren E.
Private Practice
124 Firehorn Rd.
Gulf Breeze, FL  32561

Young, Michael D.
St. Johns River W.M.D.
Rt. 2, Box 93
Mel rose, FL  32666

Yousef, Yousef A.
University of Central Florida
CEES Department
Orlando, FL  32816

Zanoni, Al E.
Marquette University
1515 W. Wisconsin Ave.
Milwaukee, WI  53233
                                     663

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
1. REPORT NO.

  EPA-600/9- 80-056
                              2.
                                                           3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
                   URBAN STORMWATER AND COMBINED SEWER
 OVERFLOW  IMPACT ON RECEIVING WATER BODIES  -  Proceedings
 of the National  Conference, Orlando, Florida
 November  26-28,  1979
               5. REPORT DATE
                 December 1980
               6. PERFORMING ORGANIZATION CODE
7, AUTHOR(S)
 Yousef A. Yousef,  Martin P. Wanielista, Waldron M.
 McLellon and James S.  Taylor - Editors
                                                           8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS

 University of  Central  Florida
 Orlando, Florida   32816
                                                           10. PROGRAM ELEMENT NO.
                  C35B1C
               11.
                         '/GRANT NO.
                                                              R-806715
12. SPONSORING AGENCY NAME AND ADDRESS
 Municipal Environmental  Research Laboratory -  Cinn.,  OH
 Office of Research  and Development
 U.S. Environmental  Protection Agency
 Cincinnati, Ohio  45268
               13. TYPE OF REPORT AND PERIOD COVERED
                Proceedings Nov. 26-28. 1979
               14. SPONSORING AGENCY CODE
                  EPA-600/14
15, SUPPLEMENTARY NOTES
 Project Officer:  Robert Turkeltaub, Storm and
 Environmental Research Laboratory, Cinn., OH,
    Combined Sewer  Section,  Municipal
     45268 - FTS  340-6679,  (201) 321-6679
16. ABSTRACT
     The conference  provided a forum for researchers,  practioners and others to  re-
 ceive an update on  the state-of-the-art and to  learn  about research findings dealing
 with stormwater impact.   It also served to stimulate  dialogue among those who are in-
 terested in stormwater effects and control, regarding the implication and applications
 of current research results, particularly from  those  projects supported by the  U.S.
 Environmental Protection Agency, Municipal Environmental  Research Laboratory's  Storm
 and Combined Sewer  Program.
     The main topical  areas  considered included:   Combined sewer overflow costs  vs.
 benefits; impacts on  lakes, rivers and estuaries; ecological  response to stormwater
 and methodologies for stormwater impact assessment; and stormwater management through
 the use of receiving  water  quality models for planning and abatement methodology.
     These proceedings contain the contributions from  the  scheduled speakers and an
 edited transcription  of the taped workshop conducted  on; practical applications  of re-
 search findings and future  research needs.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
 b.lDENTIFIERS/OPEN ENDED TERMS  C.  COS AT I Field/Group
 Water pollution, Water quality,  Combined
 sewer, Meetings, Impact,  Storm sewers,
 Management, Cost-effectiveness
   Urban runoff, Receiving
   water, Impacts, Control
   methodologies, Toxics,
   Biological effects, Sedi
   ment, Models, Dissolved
   oxygen, Pathogens
       13B
18, DISTRIBUTION STATEMENT


   RELEASE TO PUBLIC
  19. SECURITY CLASS (ThisReport)

   UNCLASSIFIED
21. NO. OF PAGES
  672
  20. SECURITY CLASS (Thispage)

   UNCLASSIFIED
                            22. PRICE
EPA Fotm 2220-1 (Rev. 4-77)
664
                                                             U.S. GOVERNMENT PRINTING OFFICE: 1980—757-064/0195

-------