Watershed Protection Research Monograph No. 1 IMPACTS

IMPACTSof Impervious

Cover onAquatic Systems

Center forWatershedProtection

March 2003

Watershed Protection Research Monograph No. 1

Cover photograph Ellicott City, Maryland 2003.Courtesy Anne Kitchell, Center for Watershed Protection.

Impacts of ImperviousCover on Aquatic Systems

March 2003

Prepared by:Center for Watershed Protection

8391 Main StreetEllicott City, MD 21043

www.cwp.orgwww.stormwatercenter.net

Copyright (c) 2003 by the Center for Watershed Protection.Material may be quoted provided credit is given. Printed in the

United States of Amercia on recycled paper.

Watershed Protection Research Monograph No. 1

Impacts of Impervious Cover on Aquatic Systems i

Foreword

Foreword

We are extremely pleased to launch the firstedition of a new series called WatershedProtection Research Monographs. Eachmonograph will synthesize emerging researchwithin a major topical area in the practice ofwatershed protection. The series of periodicmonographs will replace our journalWatershed Protection Techniques, whichlapsed in 2002. We hope this new format willprovide watershed managers with the scienceand perspectives they need to better protect andrestore their local watersheds.

This monograph was written to respond tomany inquiries from watershed managers andpolicy makers seeking to understand thescientific basis behind the relationship betweenimpervious cover and the health of aquaticecosystems. It reviews more than 225 researchstudies that have explored the impact ofimpervious cover and other indicators ofurbanization on aquatic systems. This reportcomprehensively reviews the available scien-tific data on how urbanization influenceshydrologic, physical, water quality, andbiological indicators of aquatic health, as oflate 2002.

Our intention was to organize the availablescientific data in a manner that was accessibleto watershed leaders, policy-makers andagency staff. In addition, the research itself,which spans dozens of different academicdepartments and disciplines, was conducted inmany different eco-regions, climatic zones,and stream types. In order to communicate

across such a wide audience, we have resortedto some simplifications, avoided some impor-tant particulars, refrained from some jargon,and tried, wherever possible, to use consistentterminology. Thus, the interpretations andconclusions contained in this document areours alone, and our readers are encouraged toconsult the original sources when in doubt.

We would also like to note that the Center forWatershed Protection and the University ofAlabama are currently developing a majornational database on stormwater quality. Thedatabase will contain nearly 4,000 station-storm events collected by municipalities as partof the U.S. EPA’s National Pollutant DischargeElimination System (NPDES) Phase I Storm-water Permit Program. We anticipate releasinga data report in late 2003 that will provide amuch needed update of stormwater event meanconcentrations (EMCs).

As of this writing, many research efforts areunderway that will further test and refine theserelationships (most notably, the U.S. Geologi-cal Survey gradients initiative, but also manyother local, state and academic efforts). Wehope that this report provides a useful sum-mary of the existing science, suggests somedirections for new research, and stimulatesgreater discussion of this important topic inwatershed management. We also feel it is timefor a major conference or symposium, wherethis diverse community can join together todiscuss methods, findings and the importantpolicy implications of their research.

ii Impacts of Impervious Cover on Aquatic Systems

Foreword

Impacts of Impervious Cover on Aquatic Systems iii

Acknowledgments

Acknowledgments

Putting this first research monograph together took a lot of energy, editing and analysis, andmany Center staff devoted their time and energy over the last two years to get it done. Theproject team consisted of Karen Cappiella, Deb Caraco, Samantha Corbin, Heather Holland,Anne Kitchell, Stephanie Linebaugh, Paul Sturm, and Chris Swann. Special thanks are extendedto Tiffany Wright, who worked tirelessly to assemble, edit and otherwise polish the final draft.

I am also grateful to Michael Paul of Tetratech, Inc., who graciously provided us with an exten-sive literature review from his PhD days at the University of Georgia that contained manyobscure and hard to find citations. Portions of this monograph were developed as part of aliterature review conducted as part of a work assignment for the U.S. EPA Office of WastewaterManagement in 2001, which proved indispensable in our efforts. Lastly, I would like to thank thehundreds of scientists who have contributed their time and data to explore and test the relation-ships between urbanization and aquatic health.

Tom SchuelerCenter for Watershed Protection

iv Impacts of Impervious Cover on Aquatic Systems

Acknowledgments

Impacts of Impervious Cover on Aquatic Systems v

Table of Contents

Table of Contents

Foreword ..................................................................................................................................... iAcknowledgments .................................................................................................................................... iiiList of Acronyms and Abbreviations ........................................................................................................... xi

Chapter 1: Introduction1.1 Is Impervious Cover Still Important? A Review of Recent Stream Research ............................ 1

1.1.1 Strength of the Evidence for the ICM..................................................................... 31.1.2 Reinterpretation of the ICM ..................................................................................... 51.1.3 Influence of Watershed Treatment Practices on the ICM ................................... 91.1.4 Recommendations for Further ICM Research ..................................................... 12

1.2 Impacts of Urbanization on Downstream Receiving Waters .................................................... 141.2.1 Relationship Between Impervious Cover and Stormwater Quality ................... 141.2.2 Water Quality Response to Stormwater Pollution................................................ 151.2.3 Effect of Watershed Treatment on Stormwater Quality ..................................... 18

1.3 Implications of the ICM for Watershed Managers .................................................................... 211.3.1 Management of Non-Supporting Streams............................................................ 211.3.2 Use of the ICM for Urban Stream Classification .................................................. 221.3.3 Role of the ICM In Small Watershed Planning ..................................................... 22

1.4 Summary .................................................................................................................................. 24

Chapter 2: Hydrologic Impacts of Impervious Cover2.1 Introduction .................................................................................................................................. 252.2 Increased Runoff Volume ............................................................................................................. 272.3 Increased Peak Discharge Rate.................................................................................................. 302.4 Increased Bankfull Flow ................................................................................................................ 312.5 Decreased Baseflow ..................................................................................................................... 342.6 Conclusions .................................................................................................................................. 37

Chapter 3: Physical Impacts of Impervious Cover3.1 Difficulty in Measuring Habitat ..................................................................................................... 40

3.1.1 The Habitat Problem ............................................................................................... 403.2 Changes in Stream Geometry ..................................................................................................... 42

3.2.1 Channel Enlargement ............................................................................................ 423.2.2 Effect of Channel Enlargement on Sediment Yield ............................................ 45

3.3 Effect on Composite Measures of Stream Habitat .................................................................... 463.4 Effect on Individual Elements of Stream Habitat ....................................................................... 47

3.4.1 Bank Erosion and Bank Stability ............................................................................. 473.4.2 Embeddedness ....................................................................................................... 473.4.3 Large Woody Debris (LWD) .................................................................................... 493.4.4 Changes in Other Individual Stream Parameters ............................................... 49

3.5 Increased Stream Warming ......................................................................................................... 503.6 Alteration of Stream Channel Networks ..................................................................................... 52

3.6.1 Channel Modification ............................................................................................ 523.6.2 Barriers to Fish Migration ......................................................................................... 53

3.7 Conclusion .................................................................................................................................. 54

Chapter 4: Water Quality Impacts of Impervious Cover4.1 Introduction .................................................................................................................................. 554.2 Summary of National and Regional Stormwater Pollutant Concentration Data .................. 56

4.2.1 National Data .......................................................................................................... 564.2.2 Regional Differences Due to Rainfall ................................................................... 564.2.3 Cold Region Snowmelt Data ................................................................................. 58

4.3 Relationship Between Pollutant Loads and Impervious Cover: The Simple Method ............. 614.4 Sediment .................................................................................................................................. 63

vi Impacts of Impervious Cover on Aquatic Systems

Table of Contents

4.4.1 Concentrations ....................................................................................................... 634.4.2 Impacts of Sediment on Streams .......................................................................... 634.4.3 Sources and Source Areas of Sediment .............................................................. 64

4.5 Nutrients .................................................................................................................................. 674.5.1 Concentrations ....................................................................................................... 674.5.2 Impacts of Nutrients on Streams ........................................................................... 684.5.3 Sources and Source Areas of Nutrients ................................................................ 69

4.6 Trace Metals .................................................................................................................................. 714.6.1 Concentrations ....................................................................................................... 714.6.2 Impacts of Metals on Streams ............................................................................... 724.6.3 Sources and Source Areas of Trace Metals ......................................................... 73

4.7 Hydrocarbons: PAH, Oil and Grease ......................................................................................... 754.7.1 Concentrations ....................................................................................................... 754.7.2 Impacts of Hydrocarbons on Streams .................................................................. 754.7.3 Sources and Source Areas of Hydrocarbons ...................................................... 76

4.8 Bacteria & Pathogens .................................................................................................................. 774.8.1 Concentrations ....................................................................................................... 774.8.2 Impacts of Bacteria and Pathogens on Streams ................................................ 794.8.3 Sources and Source Areas of Bacteria and Pathogens .................................... 80

4.9 Organic Carbon ............................................................................................................................ 824.9.1 Concentrations ....................................................................................................... 824.9.2 Impacts of Organic Carbon on Streams .............................................................. 824.9.3 Sources and Source Areas of Total Organic Carbon ......................................... 82

4.10 MTBE .................................................................................................................................. 834.10.1 Concentrations ....................................................................................................... 834.10.2 Impacts of MTBE on Streams.................................................................................. 834.10.3 Sources and Source Areas of MTBE ...................................................................... 84

4.11 Pesticides ............................................................................................................................... 854.11.1 Concentrations ....................................................................................................... 864.11.2 Impacts of Pesticides on Streams ......................................................................... 864.11.3 Sources and Source Areas of Pesticides .............................................................. 87

4.12 Deicers .................................................................................................................................. 884.12.1 Concentrations ....................................................................................................... 894.12.2 Impacts of Deicers on Streams ............................................................................. 894.12.3 Sources and Source Areas of Deicers .................................................................. 90

4.13 Conclusion ............................................................................................................................. 91

Chapter 5: Biological Impacts of Impervious Cover5.1 Introduction .................................................................................................................................. 935.2 Indicators and General Trends .................................................................................................... 95

5.2.1 Biological Indicators ................................................................................................. 955.2.2 Watershed Development Indices .......................................................................... 955.2.3 General Trends ......................................................................................................... 97

5.3 Effects on Aquatic Insect Diversity ............................................................................................ 1005.3.1 Findings Based on Impervious Cover Indicators .................................................. 1005.3.2 Findings Based on Other Development Indicators ............................................. 104

5.4 Effects on Fish Diversity ............................................................................................................... 1055.4.1 Findings Based on Impervious Cover Indicators .................................................. 1055.4.2 Findings Based on other Development Indicators .............................................. 110

5.5 Effects on Amphibian Diversity .................................................................................................. 1125.6 Effects on Wetland Diversity ....................................................................................................... 1145.7 Effects on Freshwater Mussel Diversity ...................................................................................... 1155.8 Conclusion ................................................................................................................................ 116

References ................................................................................................................................ 117Glossary ................................................................................................................................ 137

Impacts of Impervious Cover on Aquatic Systems vii

Table of Contents

List of Tables

1 The Strength of Evidence: A Review of Current Research on Urban StreamQuality Indicators ............................................................................................................................. 4

2 Land Use/IC Relationships for Suburban Areas of the Chesapeake Bay .................................. 93 Summary of Urban Stormwater Pollutant Loads on Quality of Receiving Waters ................... 144 The Effectiveness of Stormwater Treatment Practices in Removing

Pollutants - Percent Removal Rate ............................................................................................... 185 Median Effluent Concentrations from Stormwater Treatment Practices ................................. 196 Additional Considerations for Urban Stream Classification ....................................................... 237 Research Review of Increased Runoff Volume and Peak Discharge in Urban Streams ........ 288 Hydrologic Differences Between a Parking Lot and a Meadow .............................................. 299 Comparison of Bulk Density for Undisturbed Soils and Common Urban Conditions ............... 2910 Research Review of Increased Bankfull Discharge in Urban Streams ...................................... 3211 Research Review of Decreased Baseflow in Urban Streams .................................................... 3412 Physical Impacts of Urbanization on Streams ............................................................................. 4113 Research Review of Channel Enlargement and Sediment Transport in Urban Streams ........ 4314 Research Review of Changes in Urban Stream Habitat ............................................................ 4815 Research Review of Thermal Impacts in Urban Streams ........................................................... 5016 National EMCs for Stormwater Pollutants .................................................................................... 5717 Regional Groupings by Annual Rainfall Amount ........................................................................ 5818 Stormwater Pollutant EMCs for Different U.S. Regions ................................................................ 5919 Mean and Median Nutrient and Sediment Stormwater Concentrations for

Residential Land Use Based on Rainfall Regions ......................................................................... 5920 EPA 1986 Water Quality Standards and Percentage of Metal Concentrations

Exceeding Water Quality Standards by Rainfall Region ............................................................ 6021 Runoff and Pollutant Characteristics of Snowmelt Stages ........................................................ 6022 EMCs for Total Suspended Solids and Turbidity ........................................................................... 6323 Summary of Impacts of Suspended Sediment on the Aquatic Environment .......................... 6424 Summary of Impacts of Deposited Sediments on the Aquatic Environment .......................... 6425 Sources and Loading of Suspended Solids Sediment in Urban Areas ...................................... 6526 Source Area Geometric Mean Concentrations for Suspended Solids in Urban Areas .......... 6627 Mean TSS Inflow and Outflow at Uncontrolled, Controlled and Model Construction Sites ... 6628 EMCs of Phosphorus and Nitrogen Urban Stormwater Pollutants ............................................. 6729 Source Area Monitoring Data for Total Nitrogen and Total Phosphorus in Urban Areas ........ 6930 EMCs and Detection Frequency for Metals in Urban Stormwater ............................................ 7131 Average Total Recoverable and Dissolved Metals for 13 Stormwater Flows and

Nine Baseflow Samples from Lincoln Creek in 1994 ................................................................... 7232 Percentage of In-situ Flow-through Toxicity Tests Using Daphnia magna and

Pimephales promelas with Significant Toxic Effects from Lincoln Creek .................................. 7333 Metal Sources and Source Area “Hotspots” in Urban Areas ..................................................... 7434 Metal Source Area Concentrations in the Urban Landscape .................................................. 7435 Hydrocarbon EMCs in Urban Areas .............................................................................................. 7536 Bacteria EMCs in Urban Areas ...................................................................................................... 7837 Cryptosporidium and Giardia EMCs............................................................................................ 7938 Percent Detection of Giardia cysts and Cryptosporidium oocysts in Subwatersheds and

Wastewater Treatment Plant Effluent in the New York City Water Supply Watersheds .......... 8039 Typical Coliform Standards for Different Water Uses .................................................................. 8040 EMCs for Organic Carbon in Urban Areas .................................................................................. 8241 MTBE Detection Frequency ........................................................................................................... 8342 Median Concentrations and Detection Frequency of Herbicides and

Insecticides in Urban Streams ....................................................................................................... 8543 Use and Water Quality Effect of Snowmelt Deicers ................................................................... 8844 EMCs for Chloride in Snowmelt and Stormwater Runoff in Urban Areas ................................. 8945 Summary of State Standards for Salinity of Receiving Waters ................................................... 90

viii Impacts of Impervious Cover on Aquatic Systems

Table of Contents

46 Review of Stressors to Urban Streams and Effects on Aquatic Life ........................................... 9447 Examples of Biodiversity Metrics Used to Assess Aquatic Communities ................................... 9648 Alternate Land Use Indicators and Significant Impact Levels .................................................. 9849 Recent Research Examining the Relationship Between IC and Aquatic

Insect Diversity in Streams ............................................................................................................ 10150 Recent Research Examining the Relationship of Other Indices of Watershed

Development on Aquatic Insect Diversity in Streams .............................................................. 10251 Recent Research Examining the Relationship Between Watershed IC

and the Fish Community .............................................................................................................. 10652 Recent Research Examining Urbanization and Freshwater Fish Community Indicators ...... 10853 Recent Research on the Relationship Between Percent Watershed

Urbanization and the Amphibian Community .......................................................................... 11354 Recent Research Examining the Relationship Between Watershed Development

and Urban Wetlands .................................................................................................................... 114

List of Figures

1 Impervious Cover Model ................................................................................................................. 22 Typical Scatter Found in IC/Stream Quality Indicator Research ................................................ 53 Relationship of IC and FC in Puget Sound Subwatersheds ......................................................... 64 The Double Scatter Problem: Difficulties in Detecting the Effect of Watershed Treatment .. 105 Estimated Phosphorous Load as a Function of IC, Discounted Stormwater

Treatment and Better Site Design ................................................................................................. 196 Altered Hydrograph in Response to Urbanization ...................................................................... 257 Runoff Coefficient vs. IC ................................................................................................................ 278 Discharge for Urban and Rural Streams in North Carolina ........................................................ 309 Effect on Flood Magnitudes of 30% Basin IC ............................................................................... 3110 Relationship of Urban/Rural 100-year Peak Flow Ratio to Basin

Development Factor and IC ......................................................................................................... 3111 Increase in Bankfull Flows Due to Urbanization ........................................................................... 3212 Increase in Number of Exceedences of Bankfull Flow Over Time With Urbanization ............. 3313 Percent of Gage Reading Above Mean Annual Flow .............................................................. 3314 Relationship Between Baseflow and Watershed Impervious Cover ........................................ 3415 Baseflow Response to Urbanization ............................................................................................. 3516 Relationship Between Percentage Baseflow and Percent IC .................................................. 3517 Effect of IC on Summer Baseflow (Corrected for Catchment Area) ....................................... 3618 Effect of Watershed IC on Summer Stream Velocity .................................................................. 3619 Urban Stream Channels with Progressively Greater IC .............................................................. 4020 Increased Shear Stress from a Hydrograph ................................................................................. 4221 Stream Channel Enlargement in Watts Branch .......................................................................... 4422 Ultimate Channel Enlargement .................................................................................................... 4523 Relationship Between Habitat Quality and IC in Maine Streams .............................................. 4624 Fine Material Sediment Deposition as a Function of IC ............................................................. 4725 LWD as a Function of IC ................................................................................................................. 4926 Stream Temperature Increase in Response to Urbanization ..................................................... 5127 Drainage Network of Rock Creek, D.C. and Four Mile Creek, VA

Before and After Urbanization ...................................................................................................... 5228 Fish Migration Barriers in the Urbanized Anacostia Watershed ................................................. 5329 Snowmelt Runoff Hydrograph ....................................................................................................... 6030 The Simple Method - Basic Equations .......................................................................................... 6131 TSS from Bank Erosion vs. IC ........................................................................................................... 6532 Nitrate-Nitrogen Concentration in Stormwater Runoff .............................................................. 6833 Total Phosphorus Concentration in Stormwater ......................................................................... 6834 Total Phosphorus From Bank Erosion as a Function of IC ........................................................... 7035 Fecal Coliform Levels in Urban Stormwater ................................................................................ 77

Impacts of Impervious Cover on Aquatic Systems ix

Table of Contents

36 Relationship Between IC and Fecal Coliform Concentrations ................................................. 7937 MTBE Concentrations in Surface Water from Eight Cities .......................................................... 8338 Concentrations of Pesticides in Stormwater in King County, WA ............................................. 8739 U.S. Highway Salt Usage Data ....................................................................................................... 8840 Combined Fish and Benthic IBI vs. IC ........................................................................................... 9841 Relationship Between B-IBI, Coho/Cutthroat Ratios, and Watershed IC ................................. 9942 Index for Biological Integrity as a Function of Population Density ............................................ 9943 Trend Line Indicating Decline in Benthic IBI as IC Increases .................................................... 10344 Compilation of Puget Lowland Watershed Biological Data ................................................... 10345 IC and IBI at Stream Sites in the Patapsco River Basin, MD ..................................................... 10346 IC vs. Aquatic Insect Sensitivity - EPT scores in Delaware Streams .......................................... 10347 Average and Spring EPT Index Values vs. % IC in 20 Small Watersheds in Maine.................. 10448 Fish IBI vs. Watershed IC for Streams in Patapsco River Basin, MD .......................................... 10549 Fish IBI and Number of Species vs. % IC in Wisconsin Streams ................................................. 10750 IC and Effects on Fish Species Diversity in Four Maryland Subwatersheds ............................ 10751 Coho Salmon/Cutthroat Trout Ratio for Puget Sound .............................................................. 10952 Mean Proportion of Fish Taxa in Urban and Non-Urban Streams,

Valley Forge Watershed, PA ........................................................................................................ 11053 Relationship Between Watershed Population Density and Stream IBI Scores ....................... 11154 Amphibian Species Richness as a Function of Watershed IC in

Puget Sound Lowland Wetlands ................................................................................................. 113

x Impacts of Impervious Cover on Aquatic Systems

Table of Contents

Impacts of Impervious Cover on Aquatic Systems xi

Acronyms and Abbreviations


B-IBI Benthic Index of Biotic IntegrityBOD Biological Oxygen DemandBSD Better Site DesignC-IBI Combined Index of Biotic Integritycfs cubic feet per secondCOD Chemical Oxygen DemandCSO Combined Sewer OverflowCu CopperDOC Dissolved Organic Carbondu/ac dwelling units per acreEMC Event Mean ConcentrationEPT Ephemeroptera, Plecoptera and

TrichopteraFC Forest CoverGIS Geographic Information SystemsIBI Index of Biotic IntegrityIC Impervious CoverICM Impervious Cover Modellbs/ac pounds per acreLWD Large Woody Debrismg/kg milligrams per kilogrammg/l milligrams per liter (equal to ppm)MPN Most Probable NumberMTBE Methyl Tertiary-Butyl EtherN Number of StudiesN/R data not reportedNO

2Nitrite

NO3

Nitrate

NOx

Nitrogen OxidesNPDES National Pollutant

Discharge Elimination SystemNTU Nephelometric Turbidity UnitNURP National Urban Runoff ProgramPAH Polycylic Aromatic HydrocarbonsPCB Polychlorinated Biphenylppb Parts per billion (equal to ug/l)ppm Parts per million (equal to mg/l)RBP Rapid Bioassessment ProtocolSLAMM Source Loading Assessment/

Management ModelSPMD Semi-Permeable Membrane DeviceSSO Sanitary Sewer OverflowSTP Stormwater Treatment PracticeTC Turf CoverTDS Total Dissolved SolidsTKN Total Kjeldhal NitrogenTMDL Total Maximum Daily LoadTotal N Total NitrogenTotal P Total PhosphorousTOC Total Organic CarbonTSS Total Suspended Solidsug/l micrograms per liter (equal to ppb)VMT Vehicle Miles TraveledVOC Volatile Organic CompoundWLF Water Level FluctuationWTP Wastewater Treatment Plant

xii Impacts of Impervious Cover on Aquatic Systems


Impacts of Impervious Cover on Aquatic Systems 1

Chapter 1: Introduction


This research monograph comprehensivelyreviews the available scientific data on theimpacts of urbanization on small streams andreceiving waters. These impacts are generallyclassified according to one of four broadcategories: changes in hydrologic, physical,water quality or biological indicators. Morethan 225 research studies have documented theadverse impact of urbanization on one or moreof these key indicators. In general, mostresearch has focused on smaller watersheds,with drainage areas ranging from a few hun-dred acres up to ten square miles.

Streams vs. DownstreamReceiving Waters

Urban watershed research has traditionallypursued two core themes. One theme hasevaluated the direct impact of urbanization onsmall streams, whereas the second theme hasexplored the more indirect impact of urbaniza-tion on downstream receiving waters, such asrivers, lakes, reservoirs, estuaries and coastalareas. This report is organized to profile recentresearch progress in both thematic areas and todiscuss the implications each poses for urbanwatershed managers.

When evaluating the direct impact of urbaniza-tion on streams, researchers have emphasizedhydrologic, physical and biological indicatorsto define urban stream quality. In recent years,impervious cover (IC) has emerged as a keyparadigm to explain and sometimes predicthow severely these stream quality indicatorschange in response to different levels ofwatershed development. The Center forWatershed Protection has integrated theseresearch findings into a general watershedplanning model, known as the imperviouscover model (ICM). The ICM predicts thatmost stream quality indicators decline whenwatershed IC exceeds 10%, with severe

degradation expected beyond 25% IC. In thefirst part of this review, we critically analyzethe scientific basis for the ICM and exploresome of its more interesting technical implica-tions.

While many researchers have monitored thequality of stormwater runoff from smallwatersheds, few have directly linked thesepollutants to specific water quality problemswithin streams (e.g., toxicity, biofouling,eutrophication). Instead, the prevailing view isthat stormwater pollutants are a downstreamexport. That is, they primarily influencedownstream receiving water quality. There-fore, researchers have focused on how toestimate stormwater pollutant loads and thendetermine the water quality response of therivers, lakes and estuaries that receive them.To be sure, there is an increasing recognitionthat runoff volume can influence physical andbiological indicators within some receivingwaters, but only a handful of studies haveexplored this area. In the second part of thisreview, we review the impacts of urbanizationon downstream receiving waters, primarilyfrom the standpoint of stormwater quality. Wealso evaluate whether the ICM can be extendedto predict water quality in rivers, lakes andestuaries.

This chapter is organized as follows:

1.1 A Review of Recent Urban StreamResearch and the ICM

1.2 Impacts of Urbanization on DownstreamReceiving Waters

1.3 Implications of the ICM for WatershedManagers

2 Impacts of Impervious Cover on Aquatic Systems


1.1 A Review of Recent UrbanStream Research and the ICM

In 1994, the Center published “The Importanceof Imperviousness,” which outlined the scien-tific evidence for the relationship between ICand stream quality. At that time, about twodozen research studies documented a reason-ably strong relationship between watershed ICand various indicators of stream quality. Theresearch findings were subsequently integratedinto the ICM (Schueler, 1994a and CWP,1998). A brief summary of the basic assump-tions of the ICM can be found in Figure 1. TheICM has had a major influence in watershedplanning, stream classification and land useregulation in many communities. The ICM is adeceptively simple model that raises extremelycomplex and profound policy implications forwatershed managers.

The ICM has been widely applied in manyurban watershed settings for the purposes ofsmall watershed planning, stream classifica-tion, and supporting restrictive developmentregulations and watershed zoning. As such, theICM has stimulated intense debate among theplanning, engineering and scientific communi-

ties. This debate is likely to soon spill over intothe realm of politics and the courtroom, givenits potential implications for local land use andenvironmental regulation. It is no wonder thatthe specter of scientific uncertainty is fre-quently invoked in the ICM debate, given theland use policy issues at stake. In this light, itis helpful to review the current strength of theevidence for and against the ICM.

The ICM is based on the following assump-tions and caveats:

• Applies only to 1st, 2nd and 3rd orderstreams.

• Requires accurate estimates of percent IC,which is defined as the total amount ofimpervious cover over a subwatershedarea.

• Predicts potential rather than actual streamquality. It can and should be expected thatsome streams will depart from the predic-tions of the model. For example, monitor-ing indicators may reveal poor waterquality in a stream classified as “sensitive”or a surprisingly high biological diversity

Watershed Impervious Cover

Stre

am

Qua

lity

Figure 1: Impervious Cover Model



score in a “non-supporting” one. Conse-quently, while IC can be used to initiallydiagnose stream quality, supplementalfield monitoring is recommended toactually confirm it.

• Does not predict the precise score of anindividual stream quality indicator butrather predicts the average behavior of agroup of indicators over a range of IC.Extreme care should be exercised if theICM is used to predict the fate of indi-vidual species (e.g., trout, salmon, mus-sels).

• “Thresholds” defined as 10 and 25% IC arenot sharp “breakpoints,” but instead reflectthe expected transition of a composite ofindividual indicators in that range of IC.Thus, it is virtually impossible to distin-guish real differences in stream qualityindicators within a few percentage pointsof watershed IC (e.g., 9.9 vs. 10.1%).

• Should only be applied within theecoregions where it has been tested,including the mid-Atlantic, Northeast,Southeast, Upper Midwest, and PacificNorthwest.

• Has not yet been validated for non-streamconditions (e.g., lakes, reservoirs, aquifersand estuaries).

• Does not currently predict the impact ofwatershed treatment.

In this section, we review available streamresearch to answer four questions about theICM:

1. Does recent stream research still supportthe basic ICM?

2. What, if any, modifications need to bemade to the ICM?

3. To what extent can watershed practicesshift the predictions of the ICM?

4. What additional research is needed to testthe ICM?

1.1.1 Strength of the Evidencefor the ICM

Many researchers have investigated the IC/stream quality relationship in recent years. TheCenter recently undertook a comprehensiveanalysis of the literature to assess the scientificbasis for the ICM. As of the end of 2002, wediscovered more than 225 research studies thatmeasured 26 different urban stream indicatorswithin many regions of North America. Weclassified the research studies into three basicgroups.

The first and most important group consists ofstudies that directly test the IC/stream qualityindicator relationship by monitoring a largepopulation of small watersheds. The secondand largest group encompasses secondarystudies that indirectly support the ICM byshowing significant differences in streamquality indicators between urban and non-urban watersheds. The third and last group ofstudies includes widely accepted engineeringmodels that explicitly use IC to directly predictstream quality indicators. Examples includeengineering models that predict peak dischargeor stormwater pollutant loads as a directfunction of IC. In most cases, these relation-ships were derived from prior empiricalresearch.

Table 1 provides a condensed summary ofrecent urban stream research, which shows theimpressive growth in our understanding ofurban streams and the watershed factors thatinfluence them. A negative relationshipbetween watershed development and nearly allof the 26 stream quality indicators has beenestablished over many regions and scientificdisciplines. About 50 primary studies havetested the IC/stream quality indicator relation-ship, with the largest number looking atbiological indicators of stream health, such asthe diversity of aquatic insects or fish. Another150 or so secondary studies provide evidencethat stream quality indicators are significantlydifferent between urban and non-urban water-sheds, which lends at least indirect support forthe ICM and suggests that additional researchto directly test the IC/stream quality indicator



Table 1: The Strength of Evidence: A Review of the Current Research on Urban Stream Indicators

Stream Quality Indicator # IC UN EM RV Notes

Increased Runoff Volume 2 Y Y Y N extensive national data

Increased Peak Discharge 7 Y Y Y Y type of drainage system key

Increased Frequency of Bankfull Flow 2 ? Y N N hard to measure

Diminished Baseflow 8 ? Y N Y inconclusive data

Stream Channel Enlargement 8 Y Y N Y stream type important

Increased Channel Modification 4 Y Y N ? stream enclosure

Loss of Riparian Continuity 4 Y Y N ? can be affected by buffer

Reduced Large Woody Debris 4 Y Y N ? Pacific NW studies

Decline in Stream Habitat Quality 11 Y Y N ?

Changes in Pool Riffle/Structure 4 Y Y N ?

Reduced Channel Sinuosity 1 ? Y N ? straighter channels

Decline in Streambed Quality 2 Y Y N ? embeddedness

Increased Stream Temperature 5 Y Y N ? buffers and ponds also a factor

Increased Road Crossings 3 ? Y N ? create fish barriers

Increased Nutrient Load 30+ ? Y Y N higher stormwater EMCs

Increased Sediment Load 30+ ? Y N Y higher EMCs in arid regions

Increased Metals & Hydrocarbons 20+ ? Y Y N related to traffic/VMT

Increased Pesticide Levels 7 ? Y N Y may be related to turf cover

Increased Chloride Levels 5 ? Y N Y related to road density

Violations of Bacteria Standards 9 Y Y N Y indirect association

Decline in Aquatic Insect Diversity 33 Y Y N N IBI and EPT

Decline in Fish Diversity 19 Y Y N N regional IBI differences

Loss of Coldwater Fish Species 6 Y Y N N trout and salmon

Reduced Fish Spawning 3 Y Y N ?

Decline in Wetland Plant Diversity 2 N Y N ? water level fluctuation

Decline in Amphibian Community 5 Y Y N ? few studies

#: total number of all studies that evaluated the indicator for urban watershedsIC: does balance of studies indicate a progressive change in the indicator as IC increases? Answers: Yes, No or No data(?)UN: If the answer to IC is no, does the balance of the studies show a change in the indicator from non-urban to urbanwatersheds? Yes or No EM Is the IC/stream quality indicator relationship implicitly assumed within the framework of widely accepted engineeringmodels? Yes, No or No models yet exist (?) RV: If the relationship has been tested in more than one eco-region, does it generally show major differences betweenecoregions? Answers: Yes, No, or insufficient data (?)

Table 1: The Strength of Evidence:A Review of the Current Research on Urban Stream Indicators



relationship is warranted. In some cases, theIC/stream quality indicator relationship isconsidered so strongly established by historicalresearch that it has been directly incorporatedinto accepted engineering models. This hasbeen particularly true for hydrological andwater quality indicators.

1.1.2 Reinterpretation of the ICM

Although the balance of recent stream researchgenerally supports the ICM, it also offersseveral important insights for interpreting andapplying the ICM, which are discussed next.

Statistical VariabilityScatter is a common characteristic of most IC/stream quality indicator relationships. In most

cases, the overall trend for the indicator isdown, but considerable variation exists alongthe trend line. Often, linear regression equa-tions between IC and individual stream qualityindicators produce relatively modest correla-tion coefficients (reported r2 of 0.3 to 0.7 areoften considered quite strong).

Figure 2 shows typical examples of the IC/stream quality indicator relationship thatillustrate the pattern of statistical variability.Variation is always encountered when dealingwith urban stream data (particularly so forbiological indicators), but several patterns existthat have important implications for watershedmanagers.

d. Biological Condition vs. Total Watershed IC (Booth, 2000)

Figure 2: Typical Scatter Found in IC/Stream Quality Indicator Research

a. Fish IBI vs. IC in Fairfax, VA (Fairfax County, 2001) b. CPSS vs. IC in Montgomery County, MD (MNCPPC, 2000)

c. Large Woody Debris vs. IC (Booth et al., 1997)



The first pattern to note is that the greatestscatter in stream quality indicator scores isfrequently seen in the range of one to 10% IC.These streams, which are classified as “sensi-tive” according to the ICM, often exhibit low,moderate or high stream quality indicatorscores, as shown in Figure 2. The key interpre-tation is that sensitive streams have the poten-tial to attain high stream quality indicatorscores, but may not always realize this poten-tial.

Quite simply, the influence of IC in the one to10% range is relatively weak compared toother potential watershed factors, such aspercent forest cover, riparian continuity,historical land use, soils, agriculture, acid minedrainage or a host of other stressors. Conse-quently, watershed managers should never relyon IC alone to classify and manage streams inwatersheds with less than 10% IC. Rather, theyshould evaluate a range of supplementalwatershed variables to measure or predictactual stream quality within these lightlydeveloped watersheds.

The second important pattern is that variabilityin stream quality indicator data is usually

dampened when IC exceeds 10%, whichpresumably reflects the stronger influence ofstormwater runoff on stream quality indicators.In particular, the chance that a stream qualityindicator will attain a high quality score issharply diminished at higher IC levels. Thistrend becomes pronounced within the 10 to25% IC range and almost inevitable whenwatershed IC exceeds 25%. Once again, thispattern suggests that IC is a more robust andreliable indicator of overall stream qualitybeyond the 10% IC threshold.

Other Watershed Variables and the ICMSeveral other watershed variables can poten-tially be included in the ICM. They includeforest cover, riparian forest continuity and turfcover.

Forest cover (FC) is clearly the main rival toIC as a useful predictor of stream quality inurban watersheds, at least for humid regions ofNorth America. In some regions, FC is simplythe reciprocal of IC. For example, Horner andMay (1999) have demonstrated a stronginterrelationship between IC and FC forsubwatersheds in the Puget Sound region(Figure 3). In other regions, however, “pre-

Figure 3: Relationship of IC and FC in Puget Sound Subwatersheds(Horner and May, 1999)



development” land use represents a complexmosaic of crop land, pasture and forest.Therefore, an inverse relationship between FCand IC may not be universal for subwatershedsthat have witnessed many cycles of deforesta-tion and cultivation.

It should come as little surprise that theprogressive loss of FC has been linked todeclining stream quality indicators, given thatforested watersheds are often routinely used todefine natural reference conditions for streams(Booth, 2000 and Horner et al., 2001). Matureforest is considered to be the main benchmarkfor defining pre-development hydrology withina subwatershed, as well. Consequently, FC isperhaps the most powerful indicator to predictthe quality of streams within the “sensitive”category (zero to 10% IC).

To use an extreme example, one would expectthat stream quality indicators would respondquite differently in a subwatershed that had90% FC compared to one that had 90% cropcover. Indeed, Booth (1991) suggests thatstream quality can only be maintained when ICis limited to less than 10% and at least 65% FCis retained within a subwatershed. The keymanagement implication then is that streamhealth is best managed by simultaneouslyminimizing the creation of IC and maximizingthe preservation of native FC.

FC has also been shown to be useful in predict-ing the quality of terrestrial variables in asubwatershed. For example, the Mid-AtlanticIntegrated Assessment (USEPA, 2000) hasdocumented that watershed FC can reliablypredict the diversity of bird, reptile and am-phibian communities in the mid-Atlanticregion. Moreover, the emerging discipline oflandscape ecology provides watershed manag-ers with a strong scientific foundation fordeciding where FC should be conserved in awatershed. Conservation plans that protect andconnect large forest fragments have beenshown to be effective in conserving terrestrialspecies.

Riparian forest continuity has also shownconsiderable promise in predicting at leastsome indicators of stream quality for urban

watersheds. Researchers have yet to come upwith a standard definition of riparian continu-ity, but it is usually defined as the proportionof the perennial stream network in asubwatershed that has a fixed width of maturestreamside forest. A series of studies indicatesthat aquatic insect and fish diversity areassociated with high levels of riparian continu-ity (Horner et al., 2001; May et al., 1997;MNCPPC, 2000; Roth et al., 1998). On theother hand, not much evidence has beenpresented to support the notion that ripariancontinuity has a strong influence on hydrologyor water quality indicators.

One watershed variable that received littleattention is the fraction of watershed areamaintained in turf cover (TC). Grass oftencomprises the largest fraction of land areawithin low-density residential developmentand could play a significant role in streams thatfall within the “impacted” category (10 to 25%IC). Although lawns are pervious, they havesharply different properties than the forests andfarmlands they replace (i.e., irrigation, com-pacted soils, greater runoff, and much higherinput of fertilizers and pesticides, etc.). It isinteresting to speculate whether the combinedarea of IC and TC might provide better predic-tions about stream health than IC area alone,particularly within impacted subwatersheds.

Several other watershed variables might haveat least supplemental value in predictingstream quality. They include the presence ofextensive wetlands and/or beaverdam com-plexes in a subwatershed; the dominant formof drainage present in the watershed (tiledrains, ditches, swales, curb and gutters, stormdrain pipes); the average age of development;and the proximity of sewer lines to the stream.As far as we could discover, none of thesevariables has been systematically tested in acontrolled population of small watersheds. Wehave observed that these factors could beimportant in our field investigations and oftenmeasure them to provide greater insight intosubwatershed behavior.

Lastly, several watershed variables that areclosely related to IC have been proposed topredict stream quality. These include popula-



tion, percent urban land, housing density, roaddensity and other indices of watershed devel-opment. As might be expected, they generallytrack the same trend as IC, but each has somesignificant technical limitations and/or difficul-ties in actual planning applications (Brown,2000).

Individual vs. Multiple IndicatorsThe ICM does not predict the precise score ofindividual stream quality indicators, but ratherpredicts the average behavior of a group ofindicators over a range of IC. Extreme careshould be exercised if the ICM is used topredict the fate of individual indicators and/orspecies. This is particularly true for sensitiveaquatic species, such as trout, salmon, andfreshwater mussels. When researchers haveexamined the relationship between IC andindividual species, they have often discoveredlower thresholds for harm. For example,Boward et al. (1999) found that brook troutwere not found in subwatersheds that had morethan 4% IC in Maryland, whereas Horner andMay (1999) asserted an 8% threshold forsustaining salmon in Puget Sound streams.

The key point is that if watershed managerswant to maintain an individual species, theyshould be very cautious about adopting the10% IC threshold. The essential habitatrequirements for many sensitive or endangeredspecies are probably determined by the mostsensitive stream quality indicators, rather thanthe average behavior of all stream qualityindicators.

Direct Causality vs. AssociationA strong relationship between IC and decliningstream quality indicators does not always meanthat the IC is directly responsible for thedecline. In some cases, however, causality canbe demonstrated. For example, increasedstormwater runoff volumes are directly causedby the percentage of IC in a subwatershed,although other factors such as conveyance,slope and soils may play a role.

In other cases, the link is much more indirect.For these indicators, IC is merely an index ofthe cumulative amount of watershed develop-

ment, and more IC simply means that a greaternumber of known or unknown pollutantsources or stressors are present. In yet othercases, a causal link appears likely but has notyet been scientifically demonstrated. A goodexample is the more than 50 studies that haveexplored how fish or aquatic insect diversitychanges in response to IC. While the majorityof these studies consistently shows a verystrong negative association between IC andbiodiversity, they do not really establish whichstressor or combination of stressors contributesmost to the decline. The widely acceptedtheory is that IC changes stream hydrology,which degrades stream habitat, and in turnleads to reduced stream biodiversity.

Regional DifferencesCurrently, the ICM has been largely confirmedwithin the following regions of North America:the mid-Atlantic, the Northeast, the Southeast,the upper Midwest and the Pacific Northwest.Limited testing in Northern California, thelower Midwest and Central Texas generallyagrees with the ICM. The ICM has not beentested in Florida, the Rocky Mountain West,and the Southwest. For a number of reasons, itis not certain if the ICM accurately predictsbiological indicators in arid and semiaridclimates (Maxted, 1999).

Measuring Impervious CoverMost researchers have relied on total impervi-ous cover as the basic unit to measure IC at thesubwatershed level. The case has repeatedlybeen made that effective impervious cover isprobably a superior metric (e.g., only countingIC that is hydraulically connected to thedrainage system). Notwithstanding, mostresearchers have continued to measure total ICbecause it is generally quicker and does notrequire extensive (and often subjective)engineering judgement as to whether it isconnected or not. Researchers have used awide variety of techniques to estimatesubwatershed IC, including satellite imagery,analysis of aerial photographs, and derivationfrom GIS land use layers. Table 2 presentssome standard land use/IC relationships thatwere developed for suburban regions of theChesapeake Bay.



Three points are worth noting. First, it is fair tosay that most researchers have spent morequality control effort on their stream qualityindicator measurements than on theirsubwatershed IC estimates. At the current time,no standard protocol exists to estimatesubwatershed IC, although Cappiella andBrown (2001) presented a useful method. Atbest, the different methods used to measure ICmake it difficult to compare results fromdifferent studies, and at worst, it can introducean error term of perhaps +/- 10% from the truevalue within an individual subwatershed.Second, it is important to keep in mind that ICis not constant over time; indeed, majorchanges in subwatershed IC have been ob-served within as few as two years. Conse-quently, it is sound practice to obtainsubwatershed IC estimates from the mostrecent possible mapping data, to ensure that itcoincides with stream quality indicator mea-surements. Lastly, it is important to keep inmind that most suburban and even rural zoningcategories exceed 10% IC (see Table 2).Therefore, from a management standpoint,planners should try to project future IC, inorder to determine the future stream classifica-tion for individual subwatersheds.

1.1.3 Influence of WatershedTreatment Practices on the ICM

The most hotly debated question about theICM is whether widespread application ofwatershed practices such as stream buffers orstormwater management can mitigate theimpact of IC, thereby allowing greater devel-opment density for a given watershed. At thispoint in time, there are fewer than 10 studiesthat directly bear on this critical question.Before these are reviewed, it is instructive tolook at the difficult technical and scientificissues involved in detecting the effect ofwatershed treatment, given its enormousimplications for land use control and watershedmanagement.

The first tough issue is how to detect the effectof watershed treatment, given the inherentscatter seen in the IC/stream quality indicatorrelationship. Figure 4 illustrates the “doublescatter” problem, based on three differenturban stream research studies in Delaware,Maryland and Washington. A quick inspectionof the three plots shows how intrinsically hardit is to distinguish the watershed treatmenteffect. As can be seen, stream quality indica-tors in subwatersheds with treatment tend to

Land Use Category

SampleNumber

(N)

MeanIC (SE)

Land UseCategory

SampleNumber

(N)

MeanIC (SE)

Agriculture 8 1.9 – 0.3 Institutional 30 34.4 – 3.45

Open Urban Land 11 8.6 – 1.64 Light 20 53.4 – 2.8

2 Acre Lot Residential 12 10.6 – 0.65 Commercia 23 72.2 – 2.0

1 Acre Lot Residential 23 14.3 – 0.53 Churches 8 39.9 – 7.8 1

1/2 Acre Lot Residential 20 21.2 – 0.78 Schools 13 30.3 – 4.8

1/4 Acre Lot Residential 23 27.8 – 0.60 Municipals 9 35.4 – 6.3

1/8 Acre Lot Residential 10 32.6 – 1.6 Golf 4 5.0 – 1.7

Townhome Residential 20 40.9 – 1.39 Cemeteries 3 8.3 – 3.5

Multifamily Residential 18 44.4 – 2.0 Parks 4 12.5 – 0.7

Table 2: Land Use/IC Relationships forSuburban Areas of the Chesapeake Bay

(Cappiella and Brown, 2001)



overplot those in subwatersheds that lacktreatment. While subtle statistical differencesmay be detected, they are not visibly evident.This suggests that the impact of watershedtreatment would need to be extremely dramaticto be detected, given the inherent statisticalvariability seen in small watersheds (particu-larly so within the five to 25% IC range wherescatter is considerable).

In an ideal world, a watershed study designwould look at a controlled population of smallurban watersheds that were developed with andwithout watershed practices to detect theimpact of “treatment.” In the real world,however, it is impossible to strictly controlsubwatershed variables. Quite simply, no twosubwatersheds are ever alike. Each differsslightly with respect to drainage area, IC,

forest cover, riparian continuity, historical landuse, and percent watershed treatment. Re-searchers must also confront other real worldissues when designing their watershed treat-ment experiments.

For example, researchers must carefullychoose which indicator or group of indicatorswill be used to define stream health. IC has anegative influence on 26 stream qualityindicators, yet nearly all of the watershedtreatment research so far has focused on just afew biological indicators (e.g., aquatic insector fish diversity) to define stream health. It isconceivable that watershed treatment mighthave no effect on biological indicators, yethave a positive influence on hydrology, habitator water quality indicators. At this point, fewof these indicators have been systematically

a. Horner and May, 1999

c. Maxted and Shaver, 1997

Figure 4: The Double Scatter Problem: Difficulties in Detecting theEffect of Watershed Treatment

b. MNCPPC, 2000

a. b.

c.



tested in the field. It is extremely doubtful thatany watershed practice can simultaneouslyimprove or mitigate all 26 stream qualityindicators, so researchers must carefullyinterpret the outcomes of their watershedtreatment experiments.

The second issue involves how to quantifywatershed treatment. In reality, watershedtreatment collectively refers to dozens ofpractices that are installed at individual devel-opment sites in the many years or even decadesit takes to fully “build out” a subwatershed.Several researchers have discovered thatwatershed practices are seldom installedconsistently across an entire subwatershed. Insome cases, less than a third of the IC in asubwatershed was actually treated by anypractice, because development occurred priorto regulations; recent projects were exempted,waived or grandfathered; or practices wereinadequately constructed or maintained(Horner and May, 1999 and MNCPPC, 2000).

Even when good coverage is achieved in awatershed, such as the 65 to 90% reported instudies of stormwater ponds (Jones et al.,1996; Maxted, 1999; Maxted and Shaver,1997), it is still quite difficult to quantify theactual quality of treatment. Often, eachsubwatershed contains its own unique mix ofstormwater practices installed over severaldecades, designed under diverse design crite-ria, and utilizing widely different stormwatertechnologies. Given these inconsistencies,researchers will need to develop standardprotocols to define the extent and quality ofwatershed treatment.

Effect of Stormwater PondsWith this in mind, the effect of stormwaterponds and stream buffers can be discussed.The effect of larger stormwater ponds inmitigating the impacts of IC in small water-sheds has received the most scrutiny to date.This is not surprising, since larger ponds oftencontrol a large fraction of their contributingsubwatershed area (e.g. 100 to 1,000 acres) andare located on the stream itself, thereforelending themselves to easier monitoring. Threestudies have evaluated the impact of largestormwater ponds on downstream aquatic

insect communities (Jones et al., 1996; Maxtedand Shaver, 1997; Stribling et al., 2001). Eachof these studies was conducted in smallheadwater subwatersheds in the mid-AtlanticRegion, and none was able to detect majordifferences in aquatic insect diversity instreams with or without stormwater ponds.

Four additional studies statistically evaluatedthe stormwater treatment effect in largerpopulations of small watersheds with varyingdegrees of IC (Horner and May, 1999; Horneret al., 2001; Maxted, 1999; MNCPPC, 2000).These studies generally sampled larger water-sheds that had many stormwater practices butnot necessarily complete watershed coverage.In general, these studies detected a small butpositive effect of stormwater treatment relativeto aquatic insect diversity. This positive effectwas typically seen only in the range of five to20% IC and was generally undetected beyondabout 30% IC. Although each author washesitant about interpreting his results, allgenerally agreed that perhaps as much as 5%IC could be added to a subwatershed whilemaintaining aquatic insect diversity, giveneffective stormwater treatment. Forest reten-tion and stream buffers were found to be veryimportant, as well. Horner et al. (2001) re-ported a somewhat stronger IC threshold forvarious species of salmon in Puget Soundstreams.

Some might conclude from these initialfindings that stormwater ponds have little or novalue in maintaining biological diversity insmall streams. However, such a conclusionmay be premature for several reasons. First,the generation of stormwater ponds that wastested was not explicitly designed to protectstream habitat or to prevent downstreamchannel erosion, which would presumablypromote aquatic diversity. Several states haverecently changed their stormwater criteria torequire extended detention for the expresspurpose of preventing downstream channelerosion, and these new criteria may exert astronger influence on aquatic diversity. In-stead, their basic design objective was tomaximize pollutant removal, which they didreasonably well.



The second point to stress is that streams withlarger stormwater ponds should be considered“regulated streams” (Ward and Stanford,1979), which have a significantly alteredaquatic insect community downstream of theponds. For example, Galli (1988) has reportedthat on-stream wet stormwater ponds shift thetrophic structure of the aquatic insect commu-nity. The insect community above the pondwas dominated by shredders, while the insectcommunity below the pond was dominated byscrapers, filterers and collectors. Of particularnote, several pollution-sensitive species wereeliminated below the pond. Galli reported thatchanges in stream temperatures, carbon supplyand substrate fouling were responsible for thedownstream shift in the aquatic insect commu-nity. Thus, while it is clear that large stormwa-ter ponds can be expected to have a negativeeffect on aquatic insect diversity, they couldstill exert positive influence on other streamquality indicators.

Effect of Stream BuffersA handful of studies have evaluated biologicalindicator scores for urban streams that haveextensive forest buffers, compared to streamswhere they were mostly or completely absent(Horner and May, 1999; Horner et al., 2001;May et al., 1997; MNCPPC, 2000; Roth et al.,1998; Steedman, 1988). Biological indicatorsincluded various indices of aquatic insect, fishand salmon diversity. Each study sampled alarge population of small subwatersheds over arange of IC and derived a quantitative measureto express the continuity, width and forestcover of the riparian buffer network withineach subwatershed. Riparian forests werehypothesized to have a positive influence onstream biodiversity, given the direct ways theycontribute to stream habitat (e.g., shading,woody debris, leaf litter, bank stability, andorganic carbon supply).

All five studies detected a small to moderatepositive effect when forested stream bufferswere present (frequently defined as at leasttwo-thirds of the stream network with at least100 feet of stream side forest). The greatesteffect was reported by Horner and May (1999)and Horner et al. (2001) for salmon streams in

the Puget Sound ecoregion. If excellentriparian habitats were preserved, they generallyreported that fish diversity could be maintainedup to 15% IC, and good aquatic insect diversitycould be maintained with as much as 30% IC.Steedman (1988) reported a somewhat smallereffect for Ontario streams. MNCPPC (2000),May et al. (1997), and Roth et al. (1998) couldnot find a statistically significant relationshipbetween riparian quality and urban streamquality indicators but did report that mostoutliers (defined as higher IC subwatershedswith unusually high biological indicatorscores) were generally associated with exten-sive stream side forest.

1.1.4 Recommendations forFurther ICM Research

At this point, we recommend three researchdirections to improve the utility of the ICM forwatershed managers. The first direction is toexpand basic research on the relationshipbetween IC and stream quality indicators thathave received little scrutiny. In particular,more work is needed to define the relationshipbetween IC and hydrological and physicalindicators such as the following:

• Physical loss or alteration of the streamnetwork

• Stream habitat measures• Riparian continuity• Baseflow conditions during dry weather

In addition, more watershed research is neededin ecoregions and physiographic areas wherethe ICM has not yet been widely tested. Keyareas include Florida, arid and semiaridclimates, karst areas and mountainous regions.The basic multiple subwatershed monitoringprotocol set forth by Schueler (1994a) can beused to investigate IC/stream quality relation-ships, although it would be wise to measure awider suite of subwatershed variables beyondIC (e.g., forest cover, turf cover, and ripariancontinuity).

The second research direction is to moreclearly define the impact of watershed treat-ment on stream quality indicators. Based on



the insurmountable problems encountered incontrolling variation at the subwatershed level,it may be necessary to abandon the multiplewatershed or paired watershed samplingapproaches that have been used to date.Instead, longitudinal monitoring studies withinindividual subwatersheds may be a morepowerful tool to detect the effect of watershedtreatment. These studies could track changes instream quality indicators in individualsubwatersheds over the entire developmentcycle: pre-development land use, clearing,construction, build out, and post construction.In most cases, longitudinal studies would takefive to 10 years to complete, but they wouldallow watershed managers to measure andcontrol the inherent variability at thesubwatershed level and provide a “before andafter” test of watershed treatment. Of course, alarge population of test subwatersheds wouldbe needed to satisfactorily answer the water-shed treatment question.

The third research direction is to monitormore non-supporting streams, in order toprovide a stronger technical foundation forcrafting more realistic urban stream standardsand to see how they respond to various water-

shed restoration treatments. As a general rule,most researchers have been more interested inthe behavior of sensitive and impacted streams.The non-supporting stream category spans awide range of IC, yet we do not really under-stand how stream quality indicators behaveover the entire 25 to 100% IC range.

For example, it would be helpful to establishthe IC level at the upper end of the rangewhere streams are essentially transformed intoan artificial conveyance system (i.e., becomepipes or artificial channels). It would also beinteresting to sample more streams near thelower end of the non-supporting category (25to 35% IC) to detect whether stream qualityindicators respond to past watershed treatmentor current watershed restoration efforts. Forpractical reasons, the multiple subwatershedsampling approach is still recommended tocharacterize indicators in non-supportingstreams. However, researchers will need toscreen a large number of non-supportingsubwatersheds in order to identify a fewsubwatersheds that are adequate for subsequentsampling (i.e., to control for area, IC, develop-ment age, percent watershed treatment, type ofconveyance systems, etc.).



1.2 Impacts of Urbanization onDownstream Receiving Waters

In this section, we review the impacts ofurbanization on downstream receiving waters,primarily from the standpoint of impactscaused by poor stormwater quality. We beginby looking at the relationship between IC andstormwater pollutant loadings. Next, wediscuss the sensitivity of selected downstreamreceiving waters to stormwater pollutant loads.Lastly, we examine the effect of watershedtreatment in reducing stormwater pollutantloads.

1.2.1 Relationship BetweenImpervious Cover andStormwater Quality

Urban stormwater runoff contains a wide rangeof pollutants that can degrade downstream

water quality (Table 3). Several generalizationscan be supported by the majority of researchconducted to date. First, the unit area pollutantload delivered by stormwater runoff to receiv-ing waters increases in direct proportion towatershed IC. This is not altogether surprising,since pollutant load is the product of theaverage pollutant concentration and stormwa-ter runoff volume. Given that runoff volumeincreases in direct proportion to IC, pollutantloads must automatically increase when ICincreases, as long the average pollutant con-centration stays the same (or increases). Thisrelationship is a central assumption in mostsimple and complex pollutant loading models(Bicknell et al., 1993; Donigian and Huber,1991; Haith et al., 1992; Novotny and Chester,1981; NVPDC, 1987; Pitt and Voorhees,1989).

The second generalization is that stormwaterpollutant concentrations are generally similar

Pollutants in UrbanStormwater

WQ Impacts To: HigherUnit

Load?

Load a functionof IC?

Other Factors Important in

LoadingR L E A W

Suspended Sediment Y Y Y N Y Y [ag] Y channel erosion

Total Nitrogen N N Y Y N Y [ag] Y septic systems

Total Phosphorus Y Y N N Y Y [ag] Y tree canopy

Metals Y Y Y ? N Y Y vehicles

Hydrocarbons Y Y Y Y Y Y ? related to VMTs andhotspots

Bacteria/Pathogens Y Y Y N Y Y Y many sources

Organic Carbon N ? ? ? Y Y Y

MTBE N N N Y Y Y ? roadway, VMTs

Pesticides ? ? ? ? Y Y ? turf/landscaping

Chloride ? Y N Y Y Y ? road density

Trash/Debris Y Y Y N ? Y Y curb and gutters

Major Water Quality Impacts Reported for: R = River, L = Lake, E = Estuary, A = Aquifer, W = Surface Water Supply Higher Unit Area Load? Yes (compared to all land uses) [ag]: with exception of cropland Load a function of IC? Yes, increases proportionally with IC

Pollutants in UrbanStormwater

WQ Impacts To: HigherUnit

Load?

Load a functionof IC?

Other Factors Important in

LoadingR L E A W

Suspended Sediment Y Y Y N Y Y [ag] Y channel erosion

Total Nitrogen N N Y Y N Y [ag] Y septic systems

Total Phosphorus Y Y N N Y Y [ag] Y tree canopy

Metals Y Y Y ? N Y Y vehicles

Hydrocarbons Y Y Y Y Y Y ? related to VMTs andhotspots

Bacteria/Pathogens Y Y Y N Y Y Y many sources

Organic Carbon N ? ? ? Y Y Y

MTBE N N N Y Y Y ? roadway, VMTs

Pesticides ? ? ? ? Y Y ? turf/landscaping

Chloride ? Y N Y Y Y ? road density

Trash/Debris Y Y Y N ? Y Y curb and gutters

Major Water Quality Impacts Reported for: R = River, L = Lake, E = Estuary, A = Aquifer, W = Surface Water Supply Higher Unit Area Load? Yes (compared to all land uses) [ag]: with exception of cropland Load a function of IC? Yes, increases proportionally with IC

Table 3: Summary of Urban Stormwater Pollutant Loadson Quality of Receiving Waters



at the catchment level, regardless of the mix ofIC types monitored (e.g., residential, commer-cial, industrial or highway runoff). Severalhundred studies have examined stormwaterpollutant concentrations from small urbancatchments and have generally found that thevariation within a catchment is as great as thevariation between catchments. Runoff concen-trations tend to be log-normally distributed,and therefore the long term “average” concen-tration is best expressed by a median value. Itshould be kept in mind that researchers havediscovered sharp differences in pollutantconcentrations for smaller, individual compo-nents of IC (e.g., rooftops, parking lots, streets,driveways and the like). Since most urbancatchments are composed of many kinds of IC,this mosaic quality tempers the variability inlong term pollutant concentrations at thecatchment or subwatershed scale.

The third generalization is that median concen-trations of pollutants in urban runoff areusually higher than in stormwater runoff frommost other non-urban land uses. Consequently,the unit area nonpoint pollutant load generatedby urban land normally exceeds that of nearlyall watershed land uses that it replaces (forest,pasture, cropland, open space — see Table 3).One important exception is cropland, whichoften produces high unit area sediment andnutrient loads in many regions of the country.In these watersheds, conversion of intensivelymanaged crops to low density residentialdevelopment may actually result in a slightlydecreased sediment or nutrient load. On theother hand, more intensive land development(30% IC or more) will tend to equal or exceedcropland loadings.

The last generalization is that the effect of ICon stormwater pollutant loadings tends to beweakest for subwatersheds in the one to 10%IC range. Numerous studies have suggestedthat other watershed and regional factors mayhave a stronger influence, such as the underly-ing geology, the amount of carbonate rock inthe watershed, physiographic region, local soiltypes, and most important, the relative fractionof forest and crop cover in the subwatershed(Herlihy et al., 1998 and Liu et al., 2000). The

limited influence of IC on pollutant loads isgenerally consistent with the finding forhydrologic, habitat and biological indicatorsover this narrow range of IC. Once again,watershed managers are advised to track otherwatershed indicators in the sensitive streamcategory, such as forest or crop cover.

1.2.2 Water Quality Response toStormwater Pollution

As noted in the previous section, most ICMresearch has been done on streams, which aredirectly influenced by increased stormwater.Many managers have wondered whether theICM also applies to downstream receivingwaters, such as lakes, water supply reservoirsand small estuaries. In general, the exact waterquality response of downstream receivingwaters to increased nonpoint source pollutantloads depends on many factors, including thespecific pollutant, the existing loading gener-ated by the converted land use, and the geom-etry and hydraulics of the receiving water.Table 3 indicates the sensitivity of rivers,lakes, estuaries, aquifers and water supplyreservoirs to various stormwater pollutants.

Lakes and the ICMThe water column and sediments of urbanlakes are impacted by many stormwaterpollutants, including sediment, nutrients,bacteria, metals, hydrocarbons, chlorides, andtrash/debris. Of these pollutants, limnologistshave always regarded phosphorus as theprimary lake management concern, given thatmore than 80% of urban lakes experiencesymptoms of eutrophication (CWP, 2001a).

In general, phosphorus export steadily in-creases as IC is added to a lake watershed,although the precise amount of IC that triggerseutrophication problems is unique to eachurban lake. With a little effort, it is possible tocalculate the specific IC threshold for anindividual lake, given its internal geometry, thesize of its contributing watershed, current in-lake phosphorus concentration, degree ofwatershed treatment, and the desired waterquality goals for the lake (CWP, 2001a). As ageneral rule, most lakes are extremely sensitive



to increases in phosphorus loads caused bywatershed IC. Exceptions include lakes that areunusually deep and/or have very small drain-age area/lake area ratios. In most lakes, how-ever, even a small amount of watersheddevelopment will result in an upward shift introphic status (CWP, 2001a).

Reservoirs and the ICMWhile surface water supply reservoirs respondto stormwater pollutant loads in the samegeneral manner as lakes, they are subject tostricter standards because of their uses fordrinking water. In particular, water supplyreservoirs are particularly sensitive to in-creased turbidity, pathogens, total organiccarbon, chlorides, metals, pesticides andhydrocarbon loads, in addition to phosphorus(Kitchell, 2001). While some pollutants can beremoved or reduced through expanded filteringand treatment at drinking water intakes, themost reliable approach is to protect the sourcewaters through watershed protection andtreatment.

Consequently, we often recommend that theICM be used as a “threat index” for mostdrinking water supplies. Quite simply, ifcurrent or future development is expected toexceed 10% IC in the contributing watershed,we recommend that a very aggressive water-shed protection strategy be implemented(Kitchell, 2001). In addition, we contend thatdrinking water quality cannot be sustainedonce watershed IC exceeds 25% and have yetto find an actual watershed where a drinkingwater utility has been maintained under theseconditions.

Small Tidal Estuaries and Coves and the ICMThe aquatic resources of small tidal estuaries,creeks, and coves are often highly impacted bywatershed development and associated activi-ties, such as boating/marinas, wastewaterdischarge, septic systems, alterations infreshwater flow and wetland degradation andloss. Given the unique impacts of eutrophica-tion on the marine system and stringent waterquality standards for shellfish harvesting, thestormwater pollutants of greatest concern inthe estuarine water column are nitrogen and

fecal coliform bacteria. Metals and hydrocar-bons in stormwater runoff can also contami-nate bottom sediments, which can prove toxicto local biota (Fortner et al., 1996; Fulton etal., 1996; Kucklick et al., 1997; Lerberg et al.,2000; Sanger et al., 1999; Vernberg et al.,1992).

While numerous studies have demonstratedthat physical, hydrologic, water quality andbiological indicators differ in urban and non-urban coastal watersheds, only a handful ofstudies have used watershed IC as an indicatorof estuarine health. These studies show signifi-cant correlations with IC, although degradationthresholds may not necessarily adhere to theICM due to tidal dilution and dispersion. Giventhe limited research, it is not fully clear if theICM can be applied to coastal systems withoutmodification.

Atmospheric deposition is considered aprimary source of nitrogen loading to estuarinewatersheds. Consequently, nitrogen loads inurban stormwater are often directly linked toIC. Total nitrogen loads have also been linkedto groundwater input, especially from subsur-face discharges from septic systems, which arecommon in low density coastal development(Swann, 2001; Valiela et al., 1997; Vernberg etal., 1996a). Nitrogen is generally considered tobe the limiting nutrient in estuarine systems,and increased loading has been shown toincrease algal and phytoplankton biomass andcause shifts in the phytoplankton communityand food web structure that may increase thepotential for phytoplankton blooms and fishkills (Bowen and Valiela, 2001; Evgenidou etal., 1997; Livingston, 1996).

Increased nitrogen loads have been linked todeclining seagrass communities, finfishpopulations, zooplankton reproduction, inver-tebrate species richness, and shellfish popula-tions (Bowen and Valiela, 2001; Rutkowski etal., 1999; Short and Wyllie-Echeverria, 1996;Valiela and Costa, 1988). Multiple studieshave shown significant increases in nitrogenloading as watershed land use becomes moreurban (Valiela et al., 1997; Vernberg et al.1996a; Wahl et al., 1997). While a few studies



link nitrogen loads with building and popula-tion density, no study was found that used ICas an indicator of estuarine nitrogen loading.

The second key water quality concern in smallestuaries is high fecal coliform levels instormwater runoff, which can lead to theclosure of shellfish beds and swimmingbeaches. Waterfowl and other wildlife havealso been shown to contribute to fecal coliformloading (Wieskel et al., 1996). Recent researchhas shown that fecal coliform standards areroutinely violated during storm events at verylow levels of IC in coastal watersheds (Mallinet al., 2001; Vernberg et al., 1996b; Schueler,1999). Maiolo and Tschetter (1981) found asignificant correlation between human popula-tion and closed shellfish acreage in NorthCarolina, and Duda and Cromartie (1982)found greater fecal coliform densities whenseptic tank density and IC increased, with anapproximate threshold at 10% watershed IC.

Recently, Mallin et al. (2000) studied fivesmall North Carolina estuaries of different landuses and showed that fecal coliform levelswere significantly correlated with watershedpopulation, developed land and IC. Percent ICwas the most statistically significant indicatorand could explain 95% of the variability infecal coliform concentrations. They also foundthat shellfish bed closures were possible inwatersheds with less than 10% IC, common inwatersheds above 10% IC, and almost certainin watersheds above 20% IC. While higherfecal coliform levels were observed in devel-oped watersheds, salinity, flushing and proxim-ity to pollution sources often resulted in higherconcentrations at upstream locations and athigh tides (Mallin et al., 1999). While thesestudies support the ICM, more research isneeded to prove the reliability of the ICM inpredicting shellfish bed closures based on IC.

Several studies have also investigated theimpacts of urbanization on estuarine fish,macrobenthos and shellfish communities.Increased PAH accumulation in oysters,negative effects of growth in juvenile sheeps-head minnows, reduced molting efficiency incopepods, and reduced numbers of grass

shrimp have all been reported for urbanestuaries as compared to forested estuaries(Fulton et al., 1996). Holland et al. (1997)reported that the greatest abundance of penaidshrimp and mummichogs was observed in tidalcreeks with forested watersheds compared tothose with urban cover. Porter et al. (1997)found lower grass shrimp abundance in smalltidal creeks adjacent to commercial and urbandevelopment, as compared to non-urbanwatersheds.

Lerberg et al. (2000) studied small tidal creeksand found that highly urban watersheds (50%IC) had the lowest benthic diversity andabundance as compared to suburban andforested creeks, and benthic communities werenumerically dominated by tolerant oligocha-etes and polychaetes. Suburban watersheds (15to 35% IC) also showed signs of degradationand had some pollution tolerant macrobenthos,though not as markedly as urban creeks.Percent abundance of pollution-indicativespecies showed a marked decline at 30% IC,and the abundance of pollution-sensitivespecies also significantly correlated with IC(Lerberg et al., 2000). Holland et al. (1997)reported that the variety and food availabilityfor juvenile fish species was impacted at 15 to20% IC.

Lastly, a limited amount of research hasfocused on the direct impact of stormwaterrunoff on salinity and hypoxia in small tidalcreeks. Blood and Smith (1996) comparedurban and forested watersheds and foundhigher salinities in urban watersheds due to theincreased number of impoundments. Fluctua-tions in salinity have been shown to affectshellfish and other aquatic populations (seeVernberg, 1996b). When urban and forestedwatersheds were compared, Lerberg et al.(2000) reported that higher salinity fluctuationsoccurred most often in developed watersheds;significant correlations with salinity range andIC were also determined. Lerberg et al. (2000)also found that the most severe and frequenthypoxia occurred in impacted salt marshcreeks and that dissolved oxygen dynamics intidal creeks were comparable to dead-endcanals common in residential marina-style



Practice N TSS TP OP TN NOx Cu Zn Oil/Grease11 Bacteria

Dry Ponds 9 47 19 N/R 25 3.5 26 26 3 44

Wet Ponds 43 80 51 65 33 43 57 66 78 70Wetlands 36 76 49 48 30 67 40 44 85 78Filtering Practices2 18 86 59 57 38 -14 49 88 84 37Water QualitySwales

9 81 34 1.0 84 31 51 71 62 -25

Ditches3 9 31 -16 N/R -9.0 24 14 0 N/R 0Infiltration 6 95 80 85 51 82 N/R N/R N/R N/R1: Represents data for Oil and Grease and PAH2: Excludes vertical sand filters3: Refers to open channel practices not designed for water qualityN/R = Not Reported

coastal developments. Suburban watersheds(15 to 35% IC) exhibited signs of degradationand had some pollution-tolerant macrobenthicspecies, though not to the extent of urbanwatersheds (50% IC).

In summary, recent research suggests thatindicators of coastal watershed health arelinked to IC. However, more research isneeded to clarify the relationship between ICand estuarine indicators in small tidal estuariesand high salinity creeks.

1.2.3 Effect of Watershed Treatmenton Stormwater Quality

Over the past two decades, many communitieshave invested in watershed protection prac-tices, such as stormwater treatment practices(STPs), stream buffers, and better site design,in order to reduce pollutant loads to receivingwaters. In this section, we review the effect ofwatershed treatment on the quality of stormwa-ter runoff.

Effect of Stormwater Treatment PracticesWe cannot directly answer the question as towhether or not stormwater treatment practicescan significantly reduce water quality impactsat the watershed level, simply because nocontrolled monitoring studies have yet beenconducted at this scale. Instead, we must relyon more indirect research that has tracked thechange in mass or concentration of pollutants

as they travel through individual stormwatertreatment practices. Thankfully, we have anabundance of these performance studies, withnearly 140 monitoring studies evaluating adiverse range of STPs, including ponds,wetlands, filters, and swales (Winer, 2000).

These studies have generally shown thatstormwater practices have at least a moderateability to remove many pollutants in urbanstormwater. Table 4 provides average removalefficiency rates for a range of practices andstormwater pollutants, and Table 5 profiles themean storm outflow concentrations for variouspractices. As can be seen, some groups ofpractices perform better than others in remov-ing certain stormwater pollutants. Conse-quently, managers need to carefully choosewhich practices to apply to solve the primarywater quality problems within their water-sheds.

It is also important to keep in mind that site-based removal rates cannot be extrapolated tothe watershed level without significant adjust-ment. Individual site practices are neverimplemented perfectly or consistently across awatershed. At least three discount factors needto be considered: bypassed load, treatabilityand loss of performance over time. For areview on how these discounts are derived,consult Schueler and Caraco (2001). Evenunder the most optimistic watershed imple-mentation scenarios, overall pollutant reduc-

Table 4: The Effectiveness of Stormwater Treatment Practices in RemovingPollutants - Percent Removal Rate (Winer, 2000)



tions by STPs may need to be discounted by atleast 30% to account for partial watershedtreatment.

Even with discounting, however, it is evidentthat STPs can achieve enough pollutantreduction to mimic rural background loads formany pollutants, as long as the watershed ICdoes not exceed 30 to 35%. This capability isillustrated in Figure 5, which shows phospho-rus load as a function of IC, with and withoutstormwater treatment.

Effect of Stream Buffers/Riparian AreasForested stream buffers are thought to havevery limited capability to remove stormwaterpollutants, although virtually no systematicmonitoring data exists to test this hypothesis.

The major reason cited for their limitedremoval capacity is that stormwater generatedfrom upland IC has usually concentratedbefore it reaches the forest buffer and thereforecrosses the buffer in a channel, ditch or stormdrain pipe. Consequently, the opportunity tofilter runoff is lost in many forest buffers inurban watersheds.

Effect of Better Site DesignBetter site design (BSD) is a term fornonstructural practices that minimize IC,conserve natural areas and distribute stormwa-ter treatment across individual developmentsites. BSD is also known by many othernames, including conservation development,low-impact development, green infrastructure,and sustainable urban drainage systems. While

Practice N TSS TP OP TN NOx Cu11 Zn11

Dry Ponds2 3 28 0.18 N/R 0.86 N/R 9.0 98Wet Ponds 25 17 0.11 0.03 1.3 0.26 5.0 30

Wetlands 19 22 0.20 0.07 1.7 0.36 7.0 31Filtering Practices3 8 11 0.10 0.07 1.1 0.55 9.7 21

Water Quality Swales 7 14 0.19 0.09 1.1 0.35 10 53Ditches4 3 29 0.31 N/R 2.4 0.72 18 32

1. Units for Zn and Cu are micrograms per liter (Fg/l)2. Data available for Dry Extended Detention Ponds only3. Excludes vertical sand filters4. Refers to open channel practices not designed for water qualityN/R = Not Reported

Table 5: Median Effluent Concentrations from Stormwater Treatment Practices (mg/l) (Winer, 2000)

Figure 5: Estimated Phosphorus Load as a Function of Impervious Cover, DiscountedStormwater Treatment and Better Site Design (Schueler and Caraco, 2001)

Impervious Cover (%)



some maintain that BSD is an alternative totraditional STPs, most consider it to be animportant complement to reduce pollutantloads.

While BSD has become popular in recentyears, only one controlled research study hasevaluated its potential performance, and this isnot yet complete (i.e. Jordan Cove, CT).

Indirect estimates of the potential value ofBSD to reduce pollutant discharges have beeninferred from modeling and redesign analyses(Zielinski, 2000). A typical example is pro-vided in Figure 5, which shows the presumedimpact of BSD in reducing phosphorus load-ings. As is apparent, BSD appears to be a veryeffective strategy in the one to 25% IC range,but its benefits diminish beyond that point.



1.3 Implications of the ICMfor Watershed Managers

One of the major policy implications of theICM is that in the absence of watershedtreatment, it predicts negative stream impactsat an extremely low intensity of watersheddevelopment. To put this in perspective,consider that a watershed zoned for two-acrelot residential development will generallyexceed 10% IC, and therefore shift from asensitive to an impacted stream classification(Cappiella and Brown, 2001). Thus, if acommunity wants to protect an important waterresource or a highly regarded species (such astrout, salmon or an endangered freshwatermussel), the ICM suggests that there is amaximum limit to growth that is not only quitelow, but is usually well below the currentzoning for many suburban or even ruralwatersheds. Consequently, the ICM suggeststhe unpleasant prospect that massive down-zoning, with all of the associated political andlegal carnage involving property rights andeconomic development, may be required tomaintain stream quality.

It is not surprising, then, that the ICM debatehas quickly shifted to the issue of whether ornot watershed treatment practices can provideadequate mitigation for IC. How much reliefcan be expected from stream buffers, stormwa-ter ponds, and other watershed practices, whichmight allow greater development densitywithin a given watershed? Only a limitedamount of research has addressed this question,and the early results are not reassuring (re-viewed in section 1.1.3). At this early stage,researchers are still having trouble detectingthe impact of watershed treatment, much lessdefining it. As noted earlier, both watershedresearch techniques and practice implementa-tion need to be greatly improved if we everexpect to get a scientifically defensible answerto this crucial question. Until then, managersshould be extremely cautious in setting highexpectations for how much watershed treat-ment can mitigate IC.

1.3.1 Management ofNon-Supporting Streams

Most researchers acknowledge that streamswith more than 25% IC in their watershedscannot support their designated uses or attainwater quality standards and are severelydegraded from a physical and biologicalstandpoint. As a consequence, many of thesestreams are listed for non-attainment under theClean Water Act and are subject to TotalMaximum Daily Load (TMDL) regulations.Communities that have streams within thisregulatory class must prepare implementationplans that demonstrate that water qualitystandards can ultimately be met.

While some communities have started torestore or rehabilitate these streams in recentyears, their efforts have yielded only modestimprovements in water quality and biologicalindicators. In particular, no community has yetdemonstrated that they can achieve waterquality standards in an urban watershed thatexceeds 25% IC. Many communities aredeeply concerned that non-supporting streamsmay never achieve water quality standards,despite massive investments in watershedrestoration. The ICM suggests that waterquality standards may need to be sharplyrevised for streams with more than 25% IC, ifthey are ever to come into attainment. Whilestates have authority to create more achievablestandards for non-supporting streams withinthe regulatory framework of the Clean WaterAct (Swietlik, 2001), no state has yet exercisedthis authority. At this time, we are not aware ofany water quality standards that are based onthe ICM or similar urban stream classificationtechniques.

Two political perceptions largely explain whystates are so reticent about revising waterquality standards. The first is a concern thatthey will run afoul of anti-degradation provi-sions within the Clean Water Act or be accusedof “backsliding” by the environmental commu-nity. The second concern relates to the demo-graphics of watershed organizations across thecountry. According to recent surveys, slightlymore than half of all watershed organizations



represent moderately to highly developedwatersheds (CWP, 2001a). These urbanwatershed organizations often have a keeninterest in keeping the existing regulatorystructure intact, since it is perceived to be theonly lever to motivate municipalities toimplement restoration efforts in non-support-ing streams.

However, revised water quality standards areurgently needed to support smart growthefforts. A key premise of smart growth is thatit is more desirable to locate new developmentwithin a non-supporting subwatershed ratherthan a sensitive or impacted one (i.e., concen-trating density and IC within an existingsubwatershed helps prevent sprawl fromencroaching on a less developed one). Yetwhile smart growth is desirable on a regionalbasis, it will usually contribute to alreadyserious problems in non-supporting water-sheds, which makes it even more difficult tomeet water quality standards.

This creates a tough choice for regulators: ifthey adopt stringent development criteria fornon-supporting watersheds, their added costscan quickly become a powerful barrier todesired redevelopment. If, on the other hand,they relax or waive environmental criteria,they contribute to the further degradation ofthe watershed. To address this problem, theCenter has developed a “smart watersheds”program to ensure that any localized degrada-tion caused by development within a non-supporting subwatershed is more than compen-sated for by improvements in stream qualityachieved through municipal restoration efforts(CWP, in press). Specifically, the smartwatersheds program includes 17 public sectorprograms to treat stormwater runoff, restoreurban stream corridors and reduce pollutiondischarges in highly urban watersheds. It ishoped that communities that adopt and imple-ment smart watershed programs will be givengreater flexibility to meet state and federalwater quality regulations and standards withinnon-supporting watersheds.

1.3.2 Use of the ICM for UrbanStream Classification

The ICM has proven to be a useful tool forclassifying and managing the large inventoryof streams that most communities possess. It isnot unusual for a typical county to have severalthousand miles of headwater streams within itspolitical boundaries, and the ICM provides aunified framework to identify and managethese subwatersheds. In our watershed practice,we use the ICM to make an initial diagnosisrather than a final determination for streamclassification. Where possible, we conductrapid stream and subwatershed assessments asa final check for an individual stream classifi-cation, particularly if it borders between thesensitive and impacted category. As notedearlier, the statistical variation in the IC/streamquality indicator makes it difficult to distin-guish between a stream with 9% versus 11%IC. Some of the key criteria we use to make afinal stream classification are provided inTable 6.

1.3.3 Role of the ICM in SmallWatershed Planning

The ICM has also proven to be an extremelyimportant tool for watershed planning, since itcan rapidly project how streams will change inresponse to future land use. We routinelyestimate existing and future IC in our water-shed planning practice and find that it is anexcellent indicator of change forsubwatersheds in the zero to 30% IC range. Inparticular, the ICM often forces watershedplanners to directly confront land use planningand land conservation issues early in theplanning process.

On the other hand, we often find that the ICMhas limited planning value whensubwatersheds exceed 30% IC for two practi-cal reasons. First, the ICM does not differenti-ate stream conditions within this very largespan of IC (i.e., there is no difference in thestream quality prediction for a subwatershedthat has 39.6% IC versus one that has 58.4%IC). Second, the key management question fornon-supporting watersheds is whether or not



they are potentially restorable. More detailedanalysis and field investigations are needed todetermine, in each subwatershed, the answer tothis question. While a knowledge of IC is oftenused in these feasibility assessments, it is butone of many factors that needs to be consid-ered.

Lastly, we have come to recognize severalpractical factors when applying the ICM forsmall watershed planning. These includethoughtful delineation of subwatershed bound-aries, the proper accounting of a direct drain-age area in larger watersheds, and the criticalneed for the most recent IC data. More guid-ance on these factors can be found in Zielinski(2001).

Stream Criteria

Reported presence of rare, threatened or endangered species in the aquaticcommunity (e.g., freshwater mussels, fish, crayfish or amphibians)Confirmed spawning of cold-water fish species (e.g., trout)Fair/good, good, or good to excellent macro invertebrate scoresMore than 65% of EPT species present in macro-invertebrate surveys No barriers impede movement of fish between the subwatershed and downstreamreceiving watersStream channels show little evidence of ditching, enclosure, tile drainage orchannelizationWater quality monitoring indicates no standards violations during dry weather Stream and flood plain remain connected and regularly interactStream drains to a downstream surface water supplyStream channels are generally stable, as determined by the Rosgen level analysisStream habitat scores are rated at least fair to good

Subwatershed Criteria

Contains terrestrial species that are documented as rare, threatened and endangeredWetlands, flood plains and/or beaver complexes make up more than 10% ofsubwatershed areaInventoried conservation areas comprise more than 10% of subwatershed areaMore than 50% of the riparian forest corridor has forest cover and is either publiclyowned or regulated Large contiguous forest tracts remain in the subwatershed (more than 40% in forestcover)Significant fraction of subwatershed is in public ownership and managementSubwatershed connected to the watershed through a wide corridorFarming, ranching and livestock operations in the subwatershed utilize bestmanagement practicesPrior development in the subwatershed has utilized stormwater treatment practices

Impervious cover is not a perfect indicator ofexisting stream quality. A number of streamand subwatershed criteria should be evaluatedin the field before a final classification deci-sion is made, particularly when the stream ison the borderline between two classifications.We routinely look at the stream andsubwatershed criteria to decide whether aborderline stream should be classified assensitive or impacted. Table 6 reviews theseadditional criteria.

Table 6: Additional Considerations for Urban Stream Classification



1.4 Summary

The remainder of this report presents greaterdetail on the individual research studies thatbear on the ICM. Chapter 2 profiles researchon hydrologic indicators in urban streams,while Chapter 3 summarizes the status ofcurrent research on the impact of urbanizationon physical habitat indicators. Chapter 4

presents a comprehensive review of the impactof urbanization on ten major stormwaterpollutants. Finally, Chapter 5 reviews thegrowing body of research on the link betweenIC and biological indicators within urbanstreams and wetlands.


Chapter 2: Hydrologic Impacts of Impervious Cover

Chapter 2: Hydrologic Impacts ofImpervious Cover

The natural hydrology of streams is fundamen-tally changed by increased watershed develop-ment. This chapter reviews the impacts ofwatershed development on selected indicatorsof stream hydrology.


2.1 Introduction2.2 Increased Runoff Volume2.3 Increased Peak Discharge Rates2.4 Increased Bankfull Flow2.5 Decreased Baseflow2.6 Conclusions

2.1 Introduction

Fundamental changes in urban stream hydrol-ogy occur as a result of three changes in theurban landscape that accompany land develop-ment. First, large areas of the watershed arepaved, rendering them impervious. Second,soils are compacted during construction, whichsignificantly reduces their infiltration capabili-ties. Lastly, urban stormwater drainage sys-

tems are installed that increase the efficiencywith which runoff is delivered to the stream(i.e., curbs and gutters, and storm drain pipes).Consequently, a greater fraction of annualrainfall is converted to surface runoff, runoffoccurs more quickly, and peak flows becomelarger. Additionally, dry weather flow instreams may actually decrease because lessgroundwater recharge is available. Figure 6illustrates the change in hydrology due toincreased urban runoff as compared to pre-development conditions.

Research has demonstrated that the effect ofwatershed urbanization on peak discharge ismore marked for smaller storm events. Inparticular, the bankfull, or channel formingflow, is increased in magnitude, frequency andduration. Increased bankfull flows have strongramifications for sediment transport andchannel enlargement. All of these changes inthe natural water balance have impacts on thephysical structure of streams, and ultimatelyaffect water quality and biological diversity.

Figure 6: Altered Hydrograph in Response to Urbanization(Schueler, 1987)



The relationship between watershed IC andstream hydrology is widely accepted, and hasbeen incorporated into many hydrologicengineering models over the past three de-cades. Several articles provide a good sum-mary of these (Bicknell et al., 1993; Hirsch etal., 1990; HEC, 1977; Huber and Dickinson,1988; McCuen and Moglen, 1988; Overton andMeadows, 1976; Pitt and Voorhees, 1989;Schueler, 1987; USDA, 1992; 1986).

The primary impacts of watershed develop-ment on stream hydrology are as follows:

• Increased runoff volume• Increased peak discharge rates• Increased magnitude, frequency, and

duration of bankfull flows• Diminished baseflow



2.2 Increased Runoff Volume

Impervious cover and other urban land usealterations, such as soil compaction and stormdrain construction, alter infiltration rates andincrease runoff velocities and the efficiencywith which water is delivered to streams. Thisdecrease in infiltration and basin lag time cansignificantly increase runoff volumes. Table 7reviews research on the impact of IC on runoffvolume in urban streams. Schueler (1987)demonstrated that runoff values are directlyrelated to subwatershed IC (Figure 7). Runoffdata was derived from 44 small catchmentareas across the country for EPA’s NationwideUrban Runoff Program.

Table 8 illustrates the difference in runoffvolume between a meadow and a parking lot,as compiled from engineering models. Theparking lot produces more than 15 times morerunoff than a meadow for the same stormevent.

Urban soils are also profoundly modifiedduring the construction process. The compac-tion of urban soils and the removal of topsoilcan decrease the infiltration capacity, causingincreases in runoff volumes (Schueler, 2000).Bulk density is often used to measure soilcompaction, and Table 9 illustrates how bulkdensity increases in many urban land uses.

Figure 7: Runoff Coefficient vs. IC (Schueler, 1987)

Note: 44 small urban catchments monitored during the national NURP study



Reference Key Finding Location

Increased Runoff Volume

Schueler,1987

Runoff coefficients were found to be strongly correlated with IC at 44 sitesnationwide. U.S.

Neller, 1988Urban watershed produced more than seven times as much runoff as asimilar rural watershed. Average time to produce runoff was reduced by 63%in urban watersheds compared to rural watersheds.

Australia

Increased Peak Discharge

Hollis, 1975

Review of data from several studies showed that floods with a return periodof a year or longer are not affected by a 5% watershed IC; small floods maybe increased 10 times by urbanization; flood with a return period of 100years may be doubled in size by a 30% watershed IC.

N/A

Leopold, 1968

Data from seven nationwide studies showed that 20% IC can cause themean annual flood to double. U.S.

Neller, 1988Average peak discharge from urban watersheds was 3.5 times higher thanpeak runoff from rural watersheds. Australia

Doll et al.,2000

Peak discharge was greater for 18 urban streams versus 11 rural Piedmontstreams. NC

Sauer et al.,1983

Estimates of flood discharge for various recurrence intervals showed that lessthan 50% watershed IC can result in a doubling of the 2-year, 10-year, and100-year floods.

U.S.

Leopold,1994

Watershed development over a 29-year period caused the peak dischargeof the 10-year storm to more than double. MD

Kibler et al.,1981

Rainfall/runoff model for two watersheds showed that an increase in ICcaused a significant increase in mean annual flood.

PA

Konrad andBooth, 2002

Evaluated streamflow data at 11 streams and found that the fraction ofannual mean discharges was exceeded and maximum annualinstantaneous discharges were related to watershed development androad density for moderately and highly developed watersheds.

WA

Table 7: Research Review of Increased Runoff Volume and PeakDischarge in Urban Streams



Hydrologic or Water Quality Parameter Parking Lot Meadow

Runoff Coefficient 0.95 0.06

Time of Concentration (minutes) 4.8 14.4

Peak Discharge, two-year, 24-hour storm (cfs) 4.3 0.4

Peak Discharge Rate, 100-year storm (cfs) 12.6 3.1

Runoff Volume from one-inch storm (cu. ft) 3,450 218

Runoff Velocity @ two-year storm (ft/sec) 8 1.8

Key Assumptions:

2-yr, 24-hr storm = 3.1 in; 100-yr storm = 8.9 in.Parking Lot: 100% imperviousness; 3% slope; 200ft flow length; hydraulic radius =.03; concrete channel;suburban Washington C valuesMeadow: 1% impervious; 3% slope; 200 ft flow length; good vegetative condition; B soils; earthenchannel Source: Schueler, 1994a

Table 8: Hydrologic Differences Between a Parking Lot and a Meadow(Schueler, 1994a)

Undisturbed SoilType or Urban

Condition

Surface BulkDensity

(grams/cubiccentimeter)

Urban Condition Surface Bulk Density

(grams/cubiccentimeter)

Peat 0.2 to 0.3 Urban Lawns 1.5 to 1.9

Compost 1.0Crushed RockParking Lot

1.5 to 1.9

Sandy Soils 1.1 to 1.3 Urban Fill Soils 1.8 to 2.0

Silty Sands 1.4 Athletic Fields 1.8 to 2.0

Silt 1.3 to 1.4 Rights-of-Way andBuilding Pads (85%)

1.5 to 1.8

Silt Loams 1.2 to 1.5Rights-of-Way andBuilding Pads (95%)

1.6 to 2.1

Organic Silts/Clays 1.0 to 1.2 Concrete

Pavement 2.2

Glacial Till 1.6 to 2.0 Rock 2.65

Table 9: Comparison of Bulk Density for Undisturbed Soils andCommon Urban Conditions (Schueler, 2000)



2.3 Increased PeakDischarge Rate

Watershed development has a strong influenceon the magnitude and frequency of flooding inurban streams. Peak discharge rates are oftenused to define flooding risk. Doll et al. (2000)compared 18 urban streams with 11 ruralstreams in the North Carolina Piedmont andfound that unit area peak discharge was alwaysgreater in urban streams (Figure 8). Data fromSeneca Creek, Maryland also suggest a similarincrease in peak discharge. The watershedexperienced significant growth during the1950s and 1960s. Comparison of pre- and post-development gage records suggests that thepeak 10-year flow event more than doubledover that time (Leopold, 1994).

Hollis (1975) reviewed numerous studies onthe effects of urbanization on floods of differ-ent recurrence intervals and found that theeffect of urbanization diminishes when floodrecurrence gets longer (i.e., 50 and 100 years).Figure 9 shows the effect on flood magnitudein urban watersheds with 30% IC, and shows

the one-year peak discharge rate increasing bya factor of 10, compared to an undevelopedwatershed. In contrast, floods with a 100-yearrecurrence interval only double in size underthe same watershed conditions.

Sauer et al. (1983) evaluated the magnitude offlooding in urban watersheds throughout theUnited States. An equation was developed forestimating discharge for floods of two-year,10-year, and 100-year recurrence intervals. Theequations used IC to account for increasedrunoff volume and a basin development factorto account for sewers, curbs and gutters,channel improvements and drainage develop-ment. Sauer noted that IC is not the dominantfactor in determining peak discharge rates forextreme floods because these storm eventssaturate the soils of undeveloped watershedsand produce high peak discharge rates. Sauerfound that watersheds with 50% IC can in-crease peak discharge for the two-year flood bya factor of four, the 10-year flood by a factor ofthree, and the 100-year flood by a factor of 2.5,depending on the basin development factor(Figure 10).

Figure 8: Peak Discharge for Urban and Rural Streams in North Carolina (Doll et al., 2000)



2.4 Increased Bankfull Flow

Urbanization also increases the frequency andduration of peak discharge associated withsmaller flood events (i.e., one- to two-yearreturn storms). In terms of stream channelmorphology, these more frequent bankfullflows are actually much more important thanlarge flood events in forming the channel. Infact, Hollis (1975) demonstrated that urbaniza-tion increased the frequency and magnitude ofbankfull flow events to a greater degree thanthe larger flood events.

Figure 10: Relationship of Urban/Rural 100-Year Peak Flow Ratio to BasinDevelopment Factor and IC (Sauer et al., 1983)

Figure 9: Effect on Flood Magnitudes of 30% Basin IC (Hollis, 1975)

An example of the increase in bankfull flow inarid regions is presented by the U.S. Geologi-cal Survey (1996), which compared the peakdischarge rate from two-year storm eventsbefore and after watersheds urbanized in ParrisValley, California. Over an approximately 20-year period, watershed IC increased by 13.5%,which caused the two-year peak flow to morethan double. Table 10 reviews other researchstudies on the relationship between watershedIC and bankfull flows in urban streams.



Leopold (1968) evaluated data from sevennationwide studies and extrapolated this data toillustrate the increase in bankfull flows due tourbanization. Figure 11 summarizes therelationship between bankfull flows over a

range of watershed IC. For example, water-sheds that have 20% IC increase the number offlows equal to or greater than bankfull flow bya factor of two. Leopold (1994) also observed adramatic increase in the frequency of thebankfull event in Watts Branch, an urbansubwatershed in Rockville, Maryland. Thiswatershed experienced significant urbandevelopment during the 1950s and 1960s.Leopold compared gage records and found thatthe bankfull storm event frequency increasedfrom two to seven times per year from 1958 to1987.

More recent data on bankfull flow frequencywas reported for the Rouge River near Detroit,Michigan by Fongers and Fulcher (2001). Theynoted that channel-forming flow (1200 cfs)was exceeded more frequently as urbanizationincreased in the watershed and had becomethree times more frequent between 1930 and1990 (Figure 12).

McCuen and Moglen (1988) have documentedthe increase in duration of bankfull flows inresponse to urbanization using hydrologymodels. MacRae (1996), monitored a stream inMarkham, Ontario downstream of a stormwa-ter pond and found that the hours of


Booth andReinelt, 1993

Using a simulation model and hydrologic data from four watersheds, itwas estimated that more than 10% watershed IC may cause dischargefrom the two-year storm under current conditions to equal or exceeddischarge from the 10-year storm under forested conditions.

WA

Fongers andFulcher, 2001

Bankfull flow of 1200 cfs was exceeded more frequently over time withurbanization, and exceedence was three times as frequent from 1930s to1990s.

MI

USGS,1996

Over a 20-year period, IC increased 13.5%, and the two-year peak flowmore than doubled in a semi-arid watershed.

CA

Henshaw andBooth,2000

Two of three watersheds in the Puget Sound lowlands showed increasingflashiness over 50 years with urbanization.

WA

Leopold, 1968Using hydrologic data from a nine-year period for North BranchBrandywine Creek, it was estimated that for a 50% IC watershed, bankfullfrequency would be increased fourfold.

PA

Leopold,1994

Bankfull frequency increased two to seven times after urbanization inWatts Branch.

MD

MacRae,1996

For a site downstream of a stormwater pond in Markham, Ontario hoursof exceedence of bankfull flows increased by 4.2 times after thewatershed urbanized (34% IC)

Ontario

Figure 11: Increase in Bankfull Flows Due toUrbanization (Leopold, 1968)

Table 10: Research Review of Increased Bankfull Discharge in Urban Streams



Figure 12: Increase in Number of Exceedences of Bankfull Flow Over TimeWith Urbanization in the Rouge River, MT (Fongers and Fulcher, 2001)

exceedence of bankfull flows increased by afactor of 4.2 once watershed IC exceeded 30%.Modeling for seven streams also downstreamof stormwater ponds in Surrey, British Colum-bia also indicated an increase in bankfullflooding in response to watershed development(MacRae, 1996).

Watershed IC also increases the “flashiness” ofstream hydrographs. Flashiness is defined here

Figure 13: Percent of Gage Reading Above Mean Annual Flow for Puget SoundLowland Streams (Henshaw and Booth, 2000)

as the percent of daily flows each year thatexceeds the mean annual flow. Henshaw andBooth (2000) evaluated seven urbanizedwatersheds in the Puget Sound lowlandstreams and tracked changes in flashiness over50 years (Figure 13). The most urbanizedwatersheds experienced flashy discharges.Henshaw and Booth concluded that increasedrunoff in urban watersheds leads to higher butshorter-duration peak discharges.

River Rouge - Number of Exceedances of 1200 cfs

Decade



Reference Key Finding LocationFinkenbine et al.,

2000Summer base flow was uniformly low in 11 streams when ICreached 40% or greater.

Vancouver

Klein, 1979 Baseflow decreased as IC increased in Piedmont streams. MD

Saravanapavan, 2002

Percentage of baseflow decreased linearly as IC increased for 13subwatersheds of Shawsheen River watershed. MA

Simmons andReynolds, 1982

Dry weather flow dropped 20 to 85% after development inseveral urban watersheds on Long Island.

NY

Spinello andSimmons, 1992

Baseflow in two Long Island streams went dry as a result ofurbanization. NY

Konrad and Booth,2002

No discernable trend over many decades in the annual sevenday low flow discharge for 11 Washington streams.

WA

Wang et al., 2001Stream baseflow was negatively correlated with watershed IC in47 small streams, with an apparent breakpoint at 8 to 12% IC.

WI

Evett et al., 1994 No clear relationship between dry weather flow and urban andrural streams in 21 larger watersheds.

NC

2.5 Decreased Baseflow

As IC increases in a watershed, less groundwa-ter infiltration is expected, which can poten-tially decrease stream flow during dry periods,(i.e. baseflow). Several East Coast studiesprovide support for a decrease in baseflow as aresult of watershed development. Table 11reviews eight research studies on baseflow inurban streams.

Klein (1979) measured baseflow in 27 smallwatersheds in the Maryland Piedmont andreported an inverse relationship between ICand baseflow (Figure 14). Spinello andSimmons (1992) demonstrated that baseflow intwo urban Long Island streams declinedseasonally as a result of urbanization (Figure15). Saravanapavan (2002) also found thatpercentage of baseflow decreased in directproportion to percent IC for 13 subwatershedsof the Shawsheen River watershed in Massa-chusetts (Figure 16).

Table 11: Research Review of Decreased Baseflow in Urban Streams

Figure 14: Relationship BetweenBaseflow and Watershed IC in theStreams on Maryland Piedmont

(Klein, 1979)



Figure 15: Baseflow Response to Urbanization in Long Island Streams(Spinello and Simmons, 1992)

Figure 16: Relationship Between Percentage Baseflow and Percent IC inMassachusetts Streams (Saravanapan, 2002)



Finkebine et al. (2000) monitored summerbaseflow in 11 streams near Vancouver, BritishColumbia and found that stream base flow wasuniformly low due to decreased groundwaterrecharge in watersheds with more than 40% IC(Figure 17). Baseflow velocity also consis-tently decreased when IC increased (Figure18). The study cautioned that other factors canaffect stream baseflow, such as watershedgeology and age of development.

Other studies, however, have not been able toestablish a relationship between IC and declin-ing baseflow. For example, a study in NorthCarolina could not conclusively determine thaturbanization reduced baseflow in larger urbanand suburban watersheds in that area (Evett et

al., 1994). In some cases, stream baseflow issupported by deeper aquifers or originate inareas outside the surface watershed boundary.In others, baseflow is augmented by leakingsewers, water pipes and irrigation return flows.

This appears to be particularly true in arid andsemi-arid areas, where baseflow can actuallyincrease in response to greater IC (Hollis,1975). For instance, Crippen and Waananen(1969) found that Sharon Creek near SanFrancisco changed from an ephemeral streaminto a perennial stream after urban develop-ment. Increased infiltration from lawn wateringand return flow from sewage treatment plantsare two common sources of augmentedbaseflows in these regions (Caraco, 2000a).

Figure 18: Effect of Watershed IC on SummerStream Velocity in Vancouver Streams (Finkerbine et al., 2000)

Figure 17: Effect of IC on Summer Baseflowin Vancouver Streams (Finkerbine et al., 2000)



2.6 Conclusions

The changes in hydrology indicators caused bywatershed urbanization include increasedrunoff volume; increased peak discharge;increased magnitude, frequency and durationof bankfull flows; flashier/less predictableflows; and decreased baseflow. Many studiessupport the direct relationship between IC andthese indicators. However, at low levels ofwatershed IC, site-specific factors such asslope, soils, types of conveyance systems, ageof development, and watershed dimensionsoften play a stronger role in determining awatershed’s hydrologic response.

Overall, the following conclusions can bedrawn from the relationship between watershedIC and hydrology indicators:

• Strong evidence exists for the directrelationship between watershed IC andincreased stormwater runoff volume andpeak discharge. These relationships areconsidered so strong that they have beenincorporated into widely accepted engi-neering models.

• The relationship between IC and bankfullflow frequency has not been extensivelydocumented, although abundant data existsfor differences between urban and non-urban watersheds.

• The relationship between IC and decliningstream flow is more ambiguous andappears to vary regionally in response toclimate and geologic factors, as well aswater and sewer infrastructure.

The changes in hydrology indicators caused bywatershed urbanization directly influencephysical and habitat characteristics of streams.The next chapter reviews how urban streamsphysically respond to the major changes totheir hydrology.




Chapter 3: Physical Impacts of Impervious Cover

Chapter 3: Physical Impacts ofImpervious Cover

A growing body of scientific literature docu-ments the physical changes that occur instreams undergoing watershed urbanization.This chapter discusses the impact of watersheddevelopment on various measures of physicalhabitat in urban stream channels and is orga-nized as follows:

3.1 Difficulty in Measuring Habitat3.2 Changes in Channel Geometry3.3 Effect on Composite Indexes of

Stream Habitat3.4 Effect on Individual Elements of

Stream Habitat3.5 Increased Stream Warming3.6 Alteration of Stream Channel Network3.7 Conclusion

This chapter reviews the available evidence onstream habitat. We begin by looking at geo-morphological research that has examined howthe geometry of streams changes in response toaltered urban hydrology. The typical responseis an enlargement of the cross-sectional area ofthe stream channel through a process ofchannel incision, widening, or a combinationof both. This process triggers an increase inbank and/or bed erosion that increases sedi-ment transport from the stream, possibly forseveral decades or more.

Next, we examine the handful of studies thathave evaluated the relationship betweenwatershed development and composite indica-tors of stream habitat (such as the habitatRapid Bioassessment Protocol, or RBP). In thefourth section, we examine the dozen studiesthat have evaluated how individual habitatelements respond to watershed development.These studies show a consistent picture.Generally, streams with low levels of IC havestable banks, contain considerable large woodydebris (LWD) and possess complex habitatstructure. As watershed IC increases, however,urban streambanks become increasinglyunstable, streams lose LWD, and they developa more simple and uniform habitat structure.This is typified by reduced pool depths, loss ofpool and riffle sequences, reduced channelroughness and less channel sinuosity.

Water temperature is often regarded as a keyhabitat element, and the fifth section describesthe stream warming effect observed in urbanstreams in six studies. The last section looks atthe effect of watershed development on thestream channel network as a whole, in regardto headwater stream loss and the creation offish barriers.



3.1 Difficulty in MeasuringHabitat

The physical transformation of urban streamsis perhaps the most conspicuous impact ofwatershed development. These dramaticphysical changes are easily documented insequences of stream photos with progressivelygreater watershed IC (see Figure 19). Indeed,the network of headwater stream channelsgenerally disappears when watershed ICexceeds 60% (CWP).

3.1.1 The Habitat Problem

It is interesting to note that while the physicalimpacts of urbanization on streams are widelyaccepted, they have rarely been documented bythe research community. As a consequence, nopredictive models exist to quantify howphysical indicators of stream habitat willdecline in response to watershed IC, despitethe fact that most would agree that some kindof decline is expected (see Table 12).

Figure 19: Urban Stream Channels with Progressively Greater IC

10% IC 28% IC

31% IC 40% IC

53% IC 55% IC



The main reason for this gap is that “habitat” isextremely hard to define, and even moredifficult to measure in the field. Most indicesof physical habitat involve a visual and qualita-tive assessment of 10 or more individualhabitat elements that are perceived by fisheryand stream biologists to contribute to qualitystream habitat. Since these indices includemany different habitat elements, each of whichis given equal weight, they have not been veryuseful in discriminating watershed effects(Wang et al., 2001).

Researchers have had greater success inrelating individual habitat elements to water-shed conditions, such as large woody debris(LWD), embeddedness, or bank stability. Evenso, direct testing has been limited, partlybecause individual habitat elements are hard tomeasure and are notoriously variable in bothspace and time. Consider bank stability for amoment. It would be quite surprising to see ahighly urban stream that did not have unstablebanks. Yet, the hard question is exactly howwould bank instability be quantitativelymeasured? Where would it be measured — at apoint, a cross-section, along a reach, on the leftbank or the right?

Geomorphologists stress that no two streamreaches are exactly alike, due to differences ingradient, bed material, sediment transport,hydrology, watershed history and many otherfactors. Consequently, it is difficult to makecontrolled comparisons among differentstreams. Indeed, geomorphic theory stressesthat individual stream reaches respond in a

highly dynamic way to changes in watershedhydrology and sediment transport, and can takeseveral decades to fully adjust to a new equi-librium.

Returning to our example of defining bankstability, how might our measure of bankinstability change over time as its watershedgradually urbanizes, is built out, and possiblyreaches a new equilibrium over several de-cades? It is not very surprising that the effectof watershed development on stream habitat iswidely observed, yet rarely measured.

Specific Impacts

Sediment transport modifiedChannel enlargementChannel incisionStream embeddednessLoss of large woody debrisChanges in pool/riffle structureLoss of riparian coverReduced channel sinuosityWarmer in-stream temperatures Loss of cold water species anddiversityChannel hardeningFish blockagesLoss of 1st and 2nd order streamsthrough storm drain enclosure

Table 12: Physical Impacts ofUrbanization on Streams



3.2 Changes in StreamGeometry

As noted in the last chapter, urbanizationcauses an increase in the frequency andduration of bankfull and sub-bankfull flowevents in streams. These flow events performmore “effective work” on the stream channel,as defined by Leopold (1994). The net effect isthat an urban stream channel is exposed tomore shear stress above the critical thresholdneeded to move bank and bed sediments(Figure 20). This usually triggers a cycle ofactive bank erosion and greater sedimenttransport in urban streams. As a consequence,the stream channel adjusts by expanding itscross-sectional area, in order to effectivelyaccommodate greater flows and sedimentsupply. The stream channel can expand byincision, widening, or both. Incision refers tostream down-cutting through the streambed,whereas widening refers to lateral erosion of

the stream bank and its flood plain (Allen andNarramore, 1985; Booth, 1990; Morisawa andLaFlure, 1979).

3.2.1 Channel Enlargement

A handful of research studies have specificallyexamined the relationship between watersheddevelopment and stream channel enlargement(Table 13). These studies indicate that streamcross-sectional areas can enlarge by as much astwo to eight times in response to urbanization,although the process is complex and may takeseveral decades to complete (Pizzuto et al.,2000; Caraco, 2000b; Hammer, 1972). Anexample of channel enlargement is provided inFigure 21, which shows how a stream cross-section in Watts Branch near Rockville,Maryland has expanded in response to nearlyfive decades of urbanization (i.e., watershed ICincreased from two to 27%).

Figure 20: Increased Shear Stress from a Hydrograph(MacRae and Rowney, 1992)




% IC used as Indicator

Caraco, 2000b

Reported enlargement in ratios of 1.5 to 2.2 for 10 stream reachesin Watts Branch and computed ultimate enlargement ratios of 2.0 MD

MacCraeand De

Andrea, 1999

Introduced the concept of ultimate channel enlargement basedon watershed IC and channel characteristics.

Ontario,TX

Morse, 2001 Demonstrated increased erosion rates with increases in IC(channels were generally of the same geomorphic type).

ME

Urbanization Used as Indicator

Allen andNarramore,

1985Enlargement ratios in two urban streams ranged from 1.7 to 2.4. TX

Bledsoe, 2001Reported that channel response to urbanization depends onother factors in addition to watershed IC including geology,vegetation, sediment and flow regimes.

N/A

Booth andHenshaw,

2001

Evaluated channel cross section erosion rates and determinedthat these rates vary based on additional factors including theunderlying geology, age of development and gradient.

WA

Hammer, 1972 Enlargement ratios ranged from 0.7 to 3.8 in urban watersheds. PA

Neller, 1989Enlargement ratios in small urban catchments ranged from two to7.19, the higher enlargement ratios were primarily from incisionoccurring in small channels.

Australia

Pizzuto et al., 2000

Evaluated channel characteristics of paired urban and ruralstreams and demonstrated median bankfull cross sectionalincrease of 180%. Median values for channel sinuosity were 8%lower in urban streams; Mannings N values were found to be 10%lower in urban streams.

PA

Hession et al.,in press

Bankfull widths for urban streams were significantly wider thannon-urban streams in 26 paired streams. Forested reaches wereconsistently wider than non-forested reaches in urban streams.

MD, DE,PA

Dartiguenaveet al., 1997

Bank erosion accounted for up to 75% of the sediment transportin urban watersheds. TX

Trimble, 1997Demonstrated channel enlargement over time in an urbanizingSan Diego Creek; Bank erosion accounted for over 66% of thesediment transport.

CA

Table 13: Research Review of Channel Enlargement and SedimentTransport in Urban Streams



Some geomorphologists suggest that urbanstream channels will reach an “ultimateenlargement” relative to pre-developed chan-nels (MacRae and DeAndrea, 1999) and thatthis can be predicted based on watershed IC,age of development, and the resistance of thechannel bed and banks. A relationship betweenultimate stream channel enlargement andwatershed IC has been developed for alluvialstreams in Texas, Vermont and Maryland(Figure 22). Other geomorphologists such asBledsoe (2001) and Booth and Henshaw(2001) contend that channel response tourbanization is more complex, and also de-pends on geology, grade control, streamgradient and other factors.

Channel incision is often limited by gradecontrol caused by bedrock, cobbles, armoredsubstrates, bridges, culverts and pipelines.These features can impede the downwarderosion of the stream channel and thereby limitthe incision process. Stream incision canbecome severe in streams that have softersubstrates such as sand, gravel and clay(Booth, 1990). For example, Allen andNarramore (1985) showed that channel en-largement in chalk channels was 12 to 67%greater than in shale channels near Dallas,

Texas. They attributed the differences to thesofter substrate, greater velocities and highershear stress in the chalk channels.

Neller (1989) and Booth and Henshaw (2001)also report that incised urban stream channelspossess cross-sectional areas that are largerthan would be predicted based on watershedarea or discharge alone. This is due to the factthat larger floods are often contained withinthe stream channel rather than the floodplain.Thus, incised channels often result in greatererosion and geomorphic change. In general,stream conditions that can foster incisioninclude erodible substrates, moderate to highstream gradients, and an absence of gradecontrol features.

Channel widening occurs more frequentlywhen streams have grade control and thestream has cut into its bank, thereby expandingits cross-sectional area. Urban stream channelsoften have artificial grade controls caused byfrequent culverts and road crossings. Thesegrade controls often cause localized sedimentdeposition that can reduce the capacity ofculverts and bridge crossings to pass floodwaters.

0

1

2

3

4

5

6

7

8

9

10

0 5 10 15 20 25 30 35 40 45

Cross Section Stations (ft) - Looking Downstream

Ele

vatio

n (f

t-m

sl)

Historic Section

Current Section

Bankfull Depth

Ultimate Section ?

Historic cross-section

Current cross-section

Ultimate cross-section ?

Figure 21: Stream Channel Enlargement in Watts Branch, MD 1950-2000 (Caraco, 2000b)



The loss of flood plain and riparian vegetationhas been strongly associated with watershedurbanization (May et al., 1997). A few studieshave shown that the loss of riparian trees canresult in increased erosion and channel migra-tion rates (Beeson and Doyle, 1995 andAllmendinger et al., 1999). For example,Beeson and Doyle (1995) found that meanderbends with vegetation were five times lesslikely to experience significant erosion from amajor flood than non-vegetated meanderbends.Hession et al. (in press) observed thatforested reaches consistently had greaterbankfull widths than non-forested reaches in aseries of urban streams in Pennsylvania,Maryland and Delaware.

3.2.2 Effect of Channel Enlargementon Sediment Yield

Regardless of whether a stream incises,widens, or does both, it will greatly increasesediment transport from the watershed due toerosion. Urban stream research conducted inCalifornia and Texas suggests that 60 to 75%of the sediment yield of urban watersheds canbe derived from channel erosion (Trimble,1997 and Dartingunave et al., 1997) This canbe compared to estimates for rural streams

where channel erosion accounts for only five to20% of the annual sediment yield (Collins etal., 1997 and Walling and Woodward, 1995).

Some geomorphologists speculate that urbanstream channels will ultimately adjust to theirpost-development flow regime and sedimentsupply. Finkenbine et al. (2000) observed theseconditions in Vancouver streams, where studystreams eventually stabilized two decades afterthe watersheds were fully developed. In olderurban streams, reduced sediment transport canbe expected when urbanization has beencompleted. At this point, headwater streamchannels are replaced by storm drains andpipes, which can transport less sediment. Thelack of available sediment may cause down-stream channel erosion, due to the diminishedsediment supply found in the stream.

Figure 22: Ultimate Channel Enlargement in MD, UT and TX Alluvial Streams(MacRae and DeAndrea, 1999 and CWP, 2001b)

0.00

2.00

4.00

6.00

8.00

10.00

12.00

14.00

0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0

Imperviousness (%)

En

larg

emen

t R

atio



3.3 Effect on CompositeMeasures of Stream Habitat

Composite measures of stream habitat refer toassessments such as EPA’s Habitat RapidBioassessment Protocol (RBP) that combinemultiple habitat elements into a single score orindex (Barbour et al., 1999). For example, theRBP requires visual assessment of 10 streamhabitat elements, including embeddedness,epifaunal substrate quality, velocity/depthregime, sediment deposition, channel flowstatus, riffle frequency, bank stabilization,streambank vegetation and riparian vegetationwidth. Each habitat element is qualitativelyscored on a 20 point scale, and each element isweighted equally to derive a composite scorefor the stream reach.

To date, several studies have found a relation-ship between declining composite habitatindicator scores and increasing watershed IC indifferent eco-regions of the United States. A

typical pattern in the composite habitat scoresis provided for headwater streams in Maine(Morse, 2001; Figure 23). This general findinghas been reported in the mid-Atlantic, North-east and the Northwest (Black and Veatch,1994; Booth and Jackson, 1997; Hicks andLarson, 1997; Maxted and Shaver, 1997;Morse, 2001; Stranko and Rodney, 2001).

However, other researchers have found a muchweaker relationship between composite habitatscores and watershed IC. Wang and his col-leagues (2001) found that composite habitatscores were not correlated with watershed ICin Wisconsin streams, although it was corre-lated with individual habitat elements, such asstreambank erosion. They noted that manyagricultural and rural streams had fair to poorcomposite habitat scores, due to poor riparianmanagement and sediment deposition. Thesame basic conclusion was also reported forstreams of the Maryland Piedmont (MNCPPC,2000).

Figure 23: Relationship Between Habitat Quality and IC in Maine Streams (Morse, 2001)



3.4 Effect on IndividualElements of Stream Habitat

Roughly a dozen studies have examined theeffect of watershed development on thedegradation of individual stream habitatfeatures such as bank stability, embeddedness,riffle/pool quality, and loss of LWD (Table14). Much of this data has been acquired fromthe Pacific Northwest, where the importance ofsuch habitat for migrating salmon has been apersistent management concern.

3.4.1 Bank Erosion andBank Stability

It is somewhat surprising that we could onlyfind one study that related bank stability orbank erosion to watershed IC. Conducted byBooth (1991) in the streams of the PugetSound lowlands, the study reported that streambanks were consistently rated as stable inwatersheds with less than 10% IC, but becameprogressively more unstable above this thresh-old. Dozens of stream assessments have foundhigh rates of bank erosion in urban streams, butnone, to our knowledge, has systematicallyrelated the prevalence or severity of bankerosion to watershed IC. As noted earlier, this

may reflect the lack of a universally recog-nized method to measure comparative bankerosion in the field.

3.4.2 Embeddedness

Embeddedness is a term that describes theextent to which the rock surfaces found on thestream bottom are filled in with sand, silts andclay. In a healthy stream, the interstitial poresbetween cobbles, rock and gravel generallylack fine sediments, and are an active habitatzone and detrital processing area. The in-creased sediment transport in urban streamscan rapidly fill up these pores in a processknown as embedding. Normally,embeddedness is visually measured in rifflezones of streams. Riffles tend to be an impor-tant habitat for aquatic insects and fish (such asdarters and sculpins). Clean stream substratesare also critical to trout and salmon eggincubation and embryo development. May etal. (1997) demonstrated that the percent of finesediment particles in riffles generally increasedwith watershed IC (Figure 24). However,Finkenbine et al. (2000) reported thatembeddedness eventually decreased slightlyafter watershed land use and sediment trans-port had stabilized for 20 years.

Figure 24: Fine Material Sediment Deposition as a Function of IC in PacificNorthwest Streams (Horner et al., 1997)




% IC Used as Indicator

Black & Veatch,1994

Habitat scores were ranked as poor in five subwatersheds that hadgreater than 30% IC.

MD

Booth andJackson, 1997

Increase in degraded habitat conditions with increases in watershed IC. WA

Hicks and Larson, 1997

Reported a reduction in composite stream habitat indices with increasingwatershed IC.

MA

May et al., 1997Composite stream habitat declined most rapidly during the initial phase ofthe watershed urbanization, when percent IC exceeded the 5-10% range.

WA

Stranko andRodney, 2001

Composite index of stream habitat declined with increasing watershed ICin coastal plain streams. MD

Wang et al., 2001Composite stream habitat scores were not correlated with watershed IC in47 small watersheds, although channel erosion was. Non-urban watershedswere highly agricultural and often lacked riparian forest buffers.

WI

MNCPPC, 2000Reported that stream habitat scores were not correlated with IC insuburban watersheds. MD

Morse, 2001 Composite habitat values tended to decline with increases in watershedIC.

ME

Booth, 1991Channel stability and fish habitat quality declined rapidly after 10%watershed IC.

WA

Booth et al., 1997 Decreased LWD with increased IC. PNW

Finkenbine et al.,2000

LWD was scarce in streams with greater than 20% IC in Vancouver. B.C.

Horner & May, 1999When IC levels were >5%, average LWD densities fell below 300pieces/kilometer.

PNW

Horner et al., 1997Interstitial spaces in streambed sediments begin to fill with increasingwatershed IC. PNW


Dunne andLeopold, 1978

Natural channels replaced by storm drains and pipes; increased erosionrates observed downstream. MD

May et al., 1997 Forested riparian corridor width declines with increased watershed IC. PNW

MWCOG, 1992 Fish blockages caused by bridges and culverts noted in urban watersheds. D.C.

Pizzuto et al., 2000Urban streams had reduced pool depth, roughness, and sinuosity,compared to rural streams; Pools were 31% shallower in urban streamscompared to non-urban ones.

PA

Richey, 1982 Altered pool/riffle sequence observed in urban streams. WA

Scott et al., 1986 Loss of habitat diversity noted in urban watersheds. PNW

Spence et al., 1996 Large woody debris is important for habitat diversity and anadromous fish. PNW

Table 14: Research Review of Changes in Urban Stream Habitat



3.4.3 Large Woody Debris (LWD)

LWD is a habitat element that describes theapproximate volume of large woody material(< four inches in diameter) found in contactwith the stream. The presence and stability ofLWD is an important habitat parameter instreams. LWD can form dams and pools, trapsediment and detritus, stabilize stream chan-nels, dissipate flow energy, and promotehabitat complexity (Booth et al., 1997). LWDcreates a variety of pool features (plunge,lateral, scour and backwater); short riffles;undercut banks; side channels; and a range ofwater depths (Spence et al., 1996). Urbanstreams tend to have a low supply of LWD, asincreased stormwater flows transport LWD andclears riparian areas. Horner et al. (1997)presents evidence from Pacific Northweststreams that LWD decreases in response toincreasing watershed IC (Figure 25).

3.4.4 Changes in Other IndividualStream Parameters

One of the notable changes in urban streamhabitat is a decrease in pool depth and ageneral simplification of habitat features suchas pools, riffles and runs. For example, Richey(1982) and Scott et al. (1986) reported anincrease in the prevalence of glides and acorresponding altered riffle/pool sequence dueto urbanization. Pizzuto et al. (2000) reported amedian 31% decrease in pool depth in urbanstreams when compared to forested streams.Pizzuto et al. also reported a modest decreasein channel sinuosity and channel roughness inthe same urban streams in Pennsylvania.

Several individual stream habitat parametersappear to have received no attention in urbanstream research to date. These parametersinclude riparian shading, wetted perimeter,various measures of velocity/depth regimes,riffle frequency, and sediment deposition inpools. More systematic monitoring of theseindividual stream habitat parameters may bewarranted.

Figure 25: LWD as a Function of IC in Puget Sound Streams (Horner et al., 1997)




%IC Used as Indicator

Galli, 1990Increase in stream temperatures of five to 12 degreesFahrenheit in urban watersheds; stream warming linked to IC. MD


Johnson, 1995Up to 10 degrees Fahrenheit increases in stream temperaturesafter summer storm events in an urban area MN

LeBlanc et al., 1997 Calibrated a model predicting stream temperature increaseas a result of urbanization

Ontario

MCDEP, 2000Monitoring effect of urbanization and stormwater ponds onstream temperatures revealed stream warming associatedwith urbanization and stormwater ponds

MD

Paul et al., 2001Daily mean stream temperatures in summer increased withurban land use GA

3.5 Increased Stream Warming

IC directly influences our local weather inurban areas. This effect is obvious to anyonewalking across a parking lot on a hot summerday, when temperatures often reach a scorch-ing 110 to 120 degrees F. Parking lots andother hard surfaces tend to absorb solar energyand release it slowly. Furthermore, they lackthe normal cooling properties of trees andvegetation, which act as natural air condition-ers. Finally, urban areas release excess heat asa result of the combustion of fossil fuels forheating, cooling and transportation. As a result,highly urban areas tend to be much warmerthan their rural counterparts and are known asurban heat islands. Researchers have found thatsummer temperatures tend to be six to eightdegrees F warmer in the summer and two tofour degrees F warmer during the wintermonths.

Water temperature in headwater streams isstrongly influenced by local air temperatures.Summer temperatures in urban streams havebeen shown to increase by as much as five to12 degrees F in response to watershed develop-ment (Table 15). Increased water temperaturescan preclude temperature-sensitive speciesfrom being able to survive in urban streams.

Figure 26 shows the stream warming phenom-enon in small headwater streams in the Mary-land Piedmont.

Galli (1990) reported that stream temperaturesthroughout the summer increased in urbanwatersheds. He monitored five headwaterstreams in the Maryland Piedmont withdifferent levels of IC. Each urban stream hadmean temperatures that were consistentlywarmer than a forested reference stream, andstream warming appeared to be a directfunction of watershed IC. Other factors, suchas lack of riparian cover and the presence ofponds, were also demonstrated to amplifystream warming, but the primary contributingfactor appeared to be watershed IC.

Johnson (1995) studied how stormwaterinfluenced an urban trout stream in Minnesotaand reported up to a 10 degree F increase instream water temperatures after summer stormevents. Paul et al. (2001) evaluated streamtemperatures for 30 subwatersheds to theEtowah River in Georgia, which ranged fromfive to 61% urban land. They found a correla-tion between summer daily mean water tem-peratures and the percentage of urban land in asubwatershed.

Table 15: Research Review of Thermal Impacts in Urban Streams



Discharges from stormwater ponds can alsocontribute to stream warming in urban water-sheds. Three studies highlight the temperatureincrease that can result from stormwater ponds.A study in Ontario found that baseflow tem-peratures below wet stormwater ponds in-creased by nine to 18 degrees F in the summer(SWAMP, 2000a, b). Oberts (1997) also

Figure 26: Stream Temperature Increase in Response to IC in MarylandPiedmont Streams (Galli, 1990)

measured change in the baseflow temperatureas it flowed through a wetland/wet pondsystem in Minnesota. He concluded that thetemperature had increased by an average ofnine degrees F during the summer months.Galli (1988) also observed a mean increase oftwo to 10 degrees F in four stormwater pondslocated in Maryland.



3.6 Alteration of StreamChannel Networks

Urban stream channels are often severelyaltered by man. Channels are lined with rip rapor concrete, natural channels are straightened,and first order and ephemeral streams areenclosed in storm drain pipes. From an engi-neering standpoint, these modifications rapidlyconvey flood waters downstream and locallystabilize stream banks. Cumulatively, however,these modifications can have a dramatic effecton the length and habitat quality of headwaterstream networks.

3.6.1 Channel Modification

Over time, watershed development can alter oreliminate a significant percentage of theperennial stream network. In general, the lossof stream network becomes quite extensivewhen watershed IC exceeds 50%. This loss isstriking when pre- and post-developmentstream networks are compared (Figure 27).The first panel illustrates the loss of streamnetwork over time in a highly urban NorthernVirginia watershed; the second panel showshow the drainage network of Rock Creek haschanged in response to watershed develop-ment.

Figure 27: a. Drainage Network of Rock Creek, D.C. (Dunne and Leopold, 1978) andb. Drainage Network of Four Mile Run, VA Before and After Urbanization (NVRC, 2001)

a.

b.

1913 1964

1917 1998



In a national study of 269 gaged urban water-sheds, Sauer et al. (1983) observed thatchannelization and channel hardening wereimportant watershed variables that controlpeak discharge rates. The channel modifica-tions increase the efficiency with which runoffis transported through the stream channel,increasing critical shear stress velocities andcausing downstream channel erosion.

Figure 28: Fish Migration Barriers in the Anacostia Watershed of D.C. and MD (MWCOG, 1992)

3.6.2 Barriers to Fish Migration

Infrastructure such as bridges, dams, pipelinesand culverts can create partial or total barriersto fish migration and impair the ability of fishto move freely in a watershed. Blockages canhave localized effects on small streams wherenon-migratory fish species can be preventedfrom re-colonizing upstream areas after acutelytoxic events. The upstream movement ofanadromous fish species such as shad, herring,salmon and steelhead can also be blocked bythese barriers. Figure 28 depicts the prevalenceof fish barriers in the Anacostia Watershed(MWCOG, 1992).



3.7 Conclusion

Watershed development and the associatedincrease in IC have been found to significantlydegrade the physical habitat of urban streams.In alluvial streams, the effects of channelenlargement and sediment transport can besevere at relatively low levels of IC (10 to20%). However, the exact response of anystream is also contingent upon a combinationof other physical factors such as geology,vegetation, gradient, the age of development,sediment supply, the use and design of storm-water treatment practices, and the extent ofriparian buffers (Bledsoe, 2001).

Despite the uncertainty introduced by thesefactors, the limited geomorphic research todate suggests that physical habitat quality isalmost always degraded by higher levels ofwatershed IC. Even in bedrock-controlledchannels, where sediment transport andchannel enlargement may not be as dramatic,researchers have noted changes in streamhabitat features, such as embeddedness, loss ofLWD, and stream warming.

Overall, the following conclusions can bemade about the influence of watershed devel-opment on the physical habitat of urbanstreams:

• The major changes in physical habitat inurban streams are caused by the increasedfrequency and duration of bankfull andsub-bankfull discharges, and the attendantchanges in sediment supply and transport.As a consequence, many urban streamsexperience significant channel enlarge-ment. Generally, channel enlargement ismost evident in alluvial streams.

• Typical habitat changes observed in urbanstreams include increased embeddedness,reduced supply of LWD, and simplifica-tion of stream habitat features such aspools, riffles and runs, as well as reducedchannel sinuosity.

• Stream warming is often directly linked towatershed development, although moresystematic subwatershed sampling isneeded to precisely predict the extent ofwarming.

• Channel straightening, hardening andenclosure and the creation of fish barriersare all associated with watershed develop-ment. More systematic research is neededto establish whether these variables can bepredicted based on watershed IC.

• In general, stream habitat diminishes atabout 10% watershed IC, and becomesseverely degraded beyond 25% watershedIC.

While our understanding of the relationshipbetween stream habitat features and watersheddevelopment has improved in recent years, thetopic deserves greater research in three areas.First, more systematic monitoring of compos-ite habitat variables needs to be conductedacross the full range of watershed IC. Inparticular, research is needed to define theapproximate degree of watershed IC whereurban streams are transformed into urbandrainage systems.

Second, additional research is needed toexplore the relationship between watershed ICand individual and measurable stream habitatparameters, such as bank erosion, channelsinuosity, pool depth and wetted perimeter.Lastly, more research is needed to determine ifwatershed treatment such as stormwaterpractices and stream buffers can mitigate theimpacts of watershed IC on stream habitat.Together, these three research efforts couldprovide a technical foundation to develop amore predictive model of how watersheddevelopment influences stream habitat.


Chapter 4: Water Quality Impacts of Impervious Cover

Chapter 4: Water Quality Impacts ofImpervious Cover

This chapter presents information on pollutantconcentrations found in urban stormwaterrunoff based on a national and regional dataassessment for nine categories of pollutants.Included is a description of the SimpleMethod, which can be used to estimate pollut-ant loads based on the amount of IC found in acatchment or subwatershed. This chapter alsoaddresses specific water quality impacts ofstormwater pollutants and explores research onthe sources and source areas of stormwaterpollutants.


4.1 Introduction4.2 Summary of National and Regional

Stormwater Pollutant ConcentrationData

4.3 Relationship Between Pollutant Loadsand IC: The Simple Method

4.4 Sediment4.5 Nutrients4.6 Trace Metals4.7 Hydrocarbons (PAH and Oil and

Grease)4.8 Bacteria and Pathogens4.9 Organic Carbon4.10 MTBE4.11 Pesticides4.12 Deicers4.13 Conclusion

4.1 Introduction

Streams are usually the first aquatic system toreceive stormwater runoff, and their waterquality can be compromised by the pollutantsit contains. Stormwater runoff typicallycontains dozens of pollutants that are detect-able at some concentration, however small.Simply put, any pollutant deposited or derivedfrom an activity on land will likely end up instormwater runoff, although certain pollutantsare consistently more likely to cause water

quality problems in receiving waters. Pollut-ants that are frequently found in stormwaterrunoff can be grouped into nine broad catego-ries: sediment, nutrients, metals, hydrocarbons,bacteria and pathogens, organic carbon,MTBE, pesticides, and deicers.

The impact that stormwater pollutants exert onwater quality depends on many factors, includ-ing concentration, annual pollutant load, andcategory of pollutant. Based on nationallyreported concentration data, there is consider-able variation in stormwater pollutant concen-trations. This variation has been at leastpartially attributed to regional differences,including rainfall and snowmelt. The volumeand regularity of rainfall, the length of snowaccumulation, and the rate of snowmelt can allinfluence stormwater pollutant concentrations.

The annual pollutant load can have long-termeffects on stream water quality, and is particu-larly important information for stormwatermanagers to have when dealing with non-pointsource pollution control. The Simple Method isa model developed to estimate the pollutantload for chemical pollutants, assuming that theannual pollutant load is a function of IC. It isan effective method for determining annualsediment, nutrient, and trace metal loads. Itcannot always be applied to other stormwaterpollutants, since they are not always correlatedwith IC.

The direct water quality impact of stormwaterpollutants also depends on the type of pollut-ant, as different pollutants impact streamsdifferently. For example, sediments affectstream habitat and aquatic biodiversity;nutrients cause eutrophication; metals, hydro-carbons, deicers, and MTBE can be toxic toaquatic life; and organic carbon can lowerdissolved oxygen levels.

The impact stormwater pollutants have on



water quality can also directly influence humanuses and activities. Perhaps the pollutants ofgreatest concern are those with associatedpublic health impacts, such as bacteria andpathogens. These pollutants can affect theavailability of clean drinking water and limitconsumptive recreational activities, such asswimming or fishing. In extreme situations,these pollutants can even limit contact recre-ational activities such as boating and wading.

It should be noted that although there is muchresearch available on the effects of urbaniza-tion on water quality, the majority has not beenfocused on the impact on streams, but on theresponse of lakes, reservoirs, rivers andestuaries. It is also important to note that notall pollutants are equally represented in moni-toring conducted to date. While we possessexcellent monitoring data for sediment,nutrients and trace metals, we have relativelylittle monitoring data for pesticides, hydrocar-bons, organic carbon, deicers, and MTBE.

4.2 Summary of National andRegional Stormwater PollutantConcentration Data

4.2.1 National Data

National mean concentrations of typicalstormwater pollutants are presented in Table16. National stormwater data are compiledfrom the Nationwide Urban Runoff Program(NURP), with additional data obtained fromthe U.S.Geological Survey (USGS), as well asinitial stormwater monitoring conducted forEPA’s National Pollutant Discharge Elimina-tion System (NPDES) Phase I stormwaterprogram.

In most cases, stormwater pollutant data isreported as an event mean concentration(EMC), which represents the average concen-tration of the pollutant during an entire storm-water runoff event.

When evaluating stormwater EMC data, it isimportant to keep in mind that regional EMCscan differ sharply from the reported nationalpollutant EMCs. Differences in EMCs betweenregions are often attributed to the variation inthe amount and frequency of rainfall andsnowmelt.

4.2.2 Regional DifferencesDue to Rainfall

The frequency of rainfall is important, since itinfluences the accumulation of pollutants on ICthat are subsequently available for wash-offduring storm events. The USGS developed anational stormwater database encompassing1,123 storms in 20 metropolitan areas and usedit as the primary data source to define regionaldifferences in stormwater EMCs. Driver(1988) performed regression analysis todetermine which factors had the greatestinfluence on stormwater EMCs and determinedthat annual rainfall depth was the best overallpredictor. Driver grouped together stormwaterEMCs based on the depth of average annualrainfall, and Table 17 depicts the regionalrainfall groupings and general trends for each



Pollutant Source EMCs

Number of EventsMean Median

Sediments (mg/l)

TSS (1) 78.4 54.5 3047

Nutrients (mg/l)Total P (1) 0.32 0.26 3094

Soluble P (1) 0.13 0.10 1091

Total N (1) 2.39 2.00 2016

TKN (1) 1.73 1.47 2693

Nitrite & Nitrate (1) 0.66 0.53 2016

Metals (Fg/l)Copper (1) 13.4 11.1 1657

Lead (1) 67.5 50.7 2713

Zinc (1) 162 129 2234

Cadmium (1) 0.7 N/R 150

Chromium (4) 4 7 164

Hydrocarbons (mg/l)PAH (5) 3.5 N/R N/R

Oil and Grease (6) 3 N/R N/R

Bacteria and Pathogens (colonies/ 100ml)Fecal Coliform (7) 15,038 N/R 34

FecalStreptococci (7) 35,351 N/R 17

Organic Carbon (mg/l)TOC (11) 17 15.2 19 studies

BOD (1) 14.1 11.5 1035

COD (1) 52.8 44.7 2639

MTBE (Fg/l)

MTBE (8) N/R 1.6 592

Pesticides (Fg/l)

Diazinon(10) N/R 0.025 326

(2) N/R 0.55 76

Chlorpyrifos (10) N/R N/R 327

Atrazine (10) N/R 0.023 327

Prometon (10) N/R 0.031 327

Simazine (10) N/R 0.039 327

Chloride (mg/l)Chloride (9) N/R 397 282Sources: (1) Smullen and Cave, 1998; (2) Brush et al., 1995; (3) Baird et al., 1996; (4) Bannerman et al., 1996; (5)

Rabanal and Grizzard, 1995; (6) Crunkilton et al., 1996; (7) Schueler, 1999; (8) Delzer, 1996; (9) EnvironmentCanada, 2001; (10) USEPA, 1998; (11) CWP, 2001a N/R - Not Reported

Pollutant Source EMCs

Number of EventsMean Median

Sediments (mg/l)

TSS (1) 78.4 54.5 3047

Nutrients (mg/l)Total P (1) 0.32 0.26 3094

Soluble P (1) 0.13 0.10 1091

Total N (1) 2.39 2.00 2016

TKN (1) 1.73 1.47 2693

Nitrite & Nitrate (1) 0.66 0.53 2016

Metals (Fg/l)Copper (1) 13.4 11.1 1657

Lead (1) 67.5 50.7 2713

Zinc (1) 162 129 2234

Cadmium (1) 0.7 N/R 150

Chromium (4) 4 7 164

Hydrocarbons (mg/l)PAH (5) 3.5 N/R N/R

Oil and Grease (6) 3 N/R N/R

Bacteria and Pathogens (colonies/ 100ml)Fecal Coliform (7) 15,038 N/R 34

FecalStreptococci (7) 35,351 N/R 17

Organic Carbon (mg/l)TOC (11) 17 15.2 19 studies

BOD (1) 14.1 11.5 1035

COD (1) 52.8 44.7 2639

MTBE (Fg/l)

MTBE (8) N/R 1.6 592

Pesticides (Fg/l)

Diazinon(10) N/R 0.025 326

(2) N/R 0.55 76

Chlorpyrifos (10) N/R N/R 327

Atrazine (10) N/R 0.023 327

Prometon (10) N/R 0.031 327

Simazine (10) N/R 0.039 327

Chloride (mg/l)Chloride (9) N/R 397 282Sources: (1) Smullen and Cave, 1998; (2) Brush et al., 1995; (3) Baird et al., 1996; (4) Bannerman et al., 1996; (5)

Rabanal and Grizzard, 1995; (6) Crunkilton et al., 1996; (7) Schueler, 1999; (8) Delzer, 1996; (9) EnvironmentCanada, 2001; (10) USEPA, 1998; (11) CWP, 2001a N/R - Not Reported

MTBE (Fg/l)

592

Table 16: National EMCs for Stormwater Pollutants

region. Table 18 illustrates the distribution ofstormwater EMCs for a range of rainfallregions from 13 local studies, based on other

monitoring studies. In general, stormwaterEMCs for nutrients, suspended sediment andmetals tend to be higher in arid and semi-arid



regions and tend to decrease slightly whenannual rainfall increases (Table 19).

It is also hypothesized that a greater amount ofsediment is eroded from pervious surfaces inarid or semi-arid regions than in humid regionsdue to the sparsity of protective vegetativecover. Table 19 shows that the highest concen-trations of total suspended solids were re-corded in regions with least rainfall. In addi-tion, the chronic toxicity standards for severalmetals are most frequently exceeded duringlow rainfall regions (Table 20).

4.2.3 Cold Region Snowmelt Data

In colder regions, snowmelt can have a signifi-cant impact on pollutant concentrations. Snowaccumulation in winter coincides with pollut-ant build-up; therefore, greater concentrationsof pollutants are measured during snowmeltevents. Sources of snowpack pollution in urbanareas include wet and dry atmospheric deposi-tion, traffic emissions, urban litter, deterioratedinfrastructure, and deicing chemicals andabrasives (WERF, 1999).

Oberts et al. (1989) measured snowmeltpollutants in Minnesota streams and found thatas much as 50% of annual sediment, nutrient,hydrocarbon and metal loads could be attrib-uted to snowmelt runoff during late winter andearly spring. This trend probably applies to anyregion where snow cover persists throughmuch of the winter. Pollutants accumulate inthe snowpack and then contribute high concen-trations during snowmelt runoff. Oberts (1994)

Region Annual Rainfall States Monitored Concentration Data

Region I: Low Rainfall

<20 inches AK, CA, CO, NM,UT

Highest mean and median values forTotal N, Total P, TSS and COD

Region II: ModerateRainfall

20 40 inchesHA, IL, MI, MN, MI,

NY, TX, OR, OH,WA, WI

Higher mean and median valuesthan Region III for TSS, dissolvedphosphorus and cadmium

Region III: High Rainfall

>40 inches FL, MD, MA, NC,

NH, NY, TX, TN, AR

Lower values for many parameterslikely due to the frequency of stormsand the lack of build up in pollutants

Table 17: Regional Groupings by Annual Rainfall Amount (Driver, 1988)

described four types of snowmelt runoff eventsand the resulting pollutant characteristics(Table 21).

A typical hydrograph for winter and earlyspring snow melts in a northern cold climate isportrayed in Figure 29. The importance ofsnowpack melt on peak runoff during March1989 can clearly be seen for an urban water-shed located in St. Paul, Minnesota.

Major source areas for snowmelt pollutantsinclude snow dumps and roadside snowpacks.Pollutant concentrations in snow dumps can beas much as five times greater than typicalstormwater pollutant concentrations (Environ-ment Canada, 2001). Snow dumps and packsaccumulate pollutants over the winter monthsand can release them during a few rain or snowmelt events in the early spring. High levels ofchloride, lead, phosphorus, biochemicaloxygen demand, and total suspended solidshave been reported in snow pack runoff ( LaBarre et al, 1973; Oliver et al., 1974; Pierstorffand Bishop, 1980; Scott and Wylie, 1980; VanLoon, 1972).

Atmospheric deposition can add pollutants tosnow piles and snowpacks. Deposited pollut-ants include trace metals, nutrients and par-ticles that are primarily generated by fossil fuelcombustion and industrial emissions (Boomand Marsalek, 1988; Horkeby and Malmqvist,1977; Malmqvist, 1978; Novotny and Chester,1981; Schrimpff and Herrman, 1979).



Region Total N (median) Total P (median) TSS (mean)

Region I: Low Rainfall 4 0.45 320

Region II: Moderate Rainfall 2.3 0.31 250

Region III: High Rainfall 2.15 0.31 120

Table 19: Mean and Median Nutrient and Sediment Stormwater Concentrations forResidential Land Use Based on Rainfall Regions (Driver, 1988)

Region I - Low Rainfall Region II - ModerateRainfall

Region III - High Rainfall SnowN

atio

na

l

Ph

oe

nix

, AZ

San

Die

go

, C

A

Bois

e, I

D

De

nve

r, C

O

Da

llas,

TX

Ma

rqu

ett

e, M

I

Au

stin

, TX

MD

Lou

isvi

lle, K

Y

GA

FL MN

Reference (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (11) (12)

AnnualRainfall(in.)

N/A 7.1" 10" 11" 15" 28" 32" 32" 41" 43" 51" 52" N/R

Number ofEvents

3000 40 36 15 35 32 12 N/R 107 21 81 N/R 49

Pollutant

TSS 78.4 227 330 116 242 663 159 190 67 98 258 43 112

Total N 2.39 3.26 4.55 4.13 4.06 2.70 1.87 2.35 N/R 2.37 2.52 1.74 4.30

Total P 0.32 0.41 0.7 0.75 0.65 0.78 0.29 0.32 0.33 0.32 0.33 0.38 0.70

Soluble P 0.13 0.17 0.4 0.47 N/R N/R 0.04 0.24 N/R 0.21 0.14 0.23 0.18

Copper 14 47 25 34 60 40 22 16 18 15 32 1.4 N/R

Lead 68 72 44 46 250 330 49 38 12.5 60 28 8.5 100

Zinc 162 204 180 342 350 540 111 190 143 190 148 55 N/R

BOD 14.1 109 21 89 N/R 112 15.4 14 14.4 88 14 11 N/R

COD 52.8 239 105 261 227 106 66 98 N/R 38 73 64 112

Sources: Adapted from Caraco, 2000a: (1) Smullen and Cave, 1998; (2) Lopes et al.; 1995; (3) Schiff, 1996; (4) Kjelstrom, 1995(computed); (5) DRCOG, 1983, (6) Brush et al., 1995; (7) Steuer et al., 1997; (8) Barrett et al., 1995; (9) Barr, 1997; (10) Evaldi et al., 1992; (11)

Thomas and McClelland, 1995; (12) Oberts, 1994 N/R = Not Reported; N/A = Not Applicable

Table 18: Stormwater Pollutant Event Mean Concentration for Different U.S. Regions(Units: mg/l, except for metals which are in FFFFFg/l)



Cadmium Copper Lead Zinc

EPA Standards 10 Fg/l 12 Fg/l 32 Fg/l 47 Fg/l

Percent Exceedance of EPA Standards

Region I: Low Rainfall 1.5% 89% 97% 97%

Region II: Moderate Rainfall 0 78% 89% 85%

Region III: High Rainfall 0 75% 91% 84%

Table 20: EPA 1986 Water Quality Standards and Percentage of MetalConcentrations Exceeding Water Quality Standards by Rainfall Region (Driver, 1988)

SnowmeltStage

Duration/Frequency

RunoffVolume Pollutant Characteristics

Pavement Short, but manytimes in winter

LowAcidic, high concentrations of solublepollutants; Chloride, nitrate, lead;total load is minimal

Roadside Moderate Moderate Moderate concentrations of bothsoluble and particulate pollutants

Pervious AreaGradual, oftenmost at end of

seasonHigh

Dilute concentrations of solublepollutants; moderate to highconcentrations of particulatepollutants depending on flow

Rain-on-Snow Short Extreme

High concentrations of particulatepollutants; moderate to highconcentrations of soluble pollutants;high total load

Table 21: Runoff and Pollutant Characteristics of Snowmelt Stages (Oberts, 1994)

Figure 29: Snowmelt Runoff Hydrograph for Minneapolis Stream (Oberts, 1994)



4.3 Relationship BetweenPollutant Loads and IC:The Simple Method

Urban stormwater runoff contains a wide rangeof pollutants that can degrade downstreamwater quality. The majority of stormwatermonitoring research conducted to date supportsseveral generalizations. First, the unit areapollutant load delivered to receiving waters bystormwater runoff increases in direct propor-tion to watershed IC. This is not altogethersurprising, since pollutant load is the productof the average pollutant concentration andstormwater runoff volume. Given that runoffvolume increases in direct proportion to IC,pollutant loads must automatically increasewhen IC increases, as long the average pollut-ant concentration stays the same (or increases).

This relationship is a central assumption inmost simple and complex pollutant loadingmodels (Bicknell et al., 1993; Donigian andHuber, 1991; Haith et al., 1992; Novotny andChester, 1981; NVPDC, 1987; Pitt andVoorhees, 1989).

Recognizing the relationship between IC andpollutant loads, Schueler (1987) developed the“Simple Method” to quickly and easily esti-mate stormwater pollutant loads for smallurban watersheds (see Figure 30). Estimates ofpollutant loads are important to watershedmanagers as they grapple with costly decisionson non-point source control. The SimpleMethod is empirical in nature and utilizes theextensive regional and national database(Driscoll, 1983; MWCOG, 1983; USEPA,1983). Figure 30 provides the basic equationsto estimate pollutant loads using the Simple

Figure 30: The Simple Method - Basic Equations

The Simple Method estimates pollutant loads as the product of annual runoff volumeand pollutant EMC, as:

(1) L = 0.226 * R * C * AWhere: L = Annual load (lbs), and:

R = Annual runoff (inches)C = Pollutant concentration in stormwater, EMC (mg/l)A = Area (acres)0.226 = Unit conversion factor

For bacteria, the equation is slightly different, to account for the differences in units. Themodified equation for bacteria is:

(2) L = 1.03 *10-3 * R * C * AWhere: L = Annual load (Billion Colonies), and:

R = Annual runoff (inches)C = Bacteria concentration (#/100 ml)A = Area (acres)1.03 * 10-3 = Unit conversion factor

Annual Runoff

The Simple Method calculates the depth of annual runoff as a product of annual runoffvolume and a runoff coefficient (Rv). Runoff volume is calculated as:

(3) R = P * Pj * RvWhere: R = Annual runoff (inches), and:

P = Annual rainfall (inches)Pj = Fraction of annual rainfall events that produce runoff (usually 0.9)Rv = Runoff coefficient

In the Simple Method, the runoff coefficient is calculated based on IC in thesubwatershed. The following equation represents the best fit line for the data set (N=47,R2=0.71).

(4) Rv=0.05+0.9IaWhere: Rv = runoff coefficient, and:

Ia = Impervious fraction



Method. It assumes that loads of stormwaterpollutants are a direct function of watershedIC, as IC is the key independent variable in theequation.

The technique requires a modest amount ofinformation, including the subwatersheddrainage area, IC, stormwater runoff pollutantEMCs, and annual precipitation. With theSimple Method, the investigator can eitherdivide up land use into specific areas (i.e.residential, commercial, industrial, and road-way) and calculate annual pollutant loads foreach land use, or utilize a generic urban landuse. Stormwater pollutant EMC data can bederived from the many summary tables oflocal, regional, or national monitoring effortsprovided in this chapter (e.g., Tables 16, 18,22, 28, 30, 35, 36, 40, and 44). The model alsorequires different IC values for separate landuses within a subwatershed. Representative ICdata from Cappiella and Brown (2001) wereprovided in Table 2 (Chapter 1).

Additionally, the Simple Method should not beused to estimate annual pollutant loads ofdeicers, hydrocarbons and MTBE, becausethey have not been found to be correlated withIC. These pollutants have been linked to otherindicators. Chlorides, hydrocarbons and MTBEare often associated with road density andvehicle miles traveled (VMT). Pesticides areassociated with turf area, and traffic patternsand “hotspots” have been noted as potentialindicators for hydrocarbons and MTBE.

Limitations of the Simple MethodThe Simple Method should provide reasonableestimates of changes in pollutant exportresulting from urban development. However,several caveats should be kept in mind whenapplying this method.

The Simple Method is most appropriate forassessing and comparing the relativestormflow pollutant load changes from differ-ent land uses and stormwater treatment sce-narios. The Simple Method provides estimatesof storm pollutant export that are probablyclose to the “true” but unknown value for adevelopment site, catchment, or subwatershed.However, it is very important not to over-emphasize the precision of the load estimateobtained. For example, it would be inappropri-ate to use the Simple Method to evaluaterelatively similar development scenarios (e.g.,34.3% versus 36.9% IC). The Simple Methodprovides a general planning estimate of likelystorm pollutant export from areas at the scaleof a development site, catchment orsubwatershed. More sophisticated modeling isneeded to analyze larger and more complexwatersheds.

In addition, the Simple Method only estimatespollutant loads generated during storm events.It does not consider pollutants associated withbaseflow during dry weather. Typically,baseflow is negligible or non-existent at thescale of a single development site and can besafely neglected. However, catchments andsubwatersheds do generate significantbaseflow volume. Pollutant loads in basefloware generally low and can seldom be distin-guished from natural background levels(NVPDC, 1979).

Consequently, baseflow pollutant loadsnormally constitute only a small fraction of thetotal pollutant load delivered from an urbanarea. Nevertheless, it is important to rememberthat the load estimates refer only to stormevent derived loads and should not be confusedwith the total pollutant load from an area. Thisis particularly important when the developmentdensity of an area is low. For example, in a lowdensity residential subwatershed (IC < 5%), asmuch as 75% of the annual runoff volumecould occur as baseflow. In such a case, annualbaseflow load may be equivalent to the annualstormflow load.



4.4 Sediment

Sediment is an important and ubiquitouspollutant in urban stormwater runoff. Sedimentcan be measured in three distinct ways: TotalSuspended Solids (TSS), Total DissolvedSolids (TDS) and turbidity. TSS is a measureof the total mass suspended sediment particlesin water. The measurement of TSS in urbanstormwater helps to estimate sediment loadtransported to local and downstream receivingwaters. Table 22 summarizes stormwaterEMCs for total suspended solids, as reportedby Barrett et al. (1995), Smullen and Cave(1998), and USEPA (1983). TDS is a measureof the dissolved solids and minerals present instormwater runoff and is used as a primaryindication of the purity of drinking water.Since few stormwater monitoring efforts havefocused on TDS, they are not reported in thisdocument. Turbidity is a measure of howsuspended solids present in water reduce theability of light to penetrate the water column.Turbidity can exert impacts on aquatic biota,such as the ability of submerged aquaticvegetation to receive light and the ability offish and aquatic insects to use their gills (Table23).

4.4.1 Concentrations

TSS concentrations in stormwater across thecountry are well documented. Table 18 reviewsmean TSS EMCs from 13 communities acrossthe country and reveals a wide range of re-corded concentrations. The lowest concentra-tion of 43 mg/l was reported in Florida, whileTSS reached 663 mg/l in Dallas, Texas.

Variation in sediment concentrations has beenattributed to regional rainfall differences(Driver, 1988); construction site runoff(Leopold, 1968); and bank erosion(Dartiguenave et al., 1997). National values areprovided in Table 22.

Turbidity levels are not as frequently reportedin national and regional monitoring summaries.Barrett and Malina (1998) monitored turbidityat two sites in Austin, Texas and reported amean turbidity of 53 NTU over 34 stormevents (Table 22).

4.4.2 Impacts of Sediment onStreams

The impacts of sediment on aquatic biota arewell documented and can be divided intoimpacts caused by suspended sediment andthose caused by deposited sediments (Tables23 and 24).

In general, high levels of TSS and/or turbiditycan affect stream habitat and cause sedimenta-tion in downstream receiving waters. Depos-ited sediment can cover benthic organismssuch as aquatic insects and freshwater mus-sels. Other problems associated with highsediments loads include stream warming byreflecting radiant energy due to increasedturbidity (Kundell and Rasmussen, 1995),decreased flow capacity (Leopold, 1973), andincreasing overbank flows (Barrett and Malina,1998). Sediments also transport other pollut-ants which bind to sediment particles. Signifi-cant levels of pollutants can be transported bysediment during stormwater runoff events,

Pollutant EMCs Number

of EventsSource

Mean Median

TSS (mg/l)78.4 54.5 3047 Smullen and Cave, 1998

174 113 2000 USEPA, 1983

Turbidity (NTU) 53 N/R 423 Barrett and Malina, 1998

N/R = Not Reported

Pollutant EMCs Number

of EventsSource

Mean Median

TSS (mg/l)78.4 54.5 3047 Smullen and Cave, 1998

174 113 2000 USEPA, 1983

Turbidity (NTU) 53 N/R 423 Barrett and Malina, 1998

N/R = Not Reported

Table 22: EMCs for Total Suspended Solids and Turbidity



including trace metals, hydrocarbons andnutrients (Crunkilton et al., 1996;Dartiguenave et al., 1997; Gavin and Moore,1982; Novotny and Chester, 1989; Schueler1994b).

4.4.3 Sources and Source Areasof Sediment

Sediment sources in urban watersheds includestream bank erosion; erosion from exposedsoils, such as from construction sites; andwashoff from impervious areas (Table 25).

As noted in this chapter, streambank erosion isgenerally considered to be the primary sourceof sediment to urban streams. Recent studiesby Dartiguenave et al. (1997) and Trimble(1997) determined that streambank erosion

contributes the majority of the annual sedimentbudget of urban streams. Trimble (1997)directly measured stream cross sections,sediment aggradation and suspended sedimentloads and determined that two-thirds of theannual sediment budget of a San Diego,California watershed was supplied bystreambank erosion. Dartiguenave et al. (1997)developed a GIS based model in Austin, Texasto determine the effects of stream bank erosionon the annual sediment budget. They comparedmodeled sediment loads from the watershedwith the actual sediment loads measured atUSGS gaging stations and concluded that morethan 75% of the sediment load came fromstreambank erosion. Dartiguenave et al. (1997)reported that sediment load per unit areaincreases with increasing IC (Figure 31).

1. Physical smothering of benthic aquatic insect community2. Reduced survival rates for fish eggs3. Destruction of fish spawning areas and eggs4. Embeddedness of stream bottom reduced fish and macroinvertebrate habitat value5. Loss of trout habitat when fine sediments are deposited in spawning or riffle-runs6. Sensitive or threatened darters and dace may be eliminated from fish community7. Increase in sediment oxygen demand can deplete dissolved oxygen in streams8. Significant contributing factor in the alarming decline of freshwater mussels9. Reduced channel capacity, exacerbating downstream bank erosion and flooding10. Reduced flood transport capacity under bridges and through culverts11. Deposits diminish scenic and recreational values of waterways

Abrades and damages fish gills, increasing risk of infection and disease

Scouring of periphyton from stream (plants attached to rocks)

Loss of sensitive or threatened fish species when turbidity exceeds 25 NTU Shifts in fish community toward more sediment-tolerant species

Decline in sunfish, bass, chub and catfish when month turbidity exceeds 100 NTU Reduces sight distance for trout, with reduction in feeding efficiency

Reduces light penetration causing reduction in plankton and aquatic plant growth

Adversely impacts aquatic insects, which are the base of the food chain Slightly increases the stream temperature in the summer

Suspended sediments can be a major carrier of nutrients and metals Reduces anglers chances of catching fish

Table 23: Summary of Impacts of Suspended Sediment on theAquatic Environment (Schueler and Holland, 2000)

Table 24: Summary of Impacts of Deposited Sediments on the Aquatic Environment(Schueler and Holland, 2000)



Sediment loads are also produced by washoffof sediment particles from impervious areasand their subsequent transport in stormwaterrunoff sediment. Source areas include parkinglots, streets, rooftops, driveways and lawns.Streets and parking lots build up dirt and grimefrom the wearing of the street surface, exhaustparticulates, “blown on” soil and organicmatter, and atmospheric deposition. Lawnrunoff primarily contains soil and organicmatter. Urban source areas that produce thehighest TSS concentrations include streets,parking lots and lawns (Table 26).

Parking lots and streets are not only respon-sible for high concentrations of sediment butalso high runoff volumes. The SLAMM sourceloading model (Pitt and Voorhees, 1989) looksat runoff volume and concentrations of pollut-ants from different urban land uses and pre-dicts stream loading. When used in the Wis-consin and Michigan subwatersheds, it demon-strated that parking lots and streets wereresponsible for over 70% of the TSS deliveredto the stream. (Steuer et al., 1997;Waschbusch et al., 2000).

Figure 31: TSS from Bank Erosion vs. IC in Texas Streams (Daringuenave et al., 1997)

Sources Loading Source

Bank Erosion75% of stream sediment budget Dartinguenave et al., 1997

66% of stream sediment budget Trimble, 1997

Overland Flow- Lawns

397 mg/l (geometric mean) Bannerman et al., 1993

262 mg/l Steuer et al., 1997

11.5% (estimated; 2 sites) Waschbusch et al., 2000

Construction Sites 200 to 1200 mg/l Table 27

Washoff from ImperviousSurfaces

78 mg/l (mean) Table 16

Sources Loading Source

Bank Erosion75% of stream sediment budget Dartinguenave et al., 1997

66% of stream sediment budget Trimble, 1997

Overland Flow- Lawns

397 mg/l (geometric mean) Bannerman et al., 1993

262 mg/l Steuer et al., 1997

11.5% (estimated; 2 sites) Waschbusch et al., 2000

Construction Sites 200 to 1200 mg/l Table 27

Washoff from ImperviousSurfaces

78 mg/l (mean) Table 16

Table 25: Sources and Loading of Suspended Solids Sediment in Urban Areas



The third major source of sediment loads iserosion from construction sites. Several studieshave reported extremely high TSS concentra-tions in construction site runoff, and thesefindings are summarized in Table 27. TSSconcentrations from uncontrolled construction

SourceMean Inflow TSSConcentration

(mg/l)

Mean Outflow TSS Concentration

(mg/l) Location

Uncontrolled Sites

Horner et al., 1990 7,363 281 PNW

Schueler and Lugbill,1990 3,646 501 MD

York and Herb, 1978 4,200 N/R MD

Islam et al., 1988 2,950 N/R OH

Controlled Sites

Schueler and Lugbill, 1990 466 212 MD

Simulated Sediment Concentrations

Jarrett, 1996 9,700 800 PA

Sturm and Kirby, 1991 1,500-4,500 200-1,000 GA

Barfield and Clar, 1985 1,000-5,000 200-1,200 MD

Dartiguenave et al., 1997 N/R 600 TX

N/R = Not Reported

sites can be more than 150 times greater thanthose from undeveloped land (Leopold, 1968)and can be reduced if erosion and sedimentcontrol practices are applied to constructionsites.

Source Area Suspended Solids (mg/l)

Source (1) (2) (3)

Commercial Parking Lot 110 58 51

High Traffic Street 226 232 65

Medium Traffic Street 305 326 51

Low Traffic Street 175 662 68

Commercial Rooftop 24 15 18

Residential Rooftop 36 27 15

Residential Driveway 157 173 N/R

Residential Lawn 262 397 59

Sources: (1) Steuer et al., 1997; (2) Bannerman et al., 1993; (3) Waschbusch et al., 2000; N/R = NotReported

Table 26: Source Area Geometric Mean Concentrations for Suspended Solids in Urban Areas

Table 27: Mean TSS Inflow and Outflow at Uncontrolled, Controlled andSimulated Construction Sites



4.5 Nutrients

Nitrogen and phosphorus are essential nutrientsfor aquatic systems. However, when theyappear in excess concentrations, they can exerta negative impact on receiving waters. Nutrientconcentrations are reported in several ways.Nitrogen is often reported as nitrate (NO

3) and

nitrite (NO2), which are inorganic forms of

nitrogen; total nitrogen (Total N), which is thesum of nitrate, nitrite, organic nitrogen andammonia; and total Kjeldhal nitrogen (TKN),which is organic nitrogen plus ammonia.

Phosphates are frequently reported as solublephosphorus, which is the dissolved and reac-tive form of phosphorus that is available foruptake by plants and animals. Total phospho-rus (Total P) is also measured, which includesboth organic and inorganic forms of phospho-rus. Organic phosphorus is derived from livingplants and animals, while inorganic phosphateis comprised of phosphate ions that are oftenbound to sediments.


Many studies have indicated that nutrientconcentrations are linked to land use type, with

urban and agricultural watersheds producingthe highest nutrient loads (Chessman et al.1992; Paul et al., 2001; USGS, 2001b andWernick et al.,1998). Typical nitrogen andphosphorus EMC data in urban stormwaterrunoff are summarized in Table 28.

Some indication of the typical concentrationsof nitrate and phosphorus in stormwater runoffare evident in Figures 32 and 33. These graphsprofile average EMCs in stormwater runoffrecorded at 37 residential catchments acrossthe U.S. The average nitrate EMC is remark-ably consistent among residential neighbor-hoods, with most clustered around the mean of0.6 mg/l and a range of 0.25 to 1.4 mg/l. Theconcentration of phosphorus during storms isalso very consistent with a mean of 0.30 mg/land a rather tight range of 0.1 to 0.66 mg/l(Schueler, 1995).

The amount of annual rainfall can also influ-ence the magnitude of nutrient concentrationsin stormwater runoff. For example, bothCaraco (2000a) and Driver (1988) reported thatthe highest nutrient EMCs were found instormwater from arid or semi-arid regions.

Pollutant EMCs (mg/l) Number of

EventsSource

Mean Median

Total P0.315 0.259 3094 Smullen and Cave, 1998

0.337 0.266 1902 USEPA, 1983

Soluble P0.129 0.103 1091 Smullen and Cave, 1998

0.1 0.078 767 USEPA, 1983

Total N2.39 2.00 2016 Smullen and Cave, 1998

2.51 2.08 1234 USEPA, 1983

TKN1.73 1.47 2693 Smullen and Cave, 1998

1.67 1.41 1601 USEPA, 1983

Nitrite &Nitrate

0.658 0.533 2016 Smullen and Cave, 1998

0.837 0.666 1234 USEPA, 1983

Pollutant EMCs (mg/l) Number of

EventsSource

Mean Median

Total P0.315 0.259 3094 Smullen and Cave, 1998

0.337 0.266 1902 USEPA, 1983

Soluble P0.129 0.103 1091 Smullen and Cave, 1998

0.1 0.078 767 USEPA, 1983

Total N2.39 2.00 2016 Smullen and Cave, 1998

2.51 2.08 1234 USEPA, 1983

TKN1.73 1.47 2693 Smullen and Cave, 1998

1.67 1.41 1601 USEPA, 1983

Nitrite &Nitrate

0.658 0.533 2016 Smullen and Cave, 1998

0.837 0.666 1234 USEPA, 1983

Table 28: EMCs of Phosphorus and Nitrogen Urban Stormwater Pollutants



4.5.2 Impacts of Nutrientson Streams

Much research on the impact of nutrient loadshas been focused on lakes, reservoirs andestuaries, which can experience eutrophication.Nitrogen and phosphorus can contribute toalgae growth and eutrophic conditions, de-pending on which nutrient limits growth(USEPA, 1998). Dissolved oxygen is alsoaffected by eutrophication. When algae oraquatic plants that are stimulated by excessnutrients die off, they are broken down by

bacteria, which depletes the oxygen in thewater. Relatively few studies have specificallyexplored the impact of nutrient enrichment onurban streams. Chessman et al. (1992) studiedthe limiting nutrients for periphyton growth ina variety of streams and noted that the severityof eutrophication was related to low flowconditions. Higher flow rates in streams maycycle nutrients faster than in slow flow rates,thus diminishing the extent of stream eutrophi-cation.

Figure 32: Nitrate-Nitrogen Concentration in Stormwater Runoff at 37Sites Nationally (Schueler, 1999)

Figure 33: Total Phosphorus Concentration in Stormwater at 37Sites Nationally (Schueler, 1999)

mg

/lm

g/l



4.5.3 Sources and SourceAreas of Nutrients

Phosphorus is normally transported in surfacewater attached to sediment particles or insoluble forms. Nitrogen is normally trans-ported by surface water runoff in urban water-sheds. Sources for nitrogen and phosphorus inurban stormwater include fertilizer, pet waste,organic matter (such as leaves and detritus),and stream bank erosion. Another significantsource of nutrients is atmospheric deposition.Fossil fuel combustion by automobiles, powerplants and industry can supply nutrients in bothwet fall and dry fall. The Metropolitan Wash-ington Council of Governments (MWCOG,1983) estimated total annual atmosphericdeposition rates of 17 lbs/ac for nitrogen and0.7 lbs/ac for phosphorus in the Washington,D.C. metro area.

Research from the upper Midwest suggests“hot spot” sources can exist for both nitrogenand phosphorus in urban watersheds. Lawns, inparticular, contribute greater concentrations ofTotal N, Total P and dissolved phosphorus thanother urban source areas. Indeed, sourceresearch suggests that nutrient concentrations

in lawn runoff can be as much as four timesgreater than other urban sources such asstreets, rooftops or driveways (Bannerman etal., 1993; Steuer et al., 1997 and Waschbuschet al., 2000) (Table 29). This finding is signifi-cant, since lawns can comprise more than 50%of the total area in suburban watersheds. Lawncare, however, has seldom been directly linkedto elevated nutrient concentrations duringstorms. A very recent lakeshore study notedthat phosphorus concentrations were higher infertilized lawns compared to unfertilizedlawns, but no significant difference was notedfor nitrogen (Garn, 2002).

Wash-off of deposited nutrients from IC isthought to be a major source of nitrogen andphosphorus during storms (MWCOG, 1983).While the concentration of nitrogen andphosphorus from parking lots and streets islower than lawns, the volume of runoff issignificantly higher. In two studies using theSLAMM source loading model (Pitt andVoorhees, 1989), parking lots and streets wereresponsible for over 30% of the nitrogen andwere second behind lawns in their contribu-tions to the phosphorus load (Steuer et al.,1997; Waschbusch et al., 2000).

Source Area Total N (mg/l) Total P (mg/l)

Source (1) (1) (2) (3)

Commercial Parking Lot 1.94 0.20 N/R 0.10

High Traffic Street 2.95 0.31 0.47 0.18

Med. Traffic Street 1.62 0.23 1.07 0.22

Low Traffic Street 1.17 0.14 1.31 0.40

Commercial Rooftop 2.09 0.09 0.20 0.13

Residential Rooftop 1.46 0.06 0.15 0.07

Residential Driveway 2.10 0.35 1.16 N/R

Residential Lawn 9.70 2.33 2.67 0.79

Basin Outlet 1.87 0.29 0.66 N/R

(1) Steuer et al., 1997; (2) Bannerman et al., 1993; (3) Waschbusch et al., 2000; N/R= Not Reported

Table 29: Source Area Monitoring Data for Total Nitrogenand Total Phosphorous in Urban Areas



Streambank erosion also appears to be a majorsource of nitrogen and phosphorus in urbanstreams. Both nitrogen and phosphorus areoften attached to eroded bank sediment, asindicated in a recent study by Dartiguenave etal. (1997) in Austin, Texas. They showed thatchannel erosion contributed nearly 50% of theTotal P load shown for subwatersheds with IClevels between 10 and 60 % (Figure 34). Thesefindings suggest that prevention or reduction ofdownstream channel erosion may be animportant nutrient reduction strategy for urbanwatersheds.

Snowmelt runoff generally has higher nutrientEMCs, compared to stormwater runoff. Oberts(1994) found that TKN and nitrate EMCs weremuch higher in snowmelt at all sites. The samepattern has also been observed for phosphorusEMCs during snowmelt and stormwater runoff.Zapf-Gilje et al. (1986) found that the first

20% of snowmelt events contained 65% of thephosphorus and 90% of the nitrogen load.Ayers et al. (1985) reported that a higherpercentage of the annual nitrate, TKN andphosphorus load was derived from snowmeltrunoff compared to stormwater runoff in anurban Minnesota watershed, which presumablyreflects the accumulation of nutrients in thesnowpack during the winter.

Figure 34: Total Phosphorus from Bank Erosion as a Function of IC in Texas Streams(Dartiguenave et al., 1997)



Metal DetectionFrequency(1)(1)

EMCs(Fg/l)

Numberof

Events Source

Mean Median

Zinc 94%162 129 2234 Smullen and Cave, 1998

176 140 1281 USEPA, 1983

Copper 91%13.5 11.1 1657 Smullen and Cave, 1998

66.6 54.8 849 USEPA, 1983

Lead 94%67.5 50.7 2713 Smullen and Cave, 1998

175( 2) 131 (2) 1579 USEPA, 1983

Cadmium 48%

0.7 N/R 150 USEPA, 1983

0.5 N/R 100 USEPA, 1993

N/R0.75 R0.96 C2.1 I

30 Baird et al., 1996

3 I1U

N/R 9 Doerfer and Urbonas, 1993

Chromium 58%

4 N/R 32 Baird et al., 1996

N/R2.1 R10 C7 I


N/R 7 164 Bannerman et al., 1993

N/R = Not Reported; R- Residential, C- Commercial, I- Industrial; (1) as reprinted in USEPA, 1983; (2) Lead levels havedeclined over time with the introduction of unleaded gasoline

Metal DetectionFrequency(1)(1)

EMCs(Fg/l)

Numberof

Events Source

Mean Median

Zinc 94%162 129 2234 Smullen and Cave, 1998

176 140 1281 USEPA, 1983

Copper 91%13.5 11.1 1657 Smullen and Cave, 1998

66.6 54.8 849 USEPA, 1983

Lead 94%67.5 50.7 2713 Smullen and Cave, 1998

175( 2) 131 (2) 1579 USEPA, 1983

Cadmium 48%

0.7 N/R 150 USEPA, 1983

0.5 N/R 100 USEPA, 1993

N/R0.75 R0.96 C2.1 I


3 I1U

N/R 9 Doerfer and Urbonas, 1993

Chromium 58%

4 N/R 32 Baird et al., 1996

N/R2.1 R10 C7 I


N/R 7 164 Bannerman et al., 1993

N/R = Not Reported; R- Residential, C- Commercial, I- Industrial; (1) as reprinted in USEPA, 1983; (2) Lead levels havedeclined over time with the introduction of unleaded gasoline

4.6 Trace Metals

Many trace metals can be found at potentiallyharmful concentrations in urban stormwater.Certain metals, such as zinc, copper, lead,cadmium and chromium, are consistentlypresent at concentrations that may be ofconcern. These metals primarily result fromthe use of motor vehicles, weathering of metalsand paints, burning of fossil fuels and atmo-spheric deposition.

Metals are routinely reported as the totalrecoverable form or the dissolved form. Thedissolved form refers to the amount of metaldissolved in the water, which excludes metals

attached to suspended particles that cannotpass through a 0.45 micron filter. Total recov-erable refers to the concentration of an unfil-tered sample that is treated with hot dilutemineral acid. In general, the toxicity of metalsis related more to the dissolved form than therecoverable form.


Stormwater EMCs for zinc, copper, lead,cadmium and chromium vary regionally andare reviewed in Table 30. Regional differencesin trace metal concentrations and water qualitystandard exceedence appears to be related toclimate. In general, drier regions often have a

Table 30: EMCs and Detection Frequency for Metals in Urban Stormwater



higher risk of exceeding trace metal concentra-tion standards.

Crunkilton et al. (1996) measured recoverableand dissolved metals concentrations in LincolnCreek, Wisconsin and found higher EMCsduring storm events compared to baseflowperiods (Table 31). They also found that totalrecoverable metal concentrations were almostalways higher than the dissolved concentration(which is the more available form).

4.6.2 Impacts of Trace Metalson Streams

Although a great deal is known about theconcentration of metals in urban stormwater,much less is known about their possibletoxicity on aquatic biota. The primary concernrelated to the presence of trace metals instreams is their potential toxicity to aquaticorganisms. High concentrations can lead tobioaccumulation of metals in plants andanimals, possible chronic or acute toxicity, andcontamination of sediments, which can affectbottom dwelling organisms (Masterson andBannerman, 1994). Generally, trace metalconcentrations found in urban stormwater arenot high enough to cause acute toxicity (Fieldand Pitt, 1990). The cumulative accumulationof trace metal concentrations in bottom sedi-ments and animal tissues are of greater con-cern. Some evidence exists for trace metalaccumulation in bottom sediments of receivingwaters and for bioaccumulation in aquaticspecies (Bay and Brown, 2000 and Livingston,1996).

Relatively few studies have examined thechronic toxicity issue. Crunkilton et al. (1996)found that concentrations of lead, zinc andcopper exceeded EPA’s Chronic ToxicityCriteria more than 75% of the time instormflow in stormwater samples for LincolnCreek in Wisconsin. When exposed to stormand base flows in Lincoln Creek, Ceriodaphniadubia, a common invertebrate test species,demonstrated significant mortality in extendedflow-through tests. Around 30% mortality wasrecorded after seven days of exposure and 70%mortality was recorded after 14 days.

Crunkilton et al. (1996) also found that signifi-cant mortality in bullhead minnows occurred inonly 14% of the tests by the end of 14 days,but mortality increased to 100% during expo-sures of 17 to 61 days (see Table 32). In arelated study in the same watershed, Mastersonand Bannerman (1994) determined that cray-fish in Lincoln Creek had elevated levels oflead, cadmium, chromium and copper whencompared to crayfish from a reference stream.The Lincoln Creek research provides limitedevidence that prolonged exposure to tracemetals in urban streams may result in signifi-cant toxicity.

Most toxicity research conducted on urbanstormwater has tested for acute toxicity over ashort period of time (two to seven days).Shorter term whole effluent toxicity protocolsare generally limited to seven days (Crunkiltonet al., 1996). Research by Ellis (1986) reporteddelayed toxicity in urban streams. Field andPitt (1990) demonstrated that pollutantsdeposited to the stream during storm events

Total Recoverable Dissolved

Metal (Fg/l) Storm Flow Baseflow Storm Flow Baseflow

Lead 35 3 1.7 1.2

Zinc 133 22 13 8

Copper 23 7 5 4

Cadmium 0.6 0.1 0.1 0.1

Table 31: Average Total Recoverable and Dissolved Metals for 13 Stormwater Flowsand Nine Baseflow Samples from Lincoln Creek in 1994 (Crunkilton et al., 1996)



may take upwards of 10 to 14 days to exertinfluence. The research suggests that longerterm in-situ and flow-through monitoring areneeded to definitively answer the questionwhether metal levels in stormwater can bechronically toxic.

An additional concern is that trace metals co-occur with other pollutants found in urbanstormwater, and it is not clear whether theyinteract to increase or decrease potentialtoxicity. Hall and Anderson (1988) investi-gated the toxicity and chemical composition ofurban stormwater runoff in British Columbiaand found that the interaction of pollutantschanged the toxicity of some metals. In labora-tory analysis with Daphnia pulex, an aquaticinvertebrate, they found that the toxicity ofiron was low and that its presence reduced thetoxicity of other metals. On the other hand, thepresence of lead increased the toxicity ofcopper and zinc.

Interaction with sediment also influences theimpact of metals. Often, over half of the tracemetals are attached to sediment (MWCOG,1983). This effectively removes the metalsfrom the water column and reduces the avail-ability for biological uptake and subsequentbioaccumulation (Gavin and Moore, 1982 andOWML, 1983). However, metals accumulatedin bottom sediment can then be resuspendedduring storms (Heaney and Huber, 1978). It is

important to note that the toxic effect of metalscan be altered when found in conjunction withother substances. For instance, the presence ofchlorides can increase the toxicity of somemetals. Both metals and chlorides are commonpollutants in snowpacks (see section 4.2 formore snow melt information).

4.6.3 Sources and Source Areas ofTrace Metals

Research conducted in the Santa Clara Valleyof California suggests that cars can be thedominant loading source for many metals ofconcern, such as cadmium, chromium, copper,lead, mercury and zinc (EOA, Inc., 2001).Other sources are also important and includeatmospheric deposition, rooftops and runofffrom industrial and residential sites.

The sources and source areas for zinc, copper,lead, chromium and cadmium are listed inTable 33. Source areas for trace metals in theurban environment include streets, parkinglots, snowpacks and rooftops. Copper is oftenfound in higher concentrations on urbanstreets, because some vehicles have brake padsthat contain copper. For example, the SantaClara study estimated that 50% of the totalcopper load was due to brake pad wear (Wood-ward-Clyde, 1992). Sources of lead includeatmospheric deposition and diesel fuel emis-sions, which frequently occur along rooftops

Species Effect Percent of Tests with Significant (p<0.05) Toxic Effects as

Compared to Controls According to Exposure

48 hours 96 hours 7 days 14 days 17-61days

D. magna Mortality 0 N/R 36% 93% N/R

ReducedReproduction 0 N/R 36% 93% N/R

P. promelas Mortality N/R 0 0 14% 100%

ReducedBiomass

N/R N/R 60% 75% N/R

N/R = Not Reported

Table 32: Percentage of In-situ Flow-through Toxicity Tests Using Daphnia magna andPimephales promelas with Significant Toxic Effects from Lincoln Creek (Crunkilton et al., 1996)



and streets. Zinc in urban environments is aresult of the wear of automobile tires (esti-mated 60% in the Santa Clara study), paints,and weathering of galvanized gutters anddownspouts. Source area concentrations oftrace metals are presented in Table 34. Ingeneral, trace metal concentrations vary

Source Area DissolvedZinc

TotalZinc

DissolvedCopper

TotalCopper Dissolved Lead Total Lead

Source (1) (2) (1) (2) (2) (1) (3) (1) (3) (2)

CommercialParking Lot

64 178 10.7 9 15 N/R N/R 40 N/R 22

High TrafficStreet

73 508 11.2 18 46 2.1 1.7 37 25 50

Medium TrafficStreet

44 339 7.3 24 56 1.5 1.9 29 46 55

Low Traffic Street 24 220 7.5 9 24 1.5 .5 21 10 33

CommercialRooftop

263 330 17.8 6 9 20 N/R 48 N/R 9

ResidentialRooftop

188 149 6.6 10 15 4.4 N/R 25 N/R 21

ResidentialDriveway 27 107 11.8 9 17 2.3 N/R 52 N/R 17

Residential Lawn N/R 59 N/R 13 13 N/R N/R N/R N/R N/R

Basin Outlet 23 203 7.0 5 16 2.4 N/R 49 N/R 32

Sources: (1) Steuer et al., 1997; (2) Bannerman et al., 1993; (3) Waschbusch, 2000; N/R = Not Reported

Table 34: Metal Source Area Concentrations in the Urban Landscape (FFFFFg/l)

considerably, but the relative rank amongsource areas remains relatively constant. Forexample, a source loading model developed foran urban watershed in Michigan estimated thatparking lots, driveways and residential streetswere the primary source areas for zinc, copperand cadmium loads (Steuer et al., 1997).

Metal Sources Source Area Hotspots

Zinc tires, fuel combustion, galvanized pipes, roofs andgutters, road salts *estimate of 60% from tires

parking lots, commercial andindustrial rooftops, and streets

Copper auto brake linings, pipes and fittings, algacides, andelectroplating *estimate of 50% from brake pad wear

parking lots, commercial roofsand streets

Lead diesel fuel, paints and stains parking lots, rooftops, and streets

Cadmium component of motor oil and corrodes from alloys andplated surfaces

parking lots, rooftops, and streets

Chromium found in exterior paints and corrodes from alloys andplated surfaces

most frequently found in industrialand commercial runoff

Sources: Bannerman et al., 1993; Barr, 1997; Steuer et al., 1997; Good, 1993; Woodward - Clyde, 1992

Table 33: Metal Sources and Source Area “Hotspots” in Urban Areas



4.7 Hydrocarbons:PAH, Oil and Grease

Hydrocarbons are petroleum-based substancesand are found frequently in urban stormwater.The term “hydrocarbons” is used to refer tomeasurements of oil and grease and polycy-clic-aromatic hydrocarbons (PAH). Certaincomponents of hydrocarbons, such as pyreneand benzo[b]fluoranthene, are carcinogens andmay be toxic to biota (Menzie-Cura , 1995).Hydrocarbons normally travel attached tosediment or organic carbon. Like many pollut-ants, hydrocarbons accumulate in bottomsediments of receiving waters, such as urbanlakes and estuaries. Relatively few studies havedirectly researched the impact of hydrocarbonson streams.


Table 35 summarizes reported EMCs of PAHand oil and grease derived from storm eventmonitoring at three different areas of the U.S.The limited research on oil and grease concen-trations in urban runoff indicated that thehighest concentrations were consistently foundin commercial areas, while the lowest werefound in residential areas.

4.7.2 Impacts of Hydrocarbonson Streams

The primary concern of PAH and oil andgrease on streams is their potentialbioaccumulation and toxicity in aquaticorganisms. Bioaccumulation in crayfish, clamsand fish has been reported by Masterson andBannerman (1994); Moring and Rose (1997);and Velinsky and Cummins (1994).

HydrocarbonIndicator

EMC Numberof Events

Source LocationMean

PAH (Fg/l)

3.2* 12 Menzie-Cura, 1995 MA

7.1 19 Menzie-Cura, 1995 MA

13.4 N/R Crunkilton et al., 1996 WI

Oil andGrease (mg/l)

1.7 R** 9 C3 I

30 Baird et al., 1996TX

3 N/R USEPA, 1983 U.S.



3.89 R13.13 C7.10 I

N/R Silverman et al., 1988 CA

2.35 R5.63 C4.86 I

107 Barr, 1997 MD

N/R = Not Reported; R = Residential, C = Commercial, I = Industrial; * = geometric mean, ** = median

HydrocarbonIndicator

EMC Numberof Events

Source LocationMean

PAH (Fg/l)



13.4 N/R Crunkilton et al., 1996 WI

Oil andGrease (mg/l)

1.7 R** 9 C3 I

30 Baird et al., 1996TX

3 N/R USEPA, 1983 U.S.



3.89 R13.13 C7.10 I

N/R Silverman et al., 1988 CA

2.35 R5.63 C4.86 I

107 Barr, 1997 MD

N/R = Not Reported; R = Residential, C = Commercial, I = Industrial; * = geometric mean, ** = median

Table 35: Hydrocarbon EMCs in Urban Areas



Moring and Rose (1997) also showed that notall PAH compounds accumulate equally inurban streams. They detected 24 different PAHcompounds in semi-permeable membranedevices (SPMDs), but only three PAH com-pounds were detected in freshwater clamtissue. In addition, PAH levels in the SPMDswere significantly higher than those reported inthe clams.

While acute PAH toxicity has been reported atextremely high concentrations (Ireland et al.,1996), delayed toxicity has also been found(Ellis, 1986). Crayfish from Lincoln Creek hada PAH concentration of 360 Fg/kg, muchhigher than the concentration thought to becarcinogenic (Masterson and Bannerman,1994). By comparison, crayfish in a non-urbanstream had undetectable PAH levels. Toxiceffects from PAH compounds may be limitedsince many are attached to sediment and maybe less available, with further reductionoccurring through photodegradation (Ireland etal., 1996).

The metabolic effect of PAH compounds onaquatic life is unclear. Crunkilton et al. (1996)found potential metabolic costs to organisms,but Masterson and Bannerman (1994) andMacCoy and Black (1998) did not. The long-term effect of PAH compounds in sediments ofreceiving waters remains a question for furtherstudy.

4.7.3 Sources and SourceAreas of Hydrocarbons

In most residential stormwater runoff, hydro-carbon concentrations are generally less than5mg/l, but the concentrations can increase tofive to 10 mg/l within some commercial,industrial and highway areas (See Table 35).Specific “hotspots” for hydrocarbons includegas stations, commuter parking lots, conve-nience stores, residential parking areas andstreets (Schueler and Shepp, 1993). Theseauthors evaluated hydrocarbon concentrationswithin oil and grease separators in the Wash-ington Metropolitan area and determined thatgas stations had significantly higher concentra-tions of hydrocarbons and trace metals, ascompared to other urban source areas. Sourcearea research in an urban catchment in Michi-gan showed that commercial parking lotscontributed 64% of the total hydrocarbon load(Steuer et al., 1997). In addition, highwayswere found to be a significant contributor ofhydrocarbons by Lopes and Dionne (1998).



4.8 Bacteria and Pathogens

Bacteria are single celled organisms that aretoo small to see with the naked eye. Of particu-lar interest are coliform bacteria, typicallyfound within the digestive system of warm-blooded animals. The coliform family ofbacteria includes fecal coliform, fecal strepto-cocci and Escherichia coli, which are consis-tently found in urban stormwater runoff. Theirpresence confirms the existence of sewage oranimal wastes in the water and indicates thatother harmful bacteria, viruses or protozoansmay be present, as well. Coliform bacteria areindicators of potential public health risks andnot actual causes of disease.

A pathogen is a microbe that is actually knownto cause disease under the right conditions.Two of the most common waterborne patho-gens in the U.S. are the protozoansCryptosporidium parvum and Giardia lambia.Cryptosporidium is a waterborne intestinalparasite that infects cattle and domesticanimals and can be transmitted to humans,

causing life-threatening problems in peoplewith impaired immune systems (Xiao et al.,2001). Giardia can cause intestinal problems inhumans and animals when ingested (Bagley etal., 1998). To infect new hosts, protozoanscreate hard casings known as oocysts(Cryptosporidium) or cysts (Giardia) that areshed in feces and travel through surface watersin search of a new host.


Concentrations of fecal coliform bacteria inurban stormwater typically exceed the 200MPN/100 ml threshold set for human contactrecreation (USGS, 2001b). Bacteria concentra-tions also tend to be highly variable from stormto storm. For example, a national summary offecal coliform bacteria in stormwater runoff isshown in Figure 35 and Table 36. The variabil-ity in fecal coliform ranges from 10 to 500,000MPN/100ml with a mean of 15,038 MPN/100ml (Schueler, 1999). Another nationaldatabase of more than 1,600 stormwater eventscomputed a mean concentration of 20,000

Figure 35: Fecal Coliform Levels in Urban Stormwater ( Schueler, 1999)

Fecal Coliform Levels in Urban Stormwater:A National Review

Stormwater runoff levels from 34 small catchments in13 monitoring studies conducted:

AL, AZ, ID, KY, MD, NC, NH, NY, SD, TN, TX, WA, WI



MPN/100ml for fecal coliform (Pitt, 1998).Fecal streptococci concentrations for 17 urbansites across the country had a mean of 35,351MPN/100ml (Schueler, 1999).

Young and Thackston (1999) showed thatbacteria concentrations at four sites in metroNashville were directly related to watershedIC. Increasing IC reflects the cumulativeincrease in potential bacteria sources in theurban landscape, such as failing septic systems,sewage overflows, dogs, and inappropriatedischarges. Other studies show that concentra-tions of bacteria are typically higher in urbanareas than rural areas (USGS, 1999a), but theyare not always directly related to IC. Forexample, Hydroqual (1996) found that concen-trations of fecal coliform in sevensubwatersheds of the Kensico watershed inNew York were generally higher for moredeveloped basins, but fecal coliform concentra-

tions did not directly increase with IC in thedeveloped basins (Figure 36).

There is some evidence that higher concentra-tions of coliform are found in arid or semi-aridwatersheds. Monitoring data from semi-aridregions in Austin, San Antonio, and CorpusChristi, Texas averaged 61,000, 37,500 and40,500 MPN/100ml, respectively (Baird etal.,1996 and Chang et al. 1990). Schiff (1996),in a report of Southern California NPDESmonitoring, found that median concentrationsof fecal coliform in San Diego were 50,000MPN/100ml and averaged 130,000 MPN/100ml in Los Angeles. In all of these arid andsemi-arid regions, concentrations were signifi-cantly higher than the national average of15,000 to 20,000 MPN/100ml.

Bacteria Type

EMCs(MPN/100ml) Number of

EventsSource Location

Mean

Fecal Coliform

15,038 34 Schueler, 1999 U.S.

20,000 1600 Pitt, 1998 U.S.

7,653 27Thomas and McClelland,

1995 GA

20,000 R* 6900 C 9700 I

30* Baird et al., 1996 TX

77,970 21 watersheds Chang et al., 1990 TX

4,500 189 Varner, 1995 WA

23,500 3Young and Thackston,

1999 TN

Fecal Strep

35,351 17 Schueler, 1999 U.S.

28,864 R 27 Thomas and McClelland,1995

GA

56,000 R *18,000 C 6,100 I


N/R = Not Reported, R = Residential Area, C = Commercial Area, I = Industrial Area, * = Median

Bacteria Type

EMCs(MPN/100ml) Number of

EventsSource Location

Mean

Fecal Coliform

15,038 34 Schueler, 1999 U.S.

20,000 1600 Pitt, 1998 U.S.

7,653 27Thomas and McClelland,

1995 GA

20,000 R* 6900 C 9700 I


77,970 21 watersheds Chang et al., 1990 TX

4,500 189 Varner, 1995 WA

23,500 3Young and Thackston,

1999 TN

Fecal Strep

35,351 17 Schueler, 1999 U.S.

28,864 R 27 Thomas and McClelland,1995

GA

56,000 R *18,000 C 6,100 I


N/R = Not Reported, R = Residential Area, C = Commercial Area, I = Industrial Area, * = Median

Table 36: Bacteria EMCs in Urban Areas



Concentrations of Cryptosporidium andGiardia in urban stormwater are shown inTable 37. States et al. (1997) found highconcentrations of Cryptosporidium and Giar-dia in storm samples from a combined sewer inPittsburgh (geometric mean 2,013 oocysts/100ml and 28,881 cysts/100ml). There isevidence that urban stormwater runoff mayhave higher concentrations of Cryptosporidiumand Giardia than other surface waters, asreported in Table 38 (Stern, 1996). Bothpathogens were detected in about 50% of urbanstormwater samples, suggesting some concernfor drinking water supplies.

4.8.2 Impacts of Bacteria andPathogens on Streams

Fecal coliform bacteria indicate the potentialfor harmful bacteria, viruses, or protozoans andare used by health authorities to determinepublic health risks. These standards wereestablished to protect human health based onexposures to water during recreation anddrinking. Bacteria standards for various wateruses are presented in Table 39 and are alleasily exceeded by typical urban stormwaterconcentrations. In fact, over 80,000 miles ofstreams and rivers are currently in non-attain-

Pathogens Units EMCs Number

of EventsSource

Mean Median

Cryptosporidium oocysts 37.2 3.9 78 Stern, 1996

oocysts/100ml 2013 N/R N/R States et al., 1997

Giardia cysts 41.0 6.4 78 Stern, 1996

cysts/100ml 28,881 N/R N/R States et al., 1997

N/R= Not reported

Pathogens Units EMCs Number

of EventsSource

Mean Median

Cryptosporidium oocysts 37.2 3.9 78 Stern, 1996

oocysts/100ml 2013 N/R N/R States et al., 1997

Giardia cysts 41.0 6.4 78 Stern, 1996

cysts/100ml 28,881 N/R N/R States et al., 1997

N/R= Not reported

Table 37: Cryptosporidium and Giardia EMCs

Figure 36: Relationship Between IC and Fecal Coliform Concentrations inNew York Streams (Hydroqual, 1996)



ment status because of high fecal coliformlevels (USEPA, 1998).

4.8.3 Sources and Source Areas ofBacteria and Pathogens

Sources of coliform bacteria include wastefrom humans and wildlife, including livestockand pets. Essentially, any warm-bloodedspecies that is present in significant numbers ina watershed is a potential culprit. Sourceidentification studies, using methods such asDNA fingerprinting, have put the blame onspecies such as rats in urban areas, ducks andgeese in stormwater ponds, livestock from

hobby farms, dogs and even raccoons(Blankenship, 1996; Lim and Olivieri, 1982;Pitt, 1998; Samadpour and Checkowitz, 1998).

Transport of bacteria takes place through directsurface runoff, direct inputs to receivingwaters, or indirect secondary sources. Sourceareas in the urban environment for directrunoff include lawns and turf, driveways,parking lots and streets. For example, dogshave high concentrations of fecal coliform intheir feces and have a tendency to defecate inclose proximity to IC (Schueler, 1999).Weiskel et al. (1996) found that direct inputsof fecal coliform from waterfowl can be very

Source WaterSampled

Number ofSources/

Number ofSamples

Percent Detection

TotalGiardia

ConfirmedGiardia

TotalCryptosporidium

ConfirmedCryptosporidium

WastewaterEffluent 8/147 41.5% 12.9% 15.7% 5.4%

UrbanSubwatershed

5/78 41.0% 6.4% 37.2% 3.9%

AgriculturalSubwatershed 5/56 30.4% 3.6% 32.1% 3.6%

UndisturbedSubwatershed

5/73 26.0% 0.0% 9.6% 1.4%

Source WaterSampled

Number ofSources/

Number ofSamples

Percent Detection

TotalGiardia

ConfirmedGiardia

TotalCryptosporidium

ConfirmedCryptosporidium

WastewaterEffluent 8/147 41.5% 12.9% 15.7% 5.4%

UrbanSubwatershed

5/78 41.0% 6.4% 37.2% 3.9%

AgriculturalSubwatershed 5/56 30.4% 3.6% 32.1% 3.6%

UndisturbedSubwatershed

5/73 26.0% 0.0% 9.6% 1.4%

Water Use Microbial Indicator Typical Water Standard

Water Contact Recreation Fecal Coliform <200 MPN per 100ml

Drinking Water Supply Fecal Coliform <20 MPN per 100ml

Shellfish Harvesting Fecal Coliform <14 MPN/ 100ml

Treated Drinking Water Total ColiformNo more than 1% coliform positive

samples per month

Freshwater Swimming E.Coli <126 MPN per 100ml

Important Note: Individual state standards may employ different sampling methods, indicators, averaging periods,averaging methods, instantaneous maximums and seasonal limits. MPN = most probable number. Higher or lowerlimits may be prescribed for different water use classes.

Table 39: Typical Coliform Standards for Different Water Uses (USEPA, 1998)

Table 38: Percent Detection of Giardia cysts and Cryptosporidium oocysts inSubwatersheds and Wastewater Treatment Plant Effluent in the

New York City Water Supply Watersheds (Stern, 1996)



important; these inputs accounted for as muchas 67% of the annual coliform load to Butter-milk Bay, Massachusetts.

Indirect sources of bacteria include leakingseptic systems, illicit discharges, sanitarysewer overflows (SSOs), and combined seweroverflows (CSOs). These sources have thepotential to deliver high coliform concentra-tions to urban streams. In fact, extremely highbacteria concentrations are usually associatedwith wastewater discharges. CSOs and SSOsoccur when the flow into the sewer exceeds thecapacity of the sewer lines to drain them. CSOsresult from stormwater flow in the lines, andSSOs are a result of infiltration problems orblockages in the lines.

Illicit connections from businesses and homesto the storm drainage system can dischargesewage or washwater into receiving waters.Illicit discharges can often be identified bybaseflow sampling of storm sewer systems.Leaking septic systems are estimated tocomprise between 10 and 40% of the systems,and individual inspections are the best way todetermine failing systems (Schueler, 1999).

There is also evidence that coliform bacteriacan survive and reproduce in stream sedimentsand storm sewers (Schueler, 1999). During astorm event, they often become resuspendedand add to the in-stream bacteria load. Sourcearea studies reported that end of pipe concen-trations were an order of magnitude higherthan any source area on the land surface;therefore, it is likely that the storm sewersystem itself acts as a source of fecal coliform(Bannerman et al., 1993 and Steuer et al.,1997). Resuspension of fecal coliform fromfine stream sediments during storm events hasbeen reported in New Mexico (NMSWQB,1999). The sediments in-stream and in thestorm sewer system may be significantcontributors to the fecal coliform load.

Sources of Cryptosporidium and Giardiainclude human sewage and animal feces.Cryptosporidium is commonly found in cattle,dogs and geese. Graczyk et al. (1998) foundthat migrating Canada geese were a vector forCryptosporidium and Giardia, which hasimplications for water quality in urban pondsthat support large populations of geese.



4.9 Organic Carbon

Total organic carbon (TOC) is often used as anindicator of the amount of organic matter in awater sample. Typically, the more organicmatter present in water, the more oxygenconsumed, since oxygen is used by bacteria inthe decomposition process. Adequate levels ofdissolved oxygen in streams and receivingwaters are important because they are criticalto maintain aquatic life. Organic carbon isroutinely found in urban stormwater, and highconcentrations can result in an increase inBiochemical Oxygen Demand (BOD) andChemical Oxygen Demand (COD). BOD andCOD are measures of the oxygen demandcaused by the decay of organic matter.


Urban stormwater has a significant ability toexert a high oxygen demand on a stream orreceiving water, even two to three weeks afteran individual storm event (Field and Pitt,1990). Average concentrations of TOC, BODand COD in urban stormwater are presented inTable 40. Mean concentrations of TOC, BODand COD during storm events in nationwidestudies were 17 mg/l, 14.1 mg/l and 52.8 mg/l,respectively (Kitchell, 2001 and Smullen andCave,1998).

4.9.2 Impacts of OrganicCarbon on Streams

TOC is primarily a concern for aquatic lifebecause of its link to oxygen demand in

streams, rivers, lakes and estuaries. The initialeffect of increased concentrations of TOC,BOD or COD in stormwater runoff may be adepression in oxygen levels, which may persistfor many days after a storm, as depositedorganic matter gradually decomposes (Fieldand Pitt, 1990).

TOC is also a concern for drinking waterquality. Organic carbon reacts with chlorineduring the drinking water disinfection processand forms trihalomethanes and other disinfec-tion by-products, which can be a seriousdrinking water quality problem (Water, 1999).TOC concentrations greater than 2 mg/l intreated water and 4 mg/l in source water canresult in unacceptably high levels of disinfec-tion byproducts and must be treated to reduceTOC or remove the disinfection byproducts(USEPA, 1998). TOC can also be a carrier forother pollutants, such as trace metals, hydro-carbons and nutrients.

4.9.3 Sources and Source Areas ofTotal Organic Carbon

The primary sources of TOC in urban areasappear to be decaying leaves and other organicmatter, sediment and combustion by-products.Source areas include curbs, storm drains,streets and stream channels. Dartiguenave etal. (1997) determined that about half of theannual TOC load in urban watersheds ofAustin, TX was derived from the erodingstreambanks.

Organic Carbon SourceEMCs (mg/l) Number of

EventsSource

Mean Median

Total Organic Carbon (TOC)32.0 N/R 423 Barrett and Malina, 1998

17 15.2 19 studies Kitchell, 2001

Biological Oxygen Demand (BOD)14.1 11.5 1035 Smullen and Cave, 1998

10.4 8.4 474 USEPA, 1983

Chemical Oxygen Demand (COD)52.8 44.7 2639 Smullen and Cave, 1998

66.1 55 1538 USEPA, 1983

N/R = Not Reported

Organic Carbon SourceEMCs (mg/l) Number of

EventsSource

Mean Median

Total Organic Carbon (TOC)32.0 N/R 423 Barrett and Malina, 1998

17 15.2 19 studies Kitchell, 2001

Biological Oxygen Demand (BOD)14.1 11.5 1035 Smullen and Cave, 1998

10.4 8.4 474 USEPA, 1983

Chemical Oxygen Demand (COD)52.8 44.7 2639 Smullen and Cave, 1998

66.1 55 1538 USEPA, 1983

N/R = Not Reported

Table 40: EMCs for Organic Carbon in Urban Areas



4.10 MTBE

Methyl tertiary butyl-ether (MTBE) is avolatile organic compound (VOC) that isadded to gasoline to increase oxygen levels,which helps gas burn cleaner (called anoxygenate). MTBE has been used as a perfor-mance fuel additive since the 1970s. In 1990,the use of oxygenates was mandated by federallaw and concentrations of MTBE in gasolineincreased. Today, MTBE is primarily used inlarge metropolitan areas that experience airpollution problems. Since 1990, MTBE hasbeen detected at increasing levels in bothsurface water and groundwater and is one ofthe most frequently detected VOCs in urbanwatersheds (USGS, 2001a). EPA has declaredMTBE to be a potential human carcinogen athigh doses. In March 2000, a decision wasmade by EPA to follow California’s lead tosignificantly reduce or eliminate the use ofMTBE in gasoline.


MTBE is highly soluble in water and thereforenot easily removed once it enters surface orground water. Delzer (1999) detected the

presence of MTBE in 27% of the shallow wellsmonitored in eight urban areas across thecountry (Figure 37). Detection frequency wassignificantly higher in New England andDenver, as shown in Table 41. In a secondstudy conducted in 16 metropolitan areas,Delzer (1999) found that 83% of MTBEdetections occurred between October andMarch, the time when MTBE is primarily usedas a fuel additive. The median MTBE concen-tration was 1.5 ppb, well below EPA’s draftadvisory level of 20 ppb (Delzer, 1996).

4.10.2 Impacts of MTBE on Streams

The primary concerns regarding MTBE arethat it is a known carcinogen to small mam-mals, a suspected human carcinogen at higher

Figure 37: MTBE Concentrations in Surface Water from Eight Cities (Delzer, 1996)

Location DetectionFrequency

Source Year

211 shallow wells ineight urban areas

27% Delzer 1999

Surface watersamples in 16metro areas

7% Delzer 1996

Table 41: MTBE Detection Frequency



doses and may possibly be toxic to aquatic lifein small streams (Delzer, 1996). MTBE canalso cause taste and odor problems in drinkingwater at fairly low concentrations. EPA issueda Drinking Water Advisory in 1997 thatindicated that MTBE concentrations less than20 ppb should not cause taste and odor prob-lems for drinking water. However, the Asso-ciation of California Water Agencies reportsthat some consumers can detect MTBE atlevels as low as 2.5 ppb (ACWA, 2000).Because MTBE is frequently found in ground-water wells, it is thought to be a potentialthreat to drinking water (Delzer, 1999). Forexample, Santa Monica, California reportedlylost half of its groundwater drinking watersupply due to MTBE contamination (Bay andBrown, 2000). MTBE has also been detected inhuman blood, especially in people frequentlyexposed to gasoline, such as gas stationattendants (Squillace et al., 1995).

4.10.3 Sources and SourceAreas of MTBE

Since MTBE is a gasoline additive, its poten-tial sources include any area that produces,transports, stores, or dispenses gasoline,particularly areas that are vulnerable to leaksand spills. Leaking underground storage tanksare usually associated with the highest MTBEconcentrations in groundwater wells (Delzer,1999). Vehicle emissions are also an importantsource of MTBE. Elevated levels are fre-quently observed along road corridors anddrainage ditches. Once emitted, MTBE cantravel in stormwater runoff or groundwater.Main source areas include heavily used multi-lane highways. Gas stations may also be ahotspot source area for MTBE contamination.

Another potential source of MTBE is water-craft, since two cycle engines can discharge asmuch as 20 to 30% of their fuel through theexhaust (Boughton and Lico, 1998). MTBEconcentrations are clearly associated withincreased use of gas engines, and there isconcern that MTBE is an increasing compo-nent of atmospheric deposition (Boughton andLico, 1998 and UC Davis, 1998).



4.11 Pesticides

Pesticides are used in the urban environment tocontrol weeds, insects and other organisms thatare considered pests. EPA estimates that nearly70 million pounds of active pesticide ingredi-ents are applied to urban lawns each year asherbicides or insecticides. Herbicides are usedon urban lawns to target annual and perennialbroadleaf weeds, while insecticides are used tocontrol insects. Many types of pesticides areavailable for use in urban areas. Immermanand Drummond (1985) report that 338 differ-

ent active ingredients are applied to lawns andgardens nationally. Each pesticide varies inmobility, persistence and potential aquaticimpact. At high levels, many pesticides havebeen found to have adverse effects on ecologi-cal and human health. Several recent researchstudies by the USGS have shown that insecti-cides are detected with the greatest frequencyin urban streams, and that pesticide detectionfrequency increases in proportion to thepercentage of urban land in a watershed(Ferrari et al., 1997; USGS, 1998, 1999a-b,2001b). A national assessment by the USGS

Pollutant DetectionFrequency

MedianConcentration (Fg/l)

Number ofSamples

Source

Insecticides

Diazinon

75% 0.025 326 USGS, 1998b

92% 0.55 76 Brush et al., 1995

17% 0.0021795

Ferrari et al., 1997

Chlorpyrifos41% Non Detect 327 USGS, 1998b

14% 0.004 1218 Brush et al., 1995

Carbaryl 46% Non Detect 327 USGS, 1998b

22% 0.003 1128 Ferrari et al., 1997

Herbicides

Atrazine86% 0.023 327 USGS, 1998b

72% 0.099 2076 Ferrari et al., 1997

Prometon84% 0.031 327 USGS, 1998b

56% 0.029 1531 Ferrari et al., 1997

Simazine88% 0.039 327 USGS, 1998b

17% 0.046 1995 Ferrari et al., 1997

2,4 -D 67% 1.1 11 Dindorf, 1992

17% 0.035 786 Ferrari et al., 1997

Dicamba 22% 1.8 4 Dindorf, 1992

MCPP 56% 1.8 10 Dindorf, 1992

MCPA 28% 1.0 5 Dindorf, 1992

Pollutant DetectionFrequency

MedianConcentration (Fg/l)

Number ofSamples

Source

Insecticides

Diazinon

75% 0.025 326 USGS, 1998b

92% 0.55 76 Brush et al., 1995

17% 0.0021795

Ferrari et al., 1997

Chlorpyrifos41% Non Detect 327 USGS, 1998b

14% 0.004 1218 Brush et al., 1995

Carbaryl 46% Non Detect 327 USGS, 1998b

22% 0.003 1128 Ferrari et al., 1997

Herbicides

Atrazine86% 0.023 327 USGS, 1998b

72% 0.099 2076 Ferrari et al., 1997

Prometon84% 0.031 327 USGS, 1998b

56% 0.029 1531 Ferrari et al., 1997

Simazine88% 0.039 327 USGS, 1998b

17% 0.046 1995 Ferrari et al., 1997

2,4 -D 67% 1.1 11 Dindorf, 1992

17% 0.035 786 Ferrari et al., 1997

Dicamba 22% 1.8 4 Dindorf, 1992

MCPP 56% 1.8 10 Dindorf, 1992

MCPA 28% 1.0 5 Dindorf, 1992

Table 42: Median Concentrations and Detection Frequency of Herbicides andInsecticides in Urban Streams



(2001a) also indicates that insecticides areusually detected at higher concentrations inurban streams than in agricultural streams.


Median concentrations and detection frequencyfor common pesticides are shown in Table 42.Herbicides that are frequently detected inurban streams include atrazine; simazine;prometon; 2,4-D; dicamba; MCPP; andMCPA. Insecticides are also frequently en-countered in urban streams, includingdiazinon, chlorpyrifos, malathion, and car-baryl. A USGS (1996) study monitored 16sites in Gills Creek in Columbia, South Caro-lina over four days. This study reported thatpesticide detection frequency increased aspercent urban land increased.

Wotzka et al. (1994) monitored herbicidelevels in an urban stream in Minneapolis,Minnesota during more than 40 storms. Theyfound herbicides, such as 2,4-D; dicamba;MCPP; and MCPA in 85% of storm runoffevents sampled. Total herbicide EMCs rangedfrom less than one to 70 µg/l. Ferrari et al.(1997) analyzed 463 streams in the mid-Atlantic region for the presence of 127 pesti-cide compounds. At least one pesticide wasdetected at more than 90% of the streamssampled.

Diazinon is one of the most commonly de-tected insecticides in urban stormwater runoffand dry weather flow. Diazinon was detectedin 75% of National Water Quality Assessment(NAWQA) samples, 92% of stormflowsamples from Texas, and 100% of urbanstormflow samples in King County, Washing-ton (Brush et al., 1995 and USGS, 1999b).Diazinon is most frequently measured atconcentrations greater than freshwater aquaticlife criteria in urban stormwater (USGS,1999a). USGS reports that diazinon concentra-tions were generally higher during urbanstormflow (Ferrari et al., 1997).

4.11.2 Impacts of Pesticideson Streams

Many pesticides are known or suspectedcarcinogens and can be toxic to humans andaquatic species. However, many of the knownhealth effects require exposure to higherconcentrations than typically found in theenvironment, while the health effects ofchronic exposure to low levels are generallyunknown (Ferrari et al., 1997).

Studies that document the toxicity of insecti-cides and herbicides in urban stormwater havebeen focused largely on diazinon. Diazinon isresponsible for the majority of acute toxicity instormwater in Alameda County, California andKing County, Washington (S.R. Hansen &Associates, 1995). Concentrations of diazinonin King County stormwater frequently exceedthe freshwater aquatic life criteria (Figure 38).Similarly, research on Sacramento, Californiastreams revealed acute toxicity for diazinon in100% of stormwater samples usingCeriodaphnia as the test organism (Connor,1995). Diazinon has a half-life of 42 days andis very soluble in water, which may explain itsdetection frequency and persistence in urbanstormwater. Diazinon is also reported to attachfairly readily to organic carbon; consequently,it is likely re-suspended during storm events.

Insecticide concentrations exceeding acute andchronic toxicity thresholds for test organismssuch as Ceriodaphnia have frequently beenfound in urban stormwater in New York,Texas, California, and Washington (Scanlinand Feng, 1997; Brush et al., 1995; USGS,1999b). The possibility exists that pesticidescould have impacts on larger bodies of water,but there is a paucity of data on the subject atthis time.



4.11.3 Sources and SourceAreas of Pesticides

Sources for pesticides in urban areas includeapplications by homeowners, landscapingcontractors and road maintenance crews.Source areas for pesticides in urban areasinclude lawns in residential areas; managedturf, such as golf courses, parks, and ballfields; and rights-of-way in nonresidentialareas. Storage areas, which are subject to spillsand leaks, can also be a source area. A study inSan Francisco was able to trace high diazinonconcentrations in some streams back to just a

few households which had applied thepesticide at high levels (Scanlin and Feng,1997). Two herbicides, simazine and atra-zine, were detected in over 60% of samplesin King County, WA stormwater but werenot identified as being sold in retail stores. Itis likely these herbicides are applied tononresidential areas such as rights-of-way,parks and recreational areas (USGS, 1999b).Because pesticides are typically applied toturf, IC is not a direct indicator for pesticideconcentrations, although they can drift ontopaved surfaces and end up in stormwaterrunoff.

Figure 38: Concentrations of Pesticides in Stormwater in King County, WA(S.R. Hansen & Associates, 1995 and USGS, 1999b)



4.12 Deicers

Deicers are substances used to melt snow andice to keep roads and walking areas safe. Themost commonly used deicer is sodium chlo-ride, although it may also be blended withcalcium chloride or magnesium chloride. Otherless frequently used deicers include urea andglycol, which are primarily used at airports todeice planes. Table 43 summarizes the compo-sition, use and water quality effects of commondeicers.

Chlorides are frequently found in snowmeltand stormwater runoff in most regions thatexperience snow and ice in the winter months(Oberts, 1994 and Sherman, 1998). Figure 39shows that the application of deicer salts hasincreased since 1940 from 200,000 tons to 10to 20 million tons per year in recent years (SaltInstitute, 2001). Several U.S. and Canadianstudies indicate severe inputs of road salts onwater quality and aquatic life (EnvironmentCanada, 2001 and Novotny et al., 1999).

Figure 39: U.S. Highway Salt Usage Data (Salt Institute, 2001)

Deicer Description Use Water Quality Effect

Chlorides

Chloride baseddeicer usually

combined with Na,Ca or Mg

Road Deicer andResidential Use

Cl complexes can release heavymetals, affect soil permeability,impacts to drinking water, potentialtoxic effects to small streams

Urea Nitrogen-basedfertilizer product

Used asalternative to

glycol

Increased nitrogen in water andpotential toxicity to organisms

EthyleneGlycol

Petroleum basedorganic compounds,similar to antifreeze

Used at airportsfor deicing planes

Toxicity effects, high BOD and COD,hazardous air pollutant

Ta Table 43: Use and Water Quality Effect of Snowmelt Deicers(Ohrel, 1995; Sills and Blakeslee, 1992)



Form ofRunoff

EMCs (mg/l) Number ofEvents

Sources LocationMean

Snowmelt

116* 49 Oberts, 1994 MN

2119 N/R Sherman, 1998 Ontario

1267 R474 U

N/R Novotny et al., 1999 NY

1612 N/RMasterson and Bannerman,

1994 WI

397 282 Environment Canada, 2001Ontario,Canada

Non-winterStormEvent

42 61 Brush et al., 1995 TX


40.5 N/RMasterson and Bannerman,

1994 WI

N/R = Not Reported, R = residential, U = urban, * = Median

Form ofRunoff

EMCs (mg/l) Number ofEvents

Sources LocationMean

Snowmelt

116* 49 Oberts, 1994 MN


1267 R474 U

N/R Novotny et al., 1999 NY

1612 N/RMasterson and Bannerman,

1994 WI

397 282 Environment Canada, 2001Ontario,Canada

Non-winterStormEvent

42 61 Brush et al., 1995 TX


40.5 N/RMasterson and Bannerman,

1994 WI

N/R = Not Reported, R = residential, U = urban, * = Median


Chloride concentrations in snowmelt runoffdepend on the amount applied and the dilutionin the receiving waters. Data for snowmelt andstormwater runoff from several studies arepresented in Table 44. For example, chlorideconcentrations in Lincoln Creek in Wisconsinwere 1,612 mg/l in winter snowmelt runoff, ascompared to 40 mg/l in non-winter runoff(Novotny et al., 1999 and Masterson andBannerman, 1994). Chloride concentrations inthe range of 2,000 to 5,000 mg/l have beenreported for Canadian streams (EnvironmentCanada, 2001). Novotny et al. (1999) moni-tored chloride concentrations in snowmelt nearSyracuse, New York and found that residentialwatersheds had higher chloride concentrationsthan rural watersheds.

Concentrations of glycol in stormwater runoffare also highly variable and depend on theamount of deicer used, the presence of arecovery system, and the nature of the precipi-tation event. Corsi et al. (2001) monitoredstreams receiving stormwater runoff from aWisconsin airport. They found concentrations

of propylene glycol as high as 39,000 mg/l atairport outfall sites during deicing operationsand concentrations of up to 960 mg/l duringlow-flow sampling at an airport outfall site.

4.12.2 Impacts of Deicerson Streams

Chloride levels can harm aquatic and terrestriallife and contaminate groundwater and drinkingwater supplies (Ohrel, 1995). Generally,chloride becomes toxic to many organismswhen it reaches concentrations of 500 to1,000mg/l (Environment Canada, 2001). Theseconcentrations are common in small streams insnow regions, at least for short periods of time.Many plant species are relatively intolerant tohigh salt levels in wetland swales and roadsidecorridors. Fish are also negatively affected byhigh chloride concentrations, with sensitivityas low as 600 mg/l for some species (Scott andWylie, 1980).

Table 45 compares the maximum chlorideconcentrations for various water uses in eightstates (USEPA, 1988). Snowmelt chlorideconcentrations typically exceed these levels.

Table 44: EMCs for Chloride in Snowmelt and Stormwater Runoff in Urban Areas in



Chloride is a concern in surface drinking watersystems because it can interfere with some ofthe treatment processes and can cause tasteproblems at concentrations as low as 250 mg/l.Chloride is also extremely difficult to removeonce it enters the water.

Glycol-based deicers have been shown to behighly toxic at relatively low concentrations instreams receiving airport runoff. These deicerscontain many proprietary agents, which mayincrease their toxicity and also make it verydifficult to set standards for their use (Hartwellet al., 1995). Corsi et al. (2001) observed acutetoxicity of Ceriodaphnia dubia, Pimephelaspromelax, Hyalela azteca, and Chironimustentans in Wisconsin streams that experiencedpropylene glycol concentrations of 5,000 mg/lor more. Chronic toxicity was observed forCeriodaphnia dubia and Pimephelas promelaxat propylene glycol concentrations of 1,500mg/l in the same study. In addition, glycolexerts an extremely high BOD on receivingwaters, which can quickly reduce or eliminatedissolved oxygen. Glycol can also be toxic tosmall animals that are attracted by its sweettaste (Novotny et al., 1999).

As with many urban pollutants, the effects ofchloride can be diluted in larger waterbodies.In general, small streams are more likely toexperience chloride effects, compared torivers, which have a greater dilution ability.

4.12.3 Sources and SourceAreas of Deicers

The main sources for deicers in urban water-sheds include highway maintenance crews,airport deicing operations, and homeownerapplications. Direct road application is thelargest source of chloride, by far. Source areasinclude roads, parking lots, sidewalks, stormdrains, airport runways, and snow collectionareas. Because deicers are applied to pavedsurfaces, the primary means of transport tostreams is through stormwater and meltwaterrunoff. Therefore, concentrations of deicercompounds are typically associated withfactors such as road density or traffic patterns.

State Limiting Concentration (mg/l) Beneficial Use

CO 250* Drinking water

IL500 General water supply

250 Drinking water

IN 500 Drinking water

MA 250 Class A waters

MN250 Drinking water

500 Class A fishing and recreation

OH 250 Drinking water

SD250 Drinking water

100 Fish propagation

VA 250 Drinking water

* Monthly average

State Limiting Concentration (mg/l) Beneficial Use

CO 250* Drinking water

IL500 General water supply

250 Drinking water

IN 500 Drinking water

MA 250 Class A waters

MN250 Drinking water

500 Class A fishing and recreation

OH 250 Drinking water

SD250 Drinking water

100 Fish propagation

VA 250 Drinking water

* Monthly average

Table 45: Summary of State Standards for Salinity of Receiving Waters (USEPA, 1988)



4.13 Conclusion

IC collects and accumulates pollutants depos-ited from the atmosphere, leaked from ve-hicles, or derived from other sources. Thepollutants build up over time but are washedoff quickly during storms and are often effi-ciently delivered to downstream waters. Thiscan create water quality problems for down-stream rivers, lakes and estuaries.

As a result of local and national monitoringefforts, we now have a much better under-standing of the nature and impacts of stormwa-ter pollution. The typical sample of urbanstormwater is characterized by high levels ofmany common pollutants such as sediment,nutrients, metals, organic carbon, hydrocar-bons, pesticides, and fecal coliform bacteria.Other pollutants that have more recentlybecome a concern in urban areas includeMTBE, deicers, and the pathogensCryptosporidium and Giardia. Concentrationsof most stormwater pollutants can be charac-terized, over the long run, by event mean stormconcentrations. Monitoring techniques havealso allowed researchers to identify sourceareas for pollutants in the urban environment,including stormwater hotspots, which generatehigher pollutant loads than normal develop-ment.

In general, most monitoring data shows thatmean pollutant storm concentrations are higherin urban watersheds than in non-urban ones.For many urban pollutants, EMCs can be usedto predict stormwater pollutant loads for urbanwatersheds, using IC as the key predictivevariable. While a direct relationship betweenIC and pollutant concentrations does notusually exist, IC directly influences the volumeof stormwater and hence, the total load. A fewexceptions are worth noting. MTBE, deicers,and PAH appear to be related more to traffic orroad density than IC. Additionally, MTBE andPAH concentrations may be greater at hotspotsource areas, which are not always widely oruniformly distributed across a watershed.Pesticides, bacteria and pathogens are oftenassociated with turf areas rather than IC.Bacteria and pathogen sources also includedirect inputs from wildlife and inappropriate

sewage discharges that are not uniformlydistributed across a watershed and are notdirectly related to IC.

Further research into the relationship betweenstormwater pollutant loads and other watershedindicators may be helpful. For example, itwould be interesting to see if turf cover is agood indicator of stream quality for impactedstreams. Other important watershed indicatorsworth studying are the influence of watershedtreatment practices, such as stormwaterpractices and stream buffers.

The direct effects of stormwater pollutants onaquatic systems appears to be a function of thesize of the receiving water and the initial healthof the aquatic community. For example, asmall urban stream receiving high stormwaterpollutant concentrations would be more likelyto experience impacts than a large river, whichis diluted by other land uses. Likewise, organ-isms in sensitive streams should be moresusceptible to stormwater pollutants thanpollution-tolerant organisms found in non-supporting streams.

Overall, the following conclusions can bemade:

• Sediment, nutrient and trace metal loads instormwater runoff can be predicted as afunction of IC, although concentrations arenot tightly correlated with watershed IC.

• Violations of bacteria standards areindirectly associated with watershed IC.

• It is not clear whether loads of hydrocar-bons, pesticides or chlorides can bepredicted on the basis of IC at the smallwatershed level.

• More research needs to be conducted toevaluate the usefulness of other watershedindicators to predict stormwater pollutantloads. For example, traffic, road density orhotspots may be useful in predictingMTBE, deicer and hydrocarbon loads.Also, watershed turf cover may be usefulin predicting pesticide and bacterial loads.



• Most research on pollutants in stormwaterrunoff has been conducted at the smallwatershed level. Additional research isneeded to evaluate the impact of watershedtreatment, such as stormwater and bufferpractices to determine the degree to whichthese may change stormwater concentra-tions or loads.

• Regional differences are evident for manystormwater pollutants, and these appear tobe caused by either differences in rainfallfrequency or snowmelt.


Chapter 5: Biological Impacts of Impervious Cover

Chapter 5: Biological Impacts ofImpervious Cover

This chapter reviews research on the impact ofurbanization on the aquatic community,focusing on aquatic insects, fish, amphibians,freshwater mussels, and freshwater wetlands.Specifically, the relationship between thehealth of the aquatic community and theamount of watershed IC is analyzed within thecontext of the Impervious Cover Model (ICM).

The chapter is organized as follows:

5.1 Introduction5.2 Indicators and General Trends5.3 Effects on Aquatic Insect1 Diversity5.4 Effects on Fish Diversity5.5 Effects on Amphibian Diversity5.6 Effects on Wetland Diversity5.7 Effects on Freshwater Mussel

Diversity5.8 Conclusion

5.1 Introduction

A number of studies, crossing differentecoregions and utilizing various techniques,have examined the link between watershedurbanization and its impact on stream andwetland biodiversity. These studies reveal thata relatively small amount of urbanization has anegative effect on aquatic diversity, and that aswatersheds become highly urban, aquaticdiversity becomes extremely degraded. Asdocumented in prior chapters, hydrologic,physical, and water quality changes caused bywatershed urbanization all stress the aquaticcommunity and collectively diminish thequality and quantity of available habitat. As aresult, these stressors generally cause a declinein biological diversity, a change in trophicstructure, and a shift towards more pollution-tolerant organisms.

Many different habitat conditions are criticalfor supporting diverse aquatic ecosystems. For

example, streambed substrates are vulnerableto deposition of fine sediments, which affectsspawning, egg incubation and fry-rearing.Many aquatic insect species shelter in the largepore spaces among cobbles and boulders,particularly within riffles. When fine sedimentfills these pore spaces, it reduces the qualityand quantity of available habitat. The aquaticinsect community is typically the base of thefood chain in streams, helps break downorganic matter and serves as a food source forjuvenile fish.

Large woody debris (LWD) plays a criticalrole in the habitat of many aquatic insects andfish. For example, Bisson et al. (1988) contendthat no other structural component is moreimportant to salmon habitat than LWD,especially in the case of juvenile coho salmon.Loss of LWD due to the removal of streamside vegetation can significantly hinder thesurvival of more sensitive aquatic species.Since LWD creates different habitat types, itsquality and quantity have been linked tosalmonid rearing habitat and the ability ofmultiple fish species to coexist in streams.

The number of stream crossings (e.g., roads,sewers and pipelines) has been reported toincrease directly in proportion to IC (May etal., 1997). Such crossings can become partialor total barriers to upstream fish migration,particularly if the stream bed downcuts belowthe fixed elevation of a culvert or pipeline.Fish barriers can prevent migration andrecolonization of aquatic life in many urbanstreams.

Urbanization can also increase pollutant levelsand stream temperatures. In particular, tracemetals and pesticides often bind to sedimentparticles and may enter the food chain, particu-larly by aquatic insects that collect and filterparticles. While in-stream data is rare, somedata are available for ponds. A study of trace

1Throughout this chapter, the term “aquatic insects” is used rather than the more cumbersome but technically correct“benthic macroinvertebrates.”



Stream Change Effects on Organisms

Increased flowvolumes/ Channelforming storms

Alterations in habitat complexityChanges in availability of food organisms, related to timing ofemergence and recovery after disturbanceReduced prey diversityScour-related mortalityLong-term depletion of LWDAccelerated streambank erosion

Decreased base flowsCrowding and increased competition for foraging sitesIncreased vulnerability to predationIncreased fine sediment deposition

Increase in sedimenttransport

Reduced survival of eggs and alevins, loss of habitat due todepositionSiltation of pool areas, reduced macroinvertebratereproduction

Loss of pools and riffles Shift in the balance of species due to habitat changeLoss of deep water cover and feeding areas

Changes in substratecomposition

Reduced survival of eggsLoss of inter-gravel fry refugial spacesReduced aquatic insect production

Loss of LWD

Loss of cover from predators and high flowsReduced sediment and organic matter storageReduced pool formation and organic substrate for aquaticinsects

Increase intemperature

Changes in migration patternsIncreased metabolic activity, increased disease and parasitesusceptibility Increased mortality of sensitive fish

Creation of fishblockages

Loss of spawning habitat for adultsInability to reach overwintering sitesLoss of summer rearing habitat,Increased vulnerability to predation

Loss of vegetativerooting systems

Decreased channel stabilityLoss of undercut banksReduced streambank integrity

Channel straighteningor hardening

Increased stream scourLoss of habitat complexity

Reduction in waterquality

Reduced survival of eggs and alevinsAcute and chronic toxicity to juveniles and adult fishIncreased physiological stress

Increase in turbidityReduced survival of eggsReduced plant productivityPhysiological stress on aquatic organisms

Algae bloomsOxygen depletion due to algal blooms, increasedeutrophication rate of standing waters

metal bioaccumulation of three fish speciesfound in central Florida stormwater pondsdiscovered that trace metal levels were signifi-cantly higher in urban ponds than in non-urbancontrol ponds, often by a factor of five to 10(Campbell, 1995; see also Karouna-Renier,1995). Although typical stormwater pollutantsare rarely acutely toxic to fish, the cumulativeeffects of sublethal pollutant exposure mayinfluence the stream community (Chapter 4).

Table 46 summarizes some of the numerouschanges to streams caused by urbanization thathave the potential to alter aquatic biodiversity.For a comprehensive review of the impacts ofurbanization on stream habitat andbiodiversity, the reader should consult Woodand Armitage (1997) and Hart and Finelli(1999).

Table 46: Review of Stressors to Urban Streams and Effects on Aquatic Life



5.2 Indicators andGeneral Trends

Stream indicators are used to gauge aquatichealth in particular watersheds. The two maincategories of stream indicators are biotic anddevelopment indices. Biotic indices usestream diversity as the benchmark for aquatichealth and use measures, such as speciesabundance, taxa richness, EPT Index, nativespecies, presence of pollution-tolerant species,dominance, functional feeding group compari-sons, or proportion with disease or anomalies.Development indices evaluate the relationshipbetween the degree of watershed urbanizationand scores for the biotic indices. Commondevelopment indices include watershed IC,housing density, population density, andpercent urban land use.

5.2.1 Biological Indicators

Biotic indices are frequently used to measurethe health of the aquatic insect or fish commu-nity in urban streams. Because many aquaticinsects have limited migration patterns or asessile mode of life, they are particularly well-suited to assess stream impacts over time.Aquatic insects integrate the effects of short-term environmental variations, as most specieshave a complex but short life cycle of a year orless. Sensitive life stages respond quickly toenvironmental stressors, but the overallcommunity responds more slowly. Aquaticinsect communities are comprised of a broadrange of species, trophic levels and pollutiontolerances, thus providing strong informationfor interpreting cumulative effects. Unlike fish,aquatic insects are abundant in most small, firstand second order streams. Individuals arerelatively easy to identify to family level, andmany “intolerant” taxa can be identified tolower taxonomic levels with ease.

Fish are good stream indicators over longertime periods and broad habitat conditionsbecause they are relatively long-lived andmobile. Fish communities generally include arange of species that represents a variety oftrophic levels (omnivores, herbivores, insecti-vores, planktivores, and piscivores). Fish tend

to integrate the effects of lower trophic levels;thus, their community structure reflects theprevailing food sources and habitat conditions.Fish are relatively easy to collect and identifyto the species level. Most specimens can besorted and identified in the field by experi-enced fisheries scientists and subsequentlyreleased unharmed.

A review of the literature indicates that a widevariety of metrics are used to measure theaquatic insect and fish community. Communityindices, such as the Index of Biotic Integrity(IBI) for fish and the Benthic Index of BioticIntegrity (B-IBI) for the aquatic insect commu-nity are a weighted combination of variousmetrics that typically characterize the commu-nity from “excellent” to “poor.” Commonmetrics of aquatic community are often basedon a composite of measures, such as speciesrichness, abundance, tolerance, trophic status,and native status. Combined indices (C-IBI)measure both fish and aquatic insect metricsand a variety of physical habitat conditions toclassify streams. Table 47 lists several com-mon metrics used in stream assessments. Itshould be clearly noted that community andcombined indices rely on different measure-ments and cannot be directly compared. For acomprehensive review of aquatic communityindicators, see Barbour et al.(1999).

5.2.2 Watershed DevelopmentIndices

Watershed IC, housing density, populationdensity, and percent urban land have all beenused as indices of the degree of watersheddevelopment. In addition, reverse indicatorssuch as percent forest cover and ripariancontinuity have also been used. The majorityof studies so far have used IC to explore therelationship between urbanization and aquaticdiversity. Percent urban land has been thesecond most frequently used indicator todescribe the impact of watershed development.Table 48 compares the four watershed devel-opment indices and the thresholds wheresignificant impacts to aquatic life are typicallyobserved.



Measurement Applied to: Definition of Measurement

Abundance Fish, Aquatic InsectsTotal number of individuals in a sample; sometimes modified to excludetolerant species.

Taxa Richness Fish, Aquatic InsectsTotal number of unique taxa identified in a sample. Typically, anincrease in taxa diversity indicates better water and habitat quality.

EPT Index Aquatic Insects

Taxa belonging to the following three groups: Ephemeroptera (mayflies),Plecoptera (stoneflies), Trichoptera (caddisflies). Typically, species inthese orders are considered to be pollution-intolerant taxa and aregenerally the first to disappear with stream quality degradation.

Native Status Fish Native vs. non-native taxa in the community.

Specific HabitatFish

Riffle benthic insectivorous individuals. Total number of benthicinsectivores. Often these types of individuals, such as darters, sculpins,and dace are found in high velocity riffles and runs and are sensitive tophysical habitat degradation.

Minnow species Total number of minnow species present. Often used asan indicator of pool habitat quality. Includes all species present in thefamily Cyprinidae, such as daces, minnows, shiners, stonerollers, andchubs.

Tolerant Species Fish, Aquatic Insects

The total number of species sensitive to and the number tolerant ofdegraded conditions. Typically, intolerant species decline withdecreasing water quality and stream habitat. A common high pollution-tolerant species that is frequently used is Chironomids.

Dominance Fish, Aquatic InsectsThe proportion of individuals at each station from the single mostabundant taxa at that particular station. Typically, a communitydominated by a single taxa may be indicative of stream degradation.

FunctionalFeeding GroupComparisons

Fish

Omnivores/ Generalists: The proportion of individuals characterized asomnivores or generalists to the total number of individuals. Typically,there is a shift away from specialized feeding towards moreopportunistic feeders under degraded conditions as food sourcesbecome unreliable.

Insectivores: The proportion of individuals characterized as insectivoresto the total number of individuals. Typically, the abundance ofinsectivores decreases relative to increasing stream degradation.

Aquatic Insects

Others: The proportion of individuals characterized as shredders,scrapers, or filter feeders to the total number of individuals. Typically,changes in the proportion of functional feeders characterized asshredders can be reflective of contaminated leaf matter. In addition, anoverabundance of scrapers over filterers can be indicative of increasedbenthic algae.

Disease/Anomalies Fish

Proportion of individuals with signs of disease or abnormalities. This isascertained through gross external examination for abnormalities duringthe field identification process. Typically, this metric assumes thatincidence of disease and deformities increases with increasing streamdegradation.

* This table is not meant to provide a comprehensive listing of metrics used for diversity indices; it is intended to provideexamples of types of measures used in biological stream assessments (see Barbour et al., 1999).

Measurement Applied to: Definition of Measurement

Abundance Fish, Aquatic InsectsTotal number of individuals in a sample; sometimes modified to excludetolerant species.

Taxa Richness Fish, Aquatic InsectsTotal number of unique taxa identified in a sample. Typically, anincrease in taxa diversity indicates better water and habitat quality.

EPT Index Aquatic Insects

Taxa belonging to the following three groups: Ephemeroptera (mayflies),Plecoptera (stoneflies), Trichoptera (caddisflies). Typically, species inthese orders are considered to be pollution-intolerant taxa and aregenerally the first to disappear with stream quality degradation.

Native Status Fish Native vs. non-native taxa in the community.

Specific HabitatFish

Riffle benthic insectivorous individuals. Total number of benthicinsectivores. Often these types of individuals, such as darters, sculpins,and dace are found in high velocity riffles and runs and are sensitive tophysical habitat degradation.

Minnow species Total number of minnow species present. Often used asan indicator of pool habitat quality. Includes all species present in thefamily Cyprinidae, such as daces, minnows, shiners, stonerollers, andchubs.

Tolerant Species Fish, Aquatic Insects

The total number of species sensitive to and the number tolerant ofdegraded conditions. Typically, intolerant species decline withdecreasing water quality and stream habitat. A common high pollution-tolerant species that is frequently used is Chironomids.

Dominance Fish, Aquatic InsectsThe proportion of individuals at each station from the single mostabundant taxa at that particular station. Typically, a communitydominated by a single taxa may be indicative of stream degradation.

FunctionalFeeding GroupComparisons

Fish

Omnivores/ Generalists: The proportion of individuals characterized asomnivores or generalists to the total number of individuals. Typically,there is a shift away from specialized feeding towards moreopportunistic feeders under degraded conditions as food sourcesbecome unreliable.

Insectivores: The proportion of individuals characterized as insectivoresto the total number of individuals. Typically, the abundance ofinsectivores decreases relative to increasing stream degradation.

Aquatic Insects

Others: The proportion of individuals characterized as shredders,scrapers, or filter feeders to the total number of individuals. Typically,changes in the proportion of functional feeders characterized asshredders can be reflective of contaminated leaf matter. In addition, anoverabundance of scrapers over filterers can be indicative of increasedbenthic algae.

Disease/Anomalies Fish

Proportion of individuals with signs of disease or abnormalities. This isascertained through gross external examination for abnormalities duringthe field identification process. Typically, this metric assumes thatincidence of disease and deformities increases with increasing streamdegradation.

* This table is not meant to provide a comprehensive listing of metrics used for diversity indices; it is intended to provideexamples of types of measures used in biological stream assessments (see Barbour et al., 1999).

Table 47: Examples of Biodiversity Metrics Used to Assess Aquatic Communities



5.2.3 General Trends

Most research suggests that a decline in bothspecies abundance and diversity begins at oraround 10% watershed IC (Schueler, 1994a).However, considerable variations in aquaticdiversity are frequently observed from five to20% IC, due to historical alterations, theeffectiveness of watershed management,prevailing riparian conditions, co-occurrenceof stressors, and natural biological variation(see Chapter 1).

Figures 40 through 42 display the negativerelationship commonly seen between bioticindices and various measures of watersheddevelopment. For example, stream research inthe Maryland Piedmont indicated that IC wasthe best predictor of stream condition, based ona combined fish and aquatic insect IBI(MNCPPC, 2000). In general, streams withless than 6% watershed IC were in “excellent”condition, whereas streams in “good” condi-tion had less than 12% IC, and streams in“fair” condition had less than 20%. Figure 40shows the general boundaries and typicalvariation seen in MNCPPC stream research.

Figure 41 illustrates that B-IBI scores andCoho Salmon/Cutthroat Trout Ratio are afunction of IC for 31 streams in Puget Sound,Washington. The interesting finding was that“good” to “excellent” B-IBI scores (greater

than 25) were reported in watersheds that hadless than 10% IC, with eight notable outliers.These outliers had greater IC (25 to 35%) butsimilar B-IBI scores. These outliers are uniquein that they had a large upstream wetland and/or a large, intact riparian corridor upstream(i.e. >70% of stream corridor had buffer width>100 feet).

Figure 42 depicts the same negative relation-ship between watershed urbanization and fish-IBI scores but uses population density as theprimary metric of development (Dreher, 1997).The six-county study area included the Chi-cago metro area and outlying rural watersheds.Significant declines in fish-IBI scores werenoted when population density exceeded 1.5persons per acre.

The actual level of watershed development atwhich an individual aquatic species begins todecline depends on several variables, but maybe lower than that indicated by the ICM. Someresearchers have detected impacts for indi-vidual aquatic species at watershed IC levels aslow as 5%. Other research has suggested thatthe presence of certain stressors, such assewage treatment plant discharges (Yoder andMiltner, 2000) or construction sites (Reice,2000) may alter the ICM and lower the level ofIC at which biodiversity impacts becomeevident.



Land UseIndicator

Level at whichSignificant Impact

Observed

Typical Value forLow Density

Residential UseComments

% IC 10-20% 10%Most accurate; highest level of effortand cost

HousingDensity

>1 unit/acre 1 unit/acre

Low accuracy in areas of substantialcommercial or industrialdevelopment; less accurate at smallscales

PopulationDensity

1.5 to 8+people/acre 2.5 people/acre

Low accuracy in areas of substantialcommercial or industrialdevelopment; less accurate at smallscales

% UrbanLand Use

33% (variable) 10-100%Does not measure intensity ofdevelopment; moderately accurateat larger watershed scales

Road Density 5 miles/square mile 2 miles/square mileAppears to be a potentially usefulindicator

Figure 40: Combined Fish and Benthic IBI vs. IC in Maryland Piedmont Streams(MNCPPC, 2000)

Table 48: Alternate Land Use Indicators and Significant Impact Levels(Brown, 2000; Konrad and Booth, 2002)



Figure 41: Relationship Between B-IBI, Coho/Cutthroat Ratios, andWatershed IC in Puget Sound Streams (Horner et al., 1997)

Figure 42: Index for Biological Integrity as a Function of Population Density in Illinois(Dreher, 1997)



5.3 Effects on AquaticInsect Diversity

The diversity, richness and abundance of theaquatic insect community is frequently used toindicate urban stream quality. Aquatic insectsare a useful indicator because they form thebase of the stream food chain in most regionsof the country. For this reason, declines orchanges in aquatic insect diversity are often anearly signal of biological impact due to water-shed development. The aquatic insect commu-nity typically responds to increasing develop-ment by losing species diversity and richnessand shifting to more pollution-tolerant species.More than 30 studies illustrate how IC andurbanization affect the aquatic insect commu-nity. These are summarized in Tables 49 and50.

5.3.1 Findings Based on ICIndicators

Klein (1979) was one of the first researchers tonote that aquatic insect diversity drops sharplyin streams where watershed IC exceeded 10 to15%. While “good” to “fair” diversity wasnoted in all headwater streams with less than10% IC, nearly all streams with 12% or morewatershed IC recorded “poor” diversity. Otherstudies have confirmed this general relation-ship between IC and the decline of aquaticinsect species diversity. Their relationshipshave been an integral part in the developmentof the ICM. The sharp drop in aquatic insectdiversity at or around 12 to 15% IC was alsoobserved in streams in the coastal plain andPiedmont of Delaware (Maxted and Shaver,1997).

Impacts at development thresholds lower than10% IC have also been observed by Booth(2000), Davis (2001), Horner et al. (1997) andMorse (2001). There seems to be a generalrecognition that the high levels of variabilityobserved below 10% IC indicate that otherfactors, such as riparian condition, effluentdischarges, and pollution legacy may be betterindicators of aquatic insect diversity (Hornerand May, 1999; Kennen, 1999; Steedman,1988; Yoder et al., 1999).

The exact point at which aquatic insect diver-sity shifts from fair to poor is not known withabsolute precision, but it is clear that few, ifany, urban streams can support diverse aquaticinsect communities with more than 25% IC.Indeed, several researchers failed to findaquatic insect communities with good orexcellent diversity in any highly urban stream(Table 52). Indeed, MNCPPC (2000) reportedthat all streams with more than 20% watershedIC were rated as “poor.”

Several good examples of the relationshipbetween IC and B-IBI scores are shown inFigures 43 through 45. Figure 43 depicts thegeneral trend line in aquatic insect diversity asIC increased at 138 stream sites in NorthernVirginia (Fairfax County, 2001). The surveystudy concluded that stream degradationoccurred at low levels of IC, and that olderdevelopments lacking more efficient sitedesign and stormwater controls tended to haveparticularly degraded streams. Figures 44 and45 show similar trends in the relationshipbetween IC and aquatic insect B-IBI scores inMaryland and Washington streams. In particu-lar, note the variability in B-IBI scores ob-served below 10% IC in both research studies.

Often, shift in the aquatic insect communityfrom pollution-sensitive species to pollution-tolerant species occurs at relatively low IClevels (<10%). This shift is often tracked usingthe EPT metric, which evaluates sensitivespecies found in the urban stream communityin the orders of Ephemeroptera (mayflies),Plecoptera (stoneflies), and Trichoptera(caddisflies). EPT species frequently disappearin urban streams and are replaced by morepollution-tolerant organisms, such as chirono-mids, tubificid worms, amphipods and snails.

In undisturbed streams, aquatic insects employspecialized feeding strategies, such as shred-ding leaf litter, filtering or collecting organicmatter that flows by, or preying on otherinsects. These feeding guilds are greatlyreduced in urban streams and are replaced bygrazers, collectors and deposit feeders. Maxtedand Shaver (1997) found that 90% of sensitive



Index Key Finding (s) Source Location

CommunityIndex

Three years stream sampling across the state at 1000 sites found that when IC was>15%, stream health was never rated good based on a C-IBI.

Boward et al.,1999 MD

CommunityIndex

Insect community and habitat scores were all ranked as poor in fivesubwatersheds that were greater than 30% IC.

Black andVeatch, 1994

MD

CommunityIndex

Puget sound study finds that some degradation of aquatic invertebrate diversitycan occur at any level of human disturbance (at least as measured by IC). 65% ofwatershed forest cover usually indicates a healthy aquatic insect community.

Booth, 2000 WA

CommunityIndex

In a Puget Sound study, the steepest decline of B-IBI was observed after 6% IC. There was a steady decline, with approximately 50% reduction in B-IBI at 45% IC.

Horner et al.,1997

WA

CommunityIndex

B-IBI decreases with increasing urbanization in study involving 209 sites, with a sharpdecline at 10% IC. Riparian condition helps mitigate effects.

Steedman, 1988 Ontario

CommunityIndex

Wetlands, forest cover and riparian integrity act to mitigate the impact of IC onaquatic insect communities.

Horner et al.,2001

WA, MD,TX

CommunityIndex B-IBI declines for aquatic insect with increasing IC at more than 200 streams. Fairfax Co.,

2001 VA

CommunityIndex

Two-year stream study of eight Piedmont watersheds reported B-IBI scores declinedsharply at an IC threshold of 15-30%.

Meyer andCouch,2000

GA

CommunityIndex

Montgomery County study; subwatersheds with <12% IC generally had streams ingood to excellent condition based on a combined fish and aquatic insect IBI. Watersheds with >20% IC had streams in poor condition.

MNCPPC, 2000

MD

CommunityIndex

Study of 1st, 2nd, and 3rd order streams in the Patapsco River Basin showed negativerelationship between B-IBI and IC.

Dail et al., 1998

MD

CommunityIndex

While no specific threshold was observed, impacts were seen at even low levels ofIC. B-IBI values declined with increasing IC, with high scores observed only inreaches with <5% IC or intact riparian zones or upstream wetlands.

Horner andMay, 1999 WA

CommunityIndex

The C-IBI also decreased by 50% at 10-15% IC. These trends were particularly strongat low-density urban sites (0-30% IC).

Maxted andShaver, 1997

DE

DiversityIn both coastal plain and Piedmont streams, a sharp decline in aquatic insectdiversity was found around 10-15% IC.

Shaver et al., 1995 DE

Diversity In a comparison of Anacostia subwatersheds, there was significant decline in thediversity of aquatic insects at 10% IC.

MWCOG, 1992

DC

Diversity In several dozen Piedmont headwater streams, aquatic diversity declinedsignificantly beyond 10-12% IC. Klein, 1979 MD

EPT Value In a 10 stream study with watershed IC ranging from three to 30%, a significantdecline in EPT values was reported as IC increased (r2 = 0.76).

Davis, 2001 MO

SensitiveSpecies

In a study of 38 wadeable, non-tidal streams in the urban Piedmont, 90% of sensitiveorganisms were eliminated from the benthic community after watershed IC reaches10-15%.

Maxted andShaver, 1997

DE

SpeciesAbundanceEPT values

For streams draining 20 catchments across the state, an abrupt decline in speciesabundance and EPT taxa was observed at approximately 6% IC.

Morse, 2001 ME

Table 49: Recent Research Examining the Relationship Between IC and Aquatic Insect Diversity in Streams



Biotic Key Finding (s) Source Location

Percent Urban Land use

CommunityIndex

Study of 700 streams in 5 major drainage basins found that the amount of urbanland and total flow of municipal effluent were the most significant factors inpredicting severe impairment of the aquatic insect community. Amount offorested land in drainage area was inversely related to impairment severity.

Kennen, 1999 NJ

CommunityIndex

All 40 urban sites sampled had fair to very poor B-IBI scores, compared toundeveloped reference sites. Yoder, 1991 OH

CommunityIndex

A negative correlation between B-IBI and urban land use was noted. Communitycharacteristics show similar patterns between agricultural and forested areas themost severe degradation being in urban and suburban areas.

Meyer andCouch, 2000

GA

EPT Value,Diversity,CommunityIndex

A comparison of three stream types found urban streams had lowest diversity andrichness. Urban streams had substantially lower EPT scores (22% vs 5% as number ofall taxa, 65% vs 10% as percent abundance) and IBI scores in the poor range.

Crawford andLenat, 1989

NC

SensitiveSpecies

Urbanization associated with decline in sensitive taxa, such as mayflies, caddisfliesand amphipods while showing increases in oligochaetes.

Pitt andBozeman, 1982 CA

SensitiveSpecies

Dramatic changes in aquatic insect community were observed in most urbanizingstream sections. Changes include an abundance of pollution-tolerant aquaticinsect species in urban streams.

Kemp andSpotila, 1997

PA

Diversity As watershed development levels increased, the aquatic insect diversity declined.Richards et al.,

1993 MN

Diversity Significant negative relationship between number of aquatic insect species anddegree of urbanization in 21 Atlanta streams.

Benke et al.,1981

GA

Diversity Drop in insect taxa from 13 to 4 was noted in urban streams. Garie andMcIntosh, 1986 NJ

Diversity Aquatic insect taxa were found to be more abundant in non-urban reaches thanin urban reaches of the watershed.

Pitt andBozeman, 1982

CA

Diversity A study of five urban streams found that as watershed land use shifted from rural tourban, aquatic insect diversity decreased.

Masterson andBannerman,

1994WI

Other Land Use Indicators

CommunityIndex

Most degraded streams were found in developed areas, particularly olderdevelopments lacking newer and more efficient stormwater controls.

Fairfax Co., 2001 VA

Diversity Urban streams had sharply lower aquatic insect diversity with human populationabove four persons/acre in northern VA.

Jones andClark, 1987

VA

EPT Value

Monitoring of four construction sites in three varying regulatory settings found thatEPT richness was related to enforcement of erosion and sediment controls. Thepattern demonstrated that EPT richness was negatively affected as one movedfrom upstream to at the site, except for one site.

Reice, 2000 NC

SensitiveSpecies

In a Seattle study, aquatic insect community shifted to chironomid, oligochaetesand amphipod species that are pollution-tolerant and have simple feeding guild.

Pedersen andPerkins,1986

WA

Table 50: Recent Research Examining the Relationship of Other Indices of WatershedDevelopment on Aquatic Insect Diversity in Streams



species (based on EPT richness, % EPTabundance, and Hilsenhoff Biotic Index) wereeliminated from the aquatic insect communitywhen IC exceeded 10 to 15% in contributingwatersheds of Delaware streams (Figure 46). Ina recent study of 30 Maine watersheds, Morse(2001) found that reference streams with less

than 5% watershed IC had significantly moreEPT taxa than more urban streams. He alsoobserved no significant differences in EPTIndex values among streams with six to 27%watershed IC (Figure 47).

Figure 45: IC and B-IBI at Stream Sites in thePatapsco River Basin, MD

(Dail et al., 1998)

Figure 43: Trend Line Indicating Decline inBenthic IBI as IC Increases in Northern VA

Streams (Fairfax County, 2001)

Figure 44: Relationship Between IC and B-IBIScores in Aquatic Insects in Streams of the

Puget Sound Lowlands (Booth, 2000)

Figure 46: IC vs. Aquatic Insect Sensitivity -EPT Scores in Delaware Streams

(Maxted and Shaver, 1997)



5.3.2 Findings Based on OtherDevelopment Indicators

Development indices, such as percent urbanland use, population density, and forest andriparian cover have also been correlated withchanges in aquatic insect communities in urbanstreams. Declines in benthic IBI scores havefrequently been observed in proportion to thepercent urban land use in small watersheds(Garie and McIntosh, 1986; Kemp and Spotila,1997; Kennen, 1999; Masterson andBannerman, 1994; Richards et al., 1993;USEPA, 1982).

A study in Washington state compared aheavily urbanized stream to a stream withlimited watershed development and found thatthe diversity of the aquatic insect communitydeclined from 13 taxa in reference streams tofive taxa in more urbanized streams (Pedersenand Perkins, 1986). The aquatic insect taxa thatwere lost were poorly suited to handle thevariable erosional and depositional conditionsfound in urban streams. Similarly, a compari-son of three North Carolina streams withdifferent watershed land uses concluded theurban watershed had the least taxa and lowestEPT scores and greatest proportion of pollu-tion-tolerant species (Crawford and Lenat,1989).

Jones and Clark (1987) monitored 22 streamsin Northern Virginia and concluded thataquatic insect diversity diminished markedlyonce watershed population density exceededfour or more people per acre. The populationdensity roughly translates to ½ - 1 acre lotresidential use, or about 10 to 20 % IC. Kennen(1999) evaluated 700 New Jersey streams andconcluded that the percentage of watershedforest was positively correlated with aquaticinsect density. Meyer and Couch (2000)reported a similar cover relationship betweenaquatic insect diversity and watershed andriparian forest cover for streams in the Atlanta,GA region. A study in the Puget Sound regionfound that aquatic insect diversity declined instreams once forest cover fell below 65%(Booth, 2000).

Figure 47: Average and Spring EPT Index Values vs.% IC in 20 Small Watershedsin Maine (Morse, 2001)



5.4 Effects on Fish Diversity

Fish communities are also excellent environ-mental indicators of stream health. In general,an increase in watershed IC produces the samekind of impact on fish diversity as it does foraquatic insects. The reduction in fish diversityis typified by a reduction in total species, lossof sensitive species, a shift toward morepollution-tolerant species, and decreasedsurvival of eggs and larvae. More than 30studies have examined the relationship be-tween watershed development and fish diver-sity; they are summarized in Tables 51 and 52.About half of the research studies used IC asthe major index of watershed development,while the remainder used other indices, such aspercent urban land use, population density,housing density, and forest cover.

5.4.1 Findings Based onIC Indicators

Recent stream research shows a consistent,negative relationship between watersheddevelopment and various measures of fishdiversity, such as diversity metrics, speciesloss and structural changes.

Typically, a notable decline in fish diversityoccurs around 10 to 15% watershed IC(Boward et al., 1999; Galli, 1994; Klein, 1979;Limburg and Schmidt, 1990; MNCPPC, 2000;MWCOG, 1992; Steward, 1983). A somewhathigher threshold was observed by Meyer andCouch (2000) for Atlanta streams with 15 to30% IC; lower thresholds have also beenobserved (Horner et al., 1997 and May et al.,1997). A typical relationship between water-shed IC and fish diversity is portrayed inFigure 48, which shows data from streams inthe Patapsco River Basin in Maryland (Dail etal., 1998). Once again, note the variability infish-IBI scores observed below 10% IC.

Wang et al. (1997) evaluated 47 Wisconsinstreams and found an apparent thresholdaround 10% IC. Fish-IBI scores were “good”to “excellent” below this threshold, but wereconsistently rated as “fair” to “poor.” Addi-tionally, Wang documented that the totalnumber of fish species drops sharply when ICincreases (Figure 49). Often, researchers alsoreported that increases in IC were stronglycorrelated with several fish metrics, such asincreases in non-native and pollution-tolerantspecies in streams in Santa Clara, California(EOA, Inc., 2001).

Figure 48: Fish-IBI vs. Watershed IC for Streams in the Patapsco River Basin, MD(Dail et al., 1998)




Abundance Brown trout abundance and recruitment declined sharply at 10-15% IC. Galli, 1994 MD

Salmonids Seattle study showed marked reduction in coho salmon populations noted at 10-15%IC at nine streams.

Steward, 1983 WA

Anadromous FishEggs

Resident and anadromous fish eggs and larvae declined in 16 subwatershedsdraining to the Hudson River with >10% IC area.

Limburg andSchmidt,

1990NY

CommunityIndex

1st, 2nd, and 3rd order streams in the Patapsco River Basin showed negativerelationship between IBI and IC.

Dail et al., 1998 MD

CommunityIndex

Fish IBI and habitat scores were all ranked as poor in five subwatersheds that weregreater than 30% IC.

Black andVeatch,1994 MD

CommunityIndex

In the Potomac subregion, subwatersheds with < 12% IC generally had streams ingood to excellent condition based on a combined fish and aquatic insect IBI.

Watersheds with >20% IC had streams in poor condition.

MNCPPC,2000 MD

CommunityIndex

In a two-year study of Piedmont streams draining eight watersheds representingvarious land uses in Chattahochee River Basin, fish community quality droppedsharply at an IC threshold of 15-30%.

Meyer andCouch,

2000GA

DiversityOf 23 headwater stream stations, all draining <10% IC areas, rated as good tofair; all with >12% were rated as poor. Fish diversity declined sharply with

increasing IC between 10-12%.

Schuelerand Galli,

1992MD

Diversity, Sensitive Species

Comparison of 4 similar subwatersheds in Piedmont streams, there was significantdecline in the diversity of fish at 10% IC. Sensitive species (trout and sculpin) were lostat 10-12%.

MWCOG, 1992 MD

Diversity,CommunityIndex

In a comparison of watershed land use and fish community data for 47 streamsbetween the 1970s and 1990s, a strong negative correlation was found betweennumber species and IBI scores with effective connected IC. A threshold of 10% ICwas observed with community quality highly variable below 10% but consistently lowabove 10% IC.

Wang et al.,1997 WI

Diversity In several dozen Piedmont headwater streams fish diversity declined significantly inareas beyond 10-12% IC. Klein, 1979 MD

Diversity ,Abundance,Non-nativeSpecies

IC strongly associated with several fisheries species and individual-level metrics,including number of pollution-tolerant species, diseased individuals, native and non-native species and total species present

EOA, Inc., 2001 CA

Juvenile SalmonRatios

In Puget Sound study, the steepest decline of biological functioning was observedafter six percent IC. There was a steady decline, with approximately 50% reductionin initial biotic integrity at 45% IC area.

Horner etal., 1997 WA

Juvenile SalmonRatio

Physical and biological stream indicators declined most rapidly during the initialphase of the urbanization process as total IC area exceeded the five to 10% range.

May et al., 1997 WA

Salmonoid Negative effects of urbanization (IC) with the defacto loss of non-structural BMPs(wetland forest cover and riparian integrity) on salmon ratios

Horner etal., 2001 WA, MD, TX

Salmonoid,Sensitive Species

While no specific threshold was observed (impacts seen at even low levels of IC),Coho/cutthroat salmon ratios >2:1 were found when IC was < 5%. Ratios fell belowone at IC levels below 20 %.

Horner andMay, 1999 WA

Sensitive species,Salmonid

Three years stream sampling across the state (approximately 1000 sites), MBSS foundthat when IC was >15%, stream health was never rated good based on CBI, andpollution sensitive brook trout were never found in streams with >2% IC.

Boward etal., 1999 MD

SensitiveSpecies,Salmonids

Seattle study observed shift from less tolerant coho salmon to more tolerant cutthroattrout population between 10 and 15% IC at nine sites.

Luchetti andFeurstenburg

1993WA

Table 51: Recent Research Examining the Relationship Between Watershed IC and the Fish Community



Sensitive fish are defined as species thatstrongly depend on clean and stable bottomsubstrates for feeding and/or spawning. Sensi-tive fish often show a precipitous decline inurban streams. The loss of sensitive fishspecies and a shift in community structuretowards more pollution-tolerant species isconfirmed by multiple studies. Figure 50shows the results of a comparison of foursimilar subwatersheds in the Maryland Pied-mont that were sampled for the number of fishspecies present (MWCOG, 1992). As the levelof watershed IC increased, the number of fishspecies collected dropped. Two sensitivespecies, including sculpin, were lost when ICincreased from 10 to 12%, and four morespecies were lost when IC reached 25%.Significantly, only two species remained in thefish community at 55% watershed IC.

Salmonid fish species (trout and salmon) andanadromous fish species appear to be particu-larly impacted by watershed IC. In a study inthe Pacific Northwest, sensitive coho salmonwere seldom found in watersheds above 10 or15% IC (Luchetti and Feurstenburg, 1993 andSteward, 1983). Key stressors in urbanstreams, such as higher peak flows, lower dryweather flows, and reduction in habitat com-plexity (e.g. fewer pools, LWD, and hidingplaces) are believed to change salmon speciescomposition, favoring cutthroat trout popula-tions over the natural coho populations(WDFW, 1997).

A series of studies from the Puget Soundreported changes in the coho/cutthroat ratios ofjuvenile salmon as watershed IC increased(Figure 51). Horner et al. (1999) found Coho/Cutthroat ratios greater than 2:1 in watershedswith less than 5 % IC. Ratios fell below 1:1when IC exceeded 20%. Similar results werereported by May et al. (1997). In the mid-Atlantic region, native trout have stringenttemperature and habitat requirements and areseldom present in watersheds where IC ex-ceeds 15% (Schueler, 1994a). Declines in troutspawning success are evident above 10% IC.In a study of over 1,000 Maryland streams,Boward et al. (1999) found that sensitive brooktrout were never found in streams that had morethan 4% IC in their contributing watersheds.

Figure 49: Fish-IBI and Number of Species vs. % IC inWisconsin Streams (Wang et al., 1997)

Figure 50: IC and Effects on Fish Species Diversity in FourMaryland Subwatersheds (MWCOG, 1992)

Imperviousness (%)

Fish DiversityAnacostia River Basin




Urbanization

CommunityIndex

All 40 urban sites sampled had fair to very poor IBI scores, compared toundeveloped reference sites.

Yoder, 1991 OH

CommunityIndex

Negative correlations between biotic community and riparian conditions andforested areas were found. Similar levels of fish degradation were foundbetween suburban and agricultural; urban areas were the most severe.

Meyer andCouch, 2000 GA

CommunityIndex

Residential urban land use caused significant decrease in fish-IBI scores at 33%. In more urbanized Cuyahoga, a significant drop in IBI scores occurred around8% urban land use in the watershed. When watersheds smaller than 100mi2 wereanalyzed separately, the level of urban land associated with a significant dropin IBI scores occurred at around 15%. Above one du/ac, most sites failed toattain biocriteria regardless of degree of urbanization.

Yoder et al.,1999

OH

CommunityIndex,Abundance

As watershed development increased to about 10%, fish communities simplifiedto more habitat and trophic generalists and fish abundance and speciesrichness declined. IBI scores for the urbanized stream fell from the good tofair category.

Weaver, 1991 VA

Diversity A study of five urban streams found that as land use shifted from rural to urban,fish diversity decreased.

Mastersonand

Bannerman, 1994

WI

Diversity,CommunityIndex

A comparison of three stream types found urban streams had lowest diversityand richness. Urban streams had IBI scores in the poor range.

Crawfordand Lenat,

1989NC

SalmonSpawning,FloodingFrequency

In comparing three streams over a 25-year period (two urbanizing and oneremaining forested), increases in flooding frequencies and decreased trends insalmon spawning were observed in the two urbanizing streams, while nochanges in flooding or spawning were seen in the forested system.

Moscript andMontgomery,

1997WA

SensitiveSpecies

Observed dramatic changes in fish communities in most urbanizing streamsections, such as absence of brown trout and abundance of pollution-tolerantspecies in urban reaches.

Kemp andSpotila,1997

PA

SensitiveSpecies,Diversity

Decline in sensitive species diversity and composition and changes in trophicstructure from specialized feeders to generalists was seen in an urbanizingwatershed from 1958 to 1990. Low intensity development was found to affectwarm water stream fish communities similarly as more intense development.

Weaver andGarman,

1994VA

Warm WaterHabitatBiocriteria

25-30% urban land use defined as the upper threshold where attainment ofwarm water habitat biocriterion is effectively lost. Non-attainment also mayoccur at lower thresholds given the co-occurrence of stressors, such as pollutionlegacy, WTPs and CSOs.

Yoder andMiltner, 2000 OH

CommunityIndex, Habitat

The amount of urban land use upstream of sample sites had a strong negativerelationship with biotic integrity, and there appeared to be a threshold between10 and 20% urban land use where IBI scores declined dramatically. Watershedsabove 20% urban land invariably had scores less than 30 ( poor to verypoor ). Habitat scores were not tightly correlated with degraded fish communityattributes.

Wang et al., 1997

WI

CommunityIndex

A study in the Patapsco Basin found significant correlation of fish IBI scores withpercent urbanized land over all scales (catchment, riparian area, and localarea).

Roth et al., 1998 MD

Table 52: Recent Research Examining Urbanization and Freshwater Fish Community Indicators




Urbanization

SensitiveSpecies

Evaluated effects of runoff in both urban and non-urban streams; found thatnative species dominated the non-urban portion of the watershed butaccounted for only seven percent of species found in the urban portions of thewatershed.

Pitt, 1982 CA

Other Land Use Indicators

CommunityIndex, Habitat

Atlanta study found that as watershed population density increased, there wasa negative impact on urban fish and habitat. Urban stream IBI scores wereinversely related to watershed population density, and once density exceededfour persons/acre, urban streams were consistently rated as very poor.

Couch et al., 1997 GA

CommunityIndex

In an Atlanta stream study, modified IBI scores declined once watershedpopulation density exceeds four persons/acre in 21 urban watersheds

DeVivo et al.,1997

GA

CommunityIndex

In a six-county study (including Chicago, its suburbs and outlyingrural/agricultural areas), streams showed a strong correlation betweenpopulation density and fish community assessments such that as populationdensity increased, community assessment scores went from the better -good range to fair - poor. Significant impacts seen at 1.5 people/acre.

Dreher, 1997 IL

CommunityIndex

Similarly, negative correlations between biotic community and riparianconditions and forested areas were also found. Similar levels of fish degradationwere found between suburban and agricultural; urban areas were the mostsevere.

Meyer andCouch, 2000

GA

CommunityIndex

Amount of forested land in basin directly related to IBI scores for fish communitycondition.

Roth et al., 1996

MD

Salmonid,SensitiveSpecies

Species community changes from natural coho salmon to cutthroat troutpopulation with increases in peak flow, lower low flow, and reductions in streamcomplexity.

WDFW, 1997 WA

Table 52 (continued): Recent Research Examining Urbanization and Freshwater Fish Community Indicators

Figure 51: Coho Salmon/Cutthroat Trout Ratio for Puget Sound Streams (Horner et al., 1997)



Many fish species have poor spawning successin urban streams and poor survival of fish eggsand fry. Fish barriers, low intragravel dissolvedoxygen, sediment deposition and scour are allfactors that can diminish the ability of fishspecies to successfully reproduce. For ex-ample, Limburg and Schmidt (1990) discov-ered that the density of anadromous fish eggsand larvae declined sharply in subwatershedswith more than 10% IC.

5.4.2 Findings Based on OtherDevelopment Indicators

Urban land use has frequently been used as adevelopment indicator to evaluate the impacton fish diversity. Streams in urban watershedstypically had lower fish species diversity andrichness than streams located in less developedwatersheds. Declines in fish diversity as afunction of urban land cover have been docu-mented in numerous studies (Crawford andLenat, 1989; Masterson and Bannerman, 1994;Roth et al., 1998; Yoder, 1991, and Yoder etal., 1999). USEPA (1982) found that nativefish species dominated the fish community ofnon-urban streams, but accounted for only 7%of the fish community found in urban streams.Kemp and Spotila (1997) evaluated streams inPennsylvania and noted the loss of sensitive

species (e.g. brown trout) and the increase ofpollution-tolerant species, such as sunfish andcreek chub (Figure 52).

Wang et al. (1997) cited percentage of urbanland in Wisconsin watersheds as a strongnegative factor influencing fish-IBI scores instreams and observed strong declines in IBIscores with 10 to 20% urban land use. Weaverand Garman (1994) compared the historicalchanges in the warm-water fish community ofa Virginia stream that had undergone signifi-cant urbanization and found that many of thesensitive species present in 1958 were eitherabsent or had dropped sharply in abundancewhen the watershed was sampled in 1990.Overall abundance had dropped from 2,056fish collected in 1958 to 417 in 1990. Inaddition, the 1990 study showed that 67% ofthe catch was bluegill and common shiner, twospecies that are habitat and trophic “general-ists.” This shift in community to more habitatand trophic generalists was observed at 10%urban land use (Weaver, 1991).

Yoder et al. (1999) evaluated a series ofstreams in Ohio and reported a strong decreasein warm-water fish community scores around33% residential urban land use. In the moreurbanized Cuyahoga streams, sharp drops in

Figure 52: Mean Proportion of Fish Taxa in Urban and Non-Urban Streams, ValleyForge Watershed, PA (Kemp and Spotila, 1997)



fish-IBI scores occurred around 8% urban landuse, primarily due to certain stressors whichfunctioned to lower the non-attainment thresh-old. When watersheds smaller than 100mi2

were analyzed separately, the percentage ofurban land use associated with a sharp drop infish-IBI scores was around 15%. In a laterstudy, Yoder and Miltner (2000) described anupper threshold for quality warm-water fishhabitat at 25 to 30% urban land use.

Watershed population and housing densityhave also been used as indicators of the healthof the fish community. In a study of 21 urbanwatersheds in Atlanta, DeVivo et al. (1997)

observed a shift in mean fish-IBI scores from“good to fair” to “very poor” when watershedpopulation density exceeded four people/acre(Figure 53). A study of Midwest streams inmetropolitan Illinois also found a negativerelationship between increase in populationdensity and fish communities, with significantimpacts detected at population densities of 1.5people or greater per acre (Dreher, 1997). Inthe Columbus and Cuyahoga watersheds inOhio, Yoder et al. (1999) concluded that moststreams failed to attain fish biocriteria aboveone dwelling unit/acre.

Figure 53: Relationship Between Watershed Population Density and StreamIBI Scores in Georgia Streams (DeVivo et al., 1997)



5.5 Effects onAmphibian Diversity

Amphibians spend portions of their life cyclein aquatic systems and are frequently foundwithin riparian, wetland or littoral areas.Relatively little research has been conducted todirectly quantify the effects of watersheddevelopment on amphibian diversity. Intu-itively, it would appear that the same stressorsthat affect fish and aquatic insects would alsoaffect amphibian species, along with riparianwetland alteration. We located four researchstudies on the impacts of watershed urbaniza-tion on amphibian populations; only one wasrelated to streams (Boward et al., 1999), whileothers were related to wetlands (Table 53).

A primary factor influencing amphibiandiversity appears to be water level fluctuations(WLF) in urban wetlands that occur as a resultof increased stormwater discharges. Chin(1996) hypothesized that increased WLF andother hydrologic factors affected the abun-

dance of egg clutches and available amphibianbreeding habitat, thereby ultimately influenc-ing amphibian richness. Increased WLF canlimit reproductive success by eliminatingmating habitat and the emergent vegetation towhich amphibians attach their eggs.

Taylor (1993) examined the effect of water-shed development on 19 freshwater wetlandsin King County, WA and concluded that theadditional stormwater contributed to greaterannual WLF. When annual WLF exceededabout eight inches, the richness of both thewetland plant and amphibian communitiesdropped sharply. Large increases in WLF wereconsistently observed in freshwater wetlandswhen IC in upstream watersheds exceeded 10to 15%. Further research on streams andwetlands in the Pacific northwest by Horner etal. (1997) demonstrated the correlation be-tween watershed IC and diversity of amphibianspecies. Figure 54 illustrates the relationshipbetween amphibian species abundance andwatershed IC, as documented in the study.



Indicator Key Finding(s) Reference Year Location

% IC

Reptile and AmphibianAbundance

In a three-year stream sampling across the state(approximately 1000 sites), MBSS found onlyhardy pollution-tolerant reptiles and amphibiansin stream corridors with >25% IC drainage area.

Boward et al.,1999

MD

Amphibian DensityMean annual water fluctuation inverselycorrelated to amphibian density in urbanwetlands. Declines noted beyond 10% IC.

Taylor, 1993 WA

Other Studies

Species Richness

In 30 wetlands, species richness of reptiles andamphibians was significantly related to density ofpaved roads on lands within a two kilometerradius.

Findlay andHoulahan,1997

Ontario

Species Richness

Decline in amphibian species richness as wetlandWLF increased. While more of a continuousdecline rather than a threshold, WLF = 22centimeters may represent a tolerance boundaryfor amphibian community.

Horner et al., 1997

WA

Amphibian DensityMean annual water fluctuation inverselycorrelated to amphibian density in urbanwetlands.

Taylor, 1993 WA

Table 53: Recent Research on the Relationship Between Percent WatershedUrbanization and the Amphibian Community

Figure 54: Amphibian Species Richness as a Function of Watershed IC inPuget Sound Lowland Wetlands (Horner et al., 1997)



5.6 Effects onWetland Diversity

We found a limited number of studies thatevaluated the impact of watershed urbanizationon wetland plant diversity (Table 54). Twostudies used IC as an index of watersheddevelopment and observed reduced wetlandplant diversity around or below 10% IC (Hicksand Larson, 1997 and Taylor, 1993). WLF androad density were also used as indicators(Findlay and Houlahan, 1997; Horner et al.,1997; Taylor, 1993).

Horner et al. (1997) reported a decline in plantspecies richness in emergent and scrub-shrubwetland zones of the Puget Sound region asWLF increased. They cautioned that speciesnumbers showed a continuous decline ratherthan a threshold value; however, it was indi-cated that WLF as small as 10 inches canrepresent a tolerance boundary for wetlandplant communities. Horner further stated thatin 90% of the cases where WLF exceeded 10inches, watershed IC exceeded 21%.

WatershedIndicator Key Finding(s) Reference Location

Biotic

% IC

InsectCommunity

Significant declines in various indicators ofwetland aquatic macro-invertebratecommunity health were observed as ICincreased to 8-9%.

Hicks andLarson, 1997

CT

WLF, WaterQuality

There is a significant increase in WLF,conductivity, fecal coliform bacteria, andtotal phosphorus in urban wetland as ICexceeds 3.5%.

Taylor et al., 1995 WA

Plant Density Declines in urban wetland plant densitynoted in areas beyond 10% IC.

Taylor, 1993 WA

Other Watershed Indicators

Plant DensityMean annual water fluctuation inverselycorrelated to plant density in urban wetlands. Taylor, 1993 WA

Plant SpeciesRichness

Decline in plant species richness in emergentand scrub-shrub wetland zones as WLFincreased. While more of a continuousdecline, rather than a threshold, WLF=22centimeters may represent a toleranceboundary for the community

Horner et al., 1997

WA


In 30 wetlands, species richness wassignificantly related to density of paved roadswithin a two kilometer radius of the wetland.Model predicted that a road density of2kilometers per hectare in paved road within1000 meters of wetland will lead to a 13%decrease in wetland plant species richness.

Findlay andHoulahan,1997 Ontario

WatershedIndicator Key Finding(s) Reference Location

Biotic

% IC

InsectCommunity

Significant declines in various indicators ofwetland aquatic macro-invertebratecommunity health were observed as ICincreased to 8-9%.

Hicks andLarson, 1997

CT

WLF, WaterQuality

There is a significant increase in WLF,conductivity, fecal coliform bacteria, andtotal phosphorus in urban wetland as ICexceeds 3.5%.

Taylor et al., 1995 WA

Plant Density Declines in urban wetland plant densitynoted in areas beyond 10% IC.

Taylor, 1993 WA

Other Watershed Indicators

Plant DensityMean annual water fluctuation inverselycorrelated to plant density in urban wetlands. Taylor, 1993 WA


Decline in plant species richness in emergentand scrub-shrub wetland zones as WLFincreased. While more of a continuousdecline, rather than a threshold, WLF=22centimeters may represent a toleranceboundary for the community

Horner et al., 1997

WA


In 30 wetlands, species richness wassignificantly related to density of paved roadswithin a two kilometer radius of the wetland.Model predicted that a road density of2kilometers per hectare in paved road within1000 meters of wetland will lead to a 13%decrease in wetland plant species richness.

Findlay andHoulahan,1997 Ontario

Table 54: Recent Research Examining the Relationship Between WatershedDevelopment and Urban Wetlands



5.7 Effects on FreshwaterMussel Diversity

Freshwater mussels are excellent indicators ofstream quality since they are filter-feeders andessentially immobile. The percentage ofimperiled mussel species in freshwaterecoregions is high (Williams et al., 1993). Ofthe 297 native mussel species in the UnitedStates, 72% are considered endangered,threatened, or of special concern, including 21mussel species that are presumed to be extinct.Seventy mussel species (24%) are consideredto have stable populations, although many ofthese have declined in abundance and distribu-tion. Modification of aquatic habitats andsedimentation are the primary reasons cited forthe decline of freshwater mussels (Williams etal., 1993).

Freshwater mussels are very susceptible tosmothering by sediment deposition. Conse-quently, increases in watershed developmentand sediment loading are suspected to be afactor leading to reduced mussel diversity. At

sublethal levels, silt interferes with feeding andmetabolism of mussels in general (Aldridge etal., 1987). Major sources of mortality and lossof diversity in mussels include impoundmentof rivers and streams, and eutrophication(Bauer, 1988). Changes in fish diversity andabundance due to dams and impoundments canalso influence the availability of mussel hosts(Williams et al., 1992).

Freshwater mussels are particularly sensitive toheavy metals and pesticides (Keller and Zam,1991). Although the effects of metals andpesticides vary from one species to another,sub-lethal levels of PCBs, DDT, Malathion,Rotenone and other compounds are generallyknown to inhibit respiratory efficiency andaccumulate in tissues (Watters, 1996). Musselsare more sensitive to pesticides than manyother animals tested and often act as “first-alerts” to toxicity long before they are seen inother organisms.

We were unable to find any empirical studiesrelating impacts of IC on the freshwater musselcommunities of streams.



5.8 Conclusion

The scientific record is quite strong withrespect to the impact of watershed urbanizationon the integrity and diversity of aquaticcommunities. We reviewed 35 studies thatindicated that increased watershed develop-ment led to declines in aquatic insect diversityand about 30 studies showing a similar impacton fish diversity. The scientific literaturegenerally shows that aquatic insect and fresh-water fish diversity declines at fairly low levelsof IC (10 to 15%), urban land use (33%),population density (1.5 to eight people/acre)and housing density (>1 du/ac). Many studiesalso suggest that sensitive elements of theaquatic community are affected at even lowerlevels of IC. Other impacts include loss ofsensitive species and reduced abundance andspawning success. Research supports the ICM,although additional research is needed toestablish the upper threshold at which water-shed development aquatic biodiversity can berestored.

One area where more research is neededinvolves determining how regional and cli-matic variations affect aquatic diversity in theICM. Generally, it appears that the 10% ICthreshold applies to streams in the East Coastand Midwest, with Pacific Northwest streamsshowing impacts at a slightly higher level. Forstreams in the arid and semi-arid Southwest, itis unclear what, if any, IC threshold existsgiven the naturally stressful conditions forthese intermittent and ephemeral streams

(Maxted, 1999). Southwestern streams arecharacterized by seasonal bursts of short butintense rainfall and tend to have aquaticcommunities that are trophically simple andrelatively low in species richness (Poff andWard, 1989).

Overall, the following conclusions can bedrawn:

• IC is the most commonly used index toassess the impacts of watershed urbaniza-tion on aquatic insect and fish diversity.Percent urban land use is also a commonindex.

• The ICM may not be sensitive enough topredict biological diversity in watershedswith low IC. For example, below 10%watershed IC, other watershed variablessuch as riparian continuity, natural forestcover, cropland, ditching and acid rain maybe better for predicting stream health.

• More research needs to be done to deter-mine the maximum level of watersheddevelopment at which stream diversity canbe restored or maintained. Additionally,the capacity of stormwater treatmentpractices and stream buffers to mitigatehigh levels of watershed IC warrants moresystematic research.

• More research is needed to test the ICM onamphibian and freshwater mussel diver-sity.


References

References

Aldridge, D., B. Payne and A. Miller. 1987.“The Effects of Intermittent Exposure toSuspended Solids and Turbulence onThree Species of Freshwater Mussels.”Environmental Pollution 45:17-28.

Allen, P. and R. Narramore. 1985. “BedrockControls on Stream Channel Enlarementwith Urbanization, North Central Texas.”American Water Resource Association21(6).

Allmendinger, N.L., J.E. Pizzuto, T.E. Johnsonand W.C. Hession. 1999. “Why Channelswith ‘Grassy’ Riparian Vegetation AreNarrower than Channels with ForestedRiparian Vegetation.” Eos (Transactions,American Geophysical Union), v. 80, FallMeeting Supplement, Abstract H32D-10.

Association of California Water Agencies(ACWA). 2000. Website. http://www.acwanet.com/news_info/testimony/tsca5-00.doc

Ayers, M., R. Brown and G. Oberts. 1985.Runoff and Chemical Loading in SmallWatersheds in the Twin Cities Metropoli-tan Area, Minnesota. U.S. GeologicalSurvey Water Resources InvestigationsReport 85-4122.

Bagley, S., M. Aver, D. Stern and M. Babiera.1998. “Sources and Fate of Giardia Cystsand Cryptosporidium Oocysts in SurfaceWaters.” Journal of Lake and ReservoirManagement 14(2-3): 379-392.

Baird, C., T. Dybala, M. Jennings andD.Ockerman. 1996. Characterization ofNonpoint Sources and Loadings to CorpusChristi National Estuary Program StudyArea. Corpus Christi Bay National EstuaryProgram. City of Corpus Christi, TX.

Bannerman, R., A. Legg and S. Greb. 1996.Quality of Wisconsin Stormwater 1989-1994. U.S. Geological Survey. Reston,VA.Open File Report 96-458.

Bannerman, R., D. Owens, R. Dodds and N.Hornewer. 1993. “Sources of Pollutants inWisconsin Stormwater.” Water Scienceand Technology 28(3-5): 241-259.

Barbour, M., J. Gerritsen, B. Snyder and J.Stribling. 1999. Rapid BioassessmentProtocols for Use in Streams and Wade-able Rivers: Periphyton, BenthicMacroinvertebrates and Fish. 2nd Edition.EPA 841-B-99-002. U.S. EPA Office ofWater. Washington, D.C.

Barfield, B. and M. Clar. 1985. Development ofNew Design Criteria for Sediment Trapsand Basins. Prepared for the MarylandResource Administration. Annapolis, MD.

Barr, R. 1997. Maryland NPDES Phase IMonitoring Data. Maryland Department ofthe Environment. Baltimore, MD.

Barrett, M. and J. Malina. 1998. Comparisonof Filtration Systems and VegetatedControls for Stormwater Treatment. 3rd

International Conference on DiffusePollution. Scottish Environment ProtectionAgency, Edinburg Scotland.

Barrett, M., R. Zuber, E. Collins and J. Malina.1995. A Review and Evaluation of Litera-ture Pertaining to the Quantity and Con-trol of Pollution from Highway Runoff andConstruction. CRWR Online Report 95-5.

Bauer, G. 1988. “Threats to the FreshwaterPearl Mussel Margaritifera margaritiferaL. in Central Europe.” Biological Conser-vation 45: 239-253.


References

Bay, S. and J. Brown. 2000. Assessment ofMTBE Discharge Impacts on CaliforniaMarine Water Quality. State Water Re-sources Control Board. Southern Califor-nia Coastal Water Research Project.Westminster, California.

Beeson, C. and P. Doyle. 1995. “Comparisonof Bank Erosion at Vegetated and Non-vegetated Bends.” Water ResourcesBulletin 31(6).

Benke, A., E. Willeke, F. Parrish and D. Stites.1981. Effects of Urbanization on StreamEcosystems. Office of Water Research andTechnology. US Department of the Inte-rior. Completion Report Project No. A-055-GA.

Bicknell, B., J. Imhoff, J. Kittle, A. Donigianand R. Johanson. 1993. Hydrologic Simu-lation Program-Fortran-HSPF. UsersManual for Release 10.0. EPA 600/3-84-066. Environmental Research Laboratory,U.S. EPA, Athens, GA.

Bisson, P., K. Sullivan, and J. Nielsen. 1988.“Channel Hydraulics, Habitat Use, andBody Form of Juvenile Coho Salmon,Steelhead, and Cutthroat Trout inStreams.” Transactions of the AmericanFisheries Society 117:262-273.

Black and Veatch. 1994. Longwell BranchRestoration-Feasibility Study Vol. 1.Carroll County, MD Office of Environ-mental Services.

Blankenship, K. 1996. “Masked Bandit Uncov-ered in Water Quality Theft.” Bay Journal6(6).

Bledsoe, B. 2001. “Relationships of StreamResponse to Hydrologic Changes.” LinkingStormwater BMP Designs and Perfor-mance to Receiving Water Impacts Mitiga-tion Proceedings. Snowmass, CO.

Blood, E. and P. Smith. 1996. “Water Qualityin Two High-Salinity Estuaries: Effects ofWatershed Alteration.” SustainableDevelopment in the Southeastern CoastalZone. F.J. Vernberg, W.B. Vernberg and T.Siewicki (eds.). Belle W. Baruch Libraryin Marine Science, No. 20. University ofSouth Carolina Press, Columbia, SC.

Boom, A. and J. Marsalek. 1988. “Accumula-tion of Polycyclic Aromatic Hydrocarbons.(PAHs) in an Urban Snowpack.” Scienceof the Total Environment 74:148.

Booth, D. 2000. “Forest Cover, ImperviousSurface Area, and the Mitigation ofUrbanization Impacts in King County,WA.” Prepared for King County Waterand Land Resource Division. University ofWashington.

Booth, D. 1991. “Urbanization and the NaturalDrainage System-Impacts, Solutions andPrognoses.” Northwest EnvironmentalJournal 7(1): 93-118.

Booth, D. 1990. “Stream Channel IncisionFollowing Drainage Basin Urbanization.”Water Resources Bulletin 26(3): 407-417.

Booth, D. and P. Henshaw. 2001. “Rates ofChannel Erosion in Small Urban Streams.”Water Science and Application 2:17-38.

Booth, D. and C. Jackson. 1997. “Urbaniza-tion of Aquatic Systems: DegradationThresholds, Stormwater Detection and theLimits of Mitigation.” Journal AWRA33(5): 1077- 1089.

Booth, D. and L. Reinelt. 1993. Consequencesof Urbanization on Aquatic Systems -Measured Effects, Degradation Thresh-olds, and Corrective Strategies. Watershed‘93 Proceedings. Alexandria, Virginia.


References

Booth, D., D. Montgomery and J. Bethel.1997. “Large Woody Debris in the UrbanStreams of the Pacific Northwest.” Effectsof Watershed Development and Manage-ment on Aquatic Ecosystems. Roesner,L.A. Editor. Proceedings of the ASCEConference. Snowbird, Utah.

Boughton, C. and M. Lico. 1998. VolatileOrganic Compounds in Lake Tahoe,Nevada and California. United StatesGeological Survey. Fact Sheet FS-055-98.

Boward, D., P. Kazyak, S. Stranko, M. Hurdand T. Prochaska. 1999. From the Moun-tains to the Sea: The State of Maryland’sFreshwater Streams. EPA 903-R-99-023.Maryland Deparment of Natural Re-sources. Annapolis, MD.

Bowen, J. and I. Valiela. 2001. “ The Ecologi-cal Effects of Urbanization of CoastalWatersheds: Historical Increases inNitrogen Loads and Eutrophication ofWaquiot Bay Estuaries.” Canadian Jour-nal of Fisheries and Aquatic Sciences58(8):1489-1500.

Brown, K. 2000. “Housing Density and UrbanLand Use as Stream Quality Indicators.”Watershed Protection Techniques 3(3):735-739.

Brown, W. 2000. “A Study of PairedCatchments Within Peavine Creek, Geor-gia.” Watershed Protection Techniques3(2):681-684.

Brush, S., M. Jennings, J. Young and H.McCreath. 1995. NPDES Monitoring –Dallas – Ft. Worth, Texas Area. In Storm-water NPDES Related Monitoring Needs.Proceedings of an Engineering FoundationConference. Edited by Harry Torno. NewYork, NY.

Campbell, K.R. 1995. “Concentrations ofHeavy Metals associated with UrbanRunoff in Fish Living in StormwaterPonds.” Archives of Environmental Con-tamination and Toxicology 27:352-356.

Cappiella, K. and K. Brown. 2001. ImperviousCover and Land Use in the ChesapeakeBay Watershed. Center for WatershedProtection. Ellicott City, MD.

Caraco, D. 2000a. “Stormwater Strategies forArid and Semi-arid Watersheds.” Water-shed Protection Techniques 3(3):695-706.

Caraco, D. 2000b. “The Dynamics of UrbanStream Channel Enlargement.” WatershedProtection Techniques 3(3):729-734.

Center for Watershed Protection (CWP). Inpress. Smart Watersheds: IntegratingLocal Programs to Achieve MeasurableProgress in Urban Watershed Restoration.Ellicott City, MD.

CWP. 2001a. “Managing Phosphorus InputsInto Lakes.” Watershed Protection Tech-niques 3(4): 769-796.

CWP. 2001b. Watts Branch Watershed Studyand Management Plan. Prepared for Cityof Rockville, Maryland. Ellicott City, MD.

CWP. 1998. Rapid Watershed PlanningManual. Ellicott City, MD.

Chang G., J. Parrish and C. Souer. 1990. TheFirst Flush of Runoff and its Effect onControl Structure Design. EnvironmentalResource Management Division - Depart-ment of Environmental and ConservationServices. Austin, TX.

Chessman, B., P. Hutton and J. Burch. 1992.“Limiting Nutrients for Periphyton Growthin Sub-alpine Forest, Agricultural andUrban Streams.” Freshwater Biology 28:349-361.

Chin, N. 1996. Watershed Urbanization Effectson Palustrine Wetlands: A Study of theHydrologic, Vegetative, and AmphibianCommunity Response Over Eight Years.M.S. Thesis. Department of Civil Engi-neering. University of Washington.


References

Collins, A, D. Walling and G. Leeks. 1997.“Source Type Ascription for FluvialSuspended Sediment Based on a Quantita-tive Composite Fingerprinting Technique.”Catena 29:1-27.

Connor, V. 1995. Pesticide Toxicity in Storm-water Runoff. Technical Memorandum.California Regional Water Quality ControlBoard, Central Valley Region. Sacramento,California.

Corsi, S., D. Hall and S. Geis. 2001. “Aircraftand Runway Deicers at General MitchellInternational Airport, Milwaukee, Wiscon-sin, USA. 2. Toxicity of Aircraft andRunway Deicers.” Environmental Toxicol-ogy and Chemistry 20(7):1483-1490.

Couch, C. et al. 1997. “Fish Dynamics inUrban Streams Near Atlanta, Georgia.”Technical Note 94. Watershed ProtectionTechniques. 2(4): 511-514.

Crawford, J. and D. Lenat. 1989. Effects ofLand Use on Water Quality and the Biotaof Three Streams in the Piedmont Provinceof North Carolina. United States Geologi-cal Service.Raleigh, NC. Water ResourcesInvestigations Report 89-4007.

Crippen and Waananen. 1969. HydrologicEffects of Suburban Development NearPalo Alto, California. Open file report.U.S. Geologic Survey, Menlo Park,Califronia.

Crunkilton, R., J. Kleist, J. Ramcheck, W.DeVita and D. Villeneuve. 1996. “Assess-ment of the Response of Aquatic Organ-isms to Long-term In Situ Exposures ofUrban Runoff.” Effects of WatershedDevelopment and Management on AquaticEcosystems. Roesner, L.A. Editor. Pro-ceedings of the ASCE Conference. Snow-bird, Utah.

Dail, H., P. Kazyak, D. Boward and S. Stranko.1998. Patapsco River Basin: Environmen-tal Assessment of Stream Conditions.Maryland Department of Natural Re-sources. Chesapeake Bay and WatershedPrograms CBP-MANTA-EA-98-4.

Dartiguenave, C., I. ECLille and D.Maidment. 1997. Water Quality MasterPlanning for Austin, TX. CRWR OnlineReport 97-6.

Davis, J. 2001. Personal communication.Department of Rural Sociology, Universityof Missouri, Columbia, MO.

Delzer, G.C. 1999. National Water-QualityAssessment Program: Quality of MethylTert-Butyl Ether (MTBE) Data forGround-water Samples Collected During1993-95. United States Geological SurveyFact Sheet. FS-101-99.

Delzer, G.C. 1996. Occurrence of the GasolineOxygenate MTBE and BTEX Compoundsin Urban Stormwater in the United States,1991-95. Untied States Geological SurveyWater-Resources Investigation Report.WRIR 96-4145.

Denver Regional Council of Governments(DRCOG). 1983. Urban Runoff Quality inthe Denver Region. Denver, CO.

DeVivo, J., C. Couch and B. Freeman. 1997.Use of Preliminary Index of Biotic Integ-rity in Urban Streams Around Atlanta,Georgia. Georgia Water Resources Con-ference. Atlanta, Georgia.

Dindorf, C. 1992. Toxic and HazardousSubstances in Urban Runoff. HennepinConservation District. Minnetonka, MN.

Doerfer, J. and B. Urbonas. 1993. StormwaterQuality Characterization in the DenverMetropolitan Area. Denver NPDES.Denver, CO.


References

Doll, B., D. Wise-Frederick, C. Buckner, S.Wilkerson, W. Harman and R. Smith.2000. “Hydraulic Geometry Relationshipsfor Urban Streams Throughout the Pied-mont of North Carolina.” Source unknown.

Donigian, A and W. Huber. 1991. Modeling ofNonpoint Source Water Quality in Urbanand Non-urban Areas. EPA/600/3-91/-39.U.S. EPA. Washington, D.C.

Dreher, D. 1997. “Watershed UrbanizationImpacts on Stream Quality Indicators inNortheastern Illinois.” Assessing theCumulative Impacts of Watershed Devel-opment on Aquatic Ecosystems and WaterQuality. D. Murray and R. Kirshner (ed.).Northeastern Illinois Planning Commis-sion. Chicago, IL.

Driscoll, E. 1983. Rainfall/ Runoff Relation-ships from the NURP Runoff Database.Stormwater and Quality Models UsersGroup Meeting. Montreal, Quebec.1983.

Driver, N. 1988. National Summary andRegression Models of Storm-Runoff Loadsand Volumes in Urban Watersheds in theUnited States. Thesis. Colorado School ofMines. Golden, Colorado.

Duda, A.M. and K.D. Cromartie. 1982.“Coastal Pollution from Septic TankDrainfields.” Journal of the Environmen-tal Engineering Division ASCE. 108:1265-1279.

Dunne, T. and L. Leopold. 1978. Water inEnvironmental Planning. W. Freeman andCompany, New York, NY.

Ellis, J. 1986. “Pollutional Aspects of UrbanRunoff.” In Urban Runoff Pollution. (eds.)H. Torno, J. Marsalek and M. Desbordes.Springer-Verlag, Berlin.

Environment Canada. 2001. Priority Sub-stances List Assessment Reports. RoadSalt. Ministry of Environment. Toronto,Canada.

EOA, Inc. 2001. Stormwater EnvironmentalIndicators. Pilot Demonstration Project.Final Report. Water Environment ResearchFoundation. Santa Clara Urban RunoffPollution Prevention Project. Santa Clara,CA.

Evaldi, R., R. Burns and B. Moore. 1992.Stormwater Data for Jefferson County,Kentucky, 1991-1992. U.S. GeologicalSurvey. Open File Report 92-638.

Evett, J., M. Love and J. Gordon. 1994. Effectsof Urbanization and Land Use Changes onLow Stream Flow. North Carolina WaterResources Research Institute. Report No.284.

Evgenidou, A., A. Konkle, A. D’Ambrosio, A.Corcoran, J. Bowen, E. Brown, D.Corcoran, C. Dearholt, S. Fern, A. Lamb,J. Michalowski, I. Ruegg and J. Cebrian.1997. “Effects of Increased NitrogenLoading on the Abundance of Diatoms andDinoflagellates in Estuarine Phytoplank-tonic Communities.” The BiologicalBulletin 197(2):292.

Fairfax County Department of Public Worksand Environmental Services (Fairfax Co).2001. Fairfax County Stream ProtectionStrategy Baseline Study. StormwaterManagement Branch, Stormwater PlanningDivision, Fairfax County, VA.

Ferrari, M., S. Altor, J. Blomquist and J.Dysart. 1997. Pesticides in the SurfaceWater of the Mid-Atlantic Region. UnitedStates Geological Survey. Water-Re-sources Investigations Report 97-4280.

Field, R. and R. Pitt. 1990. “Urban Storm-induced Discharge Impacts: US Environ-mental Protection Agency ResearchProgram Review.” Water Science Technol-ogy (22): 10-11.

Findlay, C. and J. Houlahan. 1997. “Anthropo-genic Correlates of Species Richness inSoutheastern Ontario Wetlands.” Conser-vation Biology 11(4):1000-1009.


References

Finkenbine, J., J. Atwater and D. Mavinic.2000. “Stream Health After Urbanization.”Journal of the American Water ResourcesAssociation 36(5): 1149-1160.

Fongers, D.and J. Fulcher. 2001. HydrologicImpacts Due to Development: The Needfor Adequate Runoff Detention and StreamProtection. Michigan Department ofEnvironmental Quality.

Fortner, A.R., M. Sanders and S.W. Lemire.1996. “Polynuclear Aromatic Hydrocarbonand Trace Metal Burdens in Sediment andthe Oyster, Crassostrea virginica Gmelin,from Two High-Salinity Estuaries in SouthCarolina.” In Sustainable Development inthe Southeastern Coastal Zone. F.J.Vernberg, W.B. Vernberg and T. Siewicki(eds.). Belle W. Baruch Library in MarineScience, No. 20. University of SouthCarolina Press, Columbia, SC.

Fulton, M., G. Chandler and G. Scott. 1996.“Urbanization Effects on the Fauna of aSoutheastern U.S.A. Bar-Built Estuary.” InSustainable Development in the Southeast-ern Coastal Zone. F.J. Vernberg, W.B.Vernberg and T. Siewicki (eds.). Belle W.Baruch Library in Marine Science, No. 20.University of South Carolina Press,Columbia, SC.

Galli, F. 1988. A Limnological Study of anUrban Stormwater Management Pond andStream Ecosystem. M.S. Thesis. GeorgeMason University.

Galli, J. 1994. Personal communication.Department of Environmental Programs.Metropolitan Washington Council ofGovernments. Washington, DC.

Galli, J. 1990. Thermal Impacts Associatedwith Urbanization and Stormwater Man-agement Best Management Practices.Metropolitan Washington Council ofGovernments. Maryland Department ofEnvironment. Washington, D.C.

Garie, H. and A. McIntosh. 1986. “Distributionof Benthic Macroinvertebrates in StreamsExposed to Urban Runoff.” Water Re-sources Bulletin 22:447-458.

Garn, H. 2002. Effects of Lawn Fertilizer onNutrient Concentrations in Runoff fromLakeshore Lawns, Lauderdale Lakes,Wisconsin. USGS Water-ResourcesInvestigation Report 02-4130.

Gavin, D. V. and R.K. Moore. 1982. Toxicantsin Urban Runoff. Prepared for the U.S.Environmental Protection Agency’sNationwide Urban Runoff Program.Seattle, WA.

Good, J. 1993. “Roof Runoff as a DiffuseSource of Metals and Aquatic Toxicologyin Stormwater.” Waterscience Technology28(3-5):317-322.

Graczyk, T. K., R. Fayer, J. M Trout, E. J.Lewis, C. A. Farley, I. Sulaiman and A.A.Lal. 1998. “Giardia sp. Cysts and Infec-tious Cryptosporidium parvum Oocysts inthe Feces of Migratory Canada geese(Branta canadensis).” Applied and Envi-ronmental Microbiology 64(7):2736-2738.

Haith, D., R. Mandel and R. Wu. 1992.GWLF-Generalized Watershed LoadingFunctions. Version 2.0 Users Manual.Cornell University. Agricultural Engineer-ing Department.

Hall, K. and B. Anderson. 1988. “The Toxicityand Chemical Composition of UrbanStormwater Runoff.” Canadian Journal ofCivil Engineering 15:98-106.

Hammer, T. 1972. “Stream Channel Enlarge-ment Due to Urbanization.” Water Re-sources Research 8(6): 1530-1540.

Hart, D. and C. Finelli. 1999. “Physical-Biological Coupling in Streams: thePervasive Effects of Flow on BenthicOrganisms.” Annu. Rev. Ecol. Syst.30:363-95.


References

Hartwell S., D. Jordahl, J. Evans and E. May.1995. “Toxicity of Aircraft De-icer andAnti-icer Solutions to Aquatic Organisms.”Environmental Toxicology and Chemistry14:1375-1386.

Heaney, J. and W. Huber. 1978. NationwideAssessment of Receiving Water Impactsfrom Urban Storm Water Pollution. UnitedStates Environmental Protection Agency.Cincinnati, OH.

Herlihy, A, J. Stoddard and C. Johnson. 1998.“The Relationship Between Stream Chem-istry and Watershed Land Cover in theMid-Atlantic Region, U.S.” Water, Air andSoil Pollution 105: 377-386.

Henshaw, P. and D. Booth, 2000. “NaturalRestabilization of Stream Channels inUrban Watersheds.” Journal of the Ameri-can Water Resources Association36(6):1219-1236.

Hession, W., J. Pizzuto, T. Johnson and R.Horowitz. In press. Influence of BankVegetation on Channel Morphology inRural and Urban Watersheds.

Hicks, A. and J. Larson. 1997. The Impact ofUrban Stormwater Runoff on FreshwaterWetlands and the Role of Aquatic Inverte-brate Bioassessment. The EnvironmentInstitute, University of Massachusetts.Amherst, MA.

Hirsch, R., J. Walker, J. Day and R. Kallio.1990. “The Influence of Man on Hydro-logic Systems.” Surface Water HydrologyO-1:329-347.

Hollis, F. 1975. “The Effects of Urbanizationon Floods of Different Recurrence Inter-vals.” Water Resources Research 11:431-435.

Holland, F., G. Riekerk, S. Lerberg, L.Zimmerman, D. Sanger, G. Scott, M.Fulton, B. Thompson, J. Daugomah, J.DeVane, K. Beck and A. Diaz. 1997. TheTidal Creek Project Summary Report.Marine Resources Research Institute, SCDepartment of Natural Resources.

Horkeby, B. and P. Malmqvist. 1977.Microsubstances in Urban Snow Water.IAHS-AISH. Publication 123:252-264.

Horner, R. and C. May. 1999. “RegionalStudy Supports Natural Land CoverProtection as Leading Best ManagementPractice for Maintaining Stream Ecologi-cal Integrity.” In: Comprehensive Storm-water & Aquatic Ecosystem ManagementConference Papers. First South PacificConference, February 22-26, New Zealand.Vol 1. p. 233-247.

Horner, R., D. Booth, A. Azous and C. May.1997. “Watershed Determinants of Ecosys-tem Functioning.” In Roesner, L.A.Editor. Effects of Watershed Developmentand Management on Aquatic Ecosystems.Proceedings of the ASCE Conference.Snowbird, Utah. 1996.

Horner, R., J. Guerdy and M. Kortenhoff.1990. Improving the Cost Effectiveness ofHighway Construction Site Erosion andPollution Control. Washington StateTransportation Center and the FederalHighway Administration. Seattle, WA.

Horner, R., C. May, E. Livingston and J.Maxted. 1999. “Impervious Cover,Aquatic Community Health, and Stormwa-ter BMPs: Is There a Relationship?” InProceedings of The Sixth Biennial Storm-water Research and Watershed Manage-ment Conference. Sept 14-17. 1999.Tampa Florida. Soutwest Florida WaterManagement District. Available on-line:http://www.stormwater-resources.com/proceedings_of_the_sixth_biennia.htm


References

Horner, R., C. May, E. Livingston, D. Blaha,M. Scoggins, J. Tims and J. Maxted. 2001.“Structural and Non-structural BMPs forProtecting Streams.” in Linking Stormwa-ter BMP Designs and Performance toReceiving Water Impact Mitigation. B.Urbonas (editor). Proceedings of anEngineering Research Foundation Confer-ence. Smowmass, CO. American Societyof Civil Engineers (TRS). pp. 60-77.

Huber, W. and R. Dickinson. 1988. StormWater Management Model (SWMM).Version 4. Users Manaul. EPA/600/3-88/001a). US EPA. Athens, GA.

Hydrologic Engineering Center (HEC). 1977.Storage, Treatment, Overflow and RunoffModel (STORM). Users Manual. General-ized Computer Program. 7233-S8-L7520.

Hydroqual, Inc. 1996. Design Criteria Report:Kensico Watershed Stormwater BestManagement Facilities: Appendix C.Report prepared for City of New York.Department of Environmental Protection.

Immerman, F. and D. Drummon. 1985. Na-tional Urban Pesticide ApplicationsSurvey. Research Triangle Institute.Publication No. 2764/08-01F.

Ireland, D., G. Burton, Jr. and G. Hess. 1996.“In Situ Toxicity Evaluations of Turbidityand Photoinduction of Polycyclic AromaticHydrocarbons.” Environmental Toxicologyand Chemistry 15(4): 574-581.

Islam, M., D. Tuphron and H. Urata-Halcomb.1988. Current Performance of SedimentBasins and Sediment Yield Measurement inUnincorporated Hamilton County, OH.Hamilton County Soil and Water Conser-vation District.

Jarrett, A. 1996. Sediment Basin Evaluationand Design Improvements. PennsylvaniaState University. Prepared for OrangeCounty Board of Commissioners.

Johnson, K. 1995. Urban Storm Water Impactson a Coldwater Resource. Presentation tothe Society of Environmental Toxicologyand Chemistry (SETAC) Second WorldCongress. Vancouver, B.C., Canada..

Jones, R., A. Via-Norton and D. Morgan. 1996.“Bioassessment of the BMP Effectivenessin Mitigating Stormwater Impacts onAquatic Biota.” In Roesner, L.A. Editor.Effects of Watershed Development andManagement on Aquatic Ecosystems.Proceedings of the ASCE Conference.Snowbird, Utah.

Jones, R. and C. Clark. 1987. “Impact ofWatershed Urbanization on Stream InsectCommunities.” Water Resources Bulletin15(4).

Karouna-Renier, N. 1995. An Assessment ofContaminant Toxicity to Aquatic Macro-Invertebrates in Urban Stormwater Treat-ment Ponds. M.S. Thesis. University ofMaryland. College Park, MD.

Keller, A. and S. Zam. 1991. “The AcuteToxicity of Selected Metals to the Fresh-water Mussel, Anodonta imbecillis.”Environmental Toxicology and Chemistry10: 539-546.

Kemp, S. and J. Spotila. 1997. “Effects ofUrbanization on Brown Trout Salmo trutta,Other Fishes and Macroinvertebrates inValley Creek, Valley Forge, PA.” Ameri-can Midl. Nat. 138:55-68.

Kennen, J. 1999. “Relation ofMacroinvertebrate Community Impairmentto Catchment Characteristics in NewJersey Streams.” Journal of the AmericanWater Resources Association 35(4):939-955.

Kibler, D., D. Froelich and G. Aron. 1981.“Analyzing Urbanization Impacts onPennsylvania Flood Peaks.” Water Re-sources Bulletin 17(2):270-274.


References

Kitchell, A. 2001. “Managing for a Pure WaterSupply.” Watershed Protection Techniques3(4): 797-812.

Kjelstrom, L. 1995. Data for Adjusted Re-gional Regression Models of Volume andQuality of Urban Stormwater Runoff inBoise and Garden City, Idaho, 1993-94.United States Geological Survey. WaterResources Investigations Report 95-4228.

Klein, R. 1979. “Urbanization and StreamQuality Impairment.” Water ResourcesBulletin 15(4):948-963.

Konrad, C. and D. Booth. 2002. HydrologicTrends Associated with Urban Develop-ment for Selected Streams in the PugetSound Basin - Western Washington. USGSWater Resources Investigation Report 02-4040.

Kucklick, J.K., S. Silversten, M. Sanders andG.I. Scott. 1997. “Factors InfluencingPolycyclic Aromatic Hydrocarbon Distri-butions in South Carolina EstuarineSediments.” Journal of ExperimentalMarine Biology and Ecology 213:13-30.

Kundell, J. and T. Rasmussen. 1995. Recom-mendations of the Georgia Board ofRegent’s Scientific Panel on Evaluatingthe Erosion Measurement StandardDefined by the Georgia Erosion andSedimentation Act. Proceedings of the1995 Georgia Water Resources Confer-ence. Athens, Georgia.

La Barre, N., J. Milne and B. Oliver. 1973.“Lead Contamination of Snow.” WaterResearch 7:1,215-1,218.

LeBlanc, R., R. Brown and J. FitzGibbon.1997. “Modeling the Effects of Land UseChange on the Water Temperature inUnregulated Urban Streams.” Journal ofEnvironmental Management 49: 445-469.

Leopold, L. 1994. A View of the River. HarvardUniversity Press, Cambridge, MA.

Leopold, L. 1973. “River Change with Time:An Example.” Geological Society ofAmerica Bulletin 84: 1845-1860.

Leopold, L. 1968. Hydrology for Urban landUse Planning - A Guidebook on theHydrologic Effects of Urban Land Use.Washington, D.C. Geological SurveyCircular 554.

Lerberg, S., F. Holland and D. Sanger. 2000.“Responses of Tidal Creek MacrobenthicCommunities to the Effects of WatershedDevelopment.” Estuaries 23(6):838-853.

Lim, S. and V. Olivieri. 1982. Sources ofMicroorganisms in Urban Runoff. JohnsHopkins School of Public Health andHygiene. Jones Falls Urban RunoffProject. Baltimore, MD.

Limburg, K. and R. Schmidt. 1990. “Patternsof Fish Spawning in Hudson River Tribu-taries-Response to an Urban Gradient?”Ecology 71(4): 1231-1245.

Liu, Z., D. Weller, D. Correll and T. Jordan.2000. “Effects of Land Cover and Geologyon Stream Chemistry in Watersheds ofChesapeake Bay.” Journal of AmericanWater Resources Association 36(6): 1349-1365.

Livingston, R. 1996. “Eutrophication inEstuaries and Coastal Systems: Relation-ship of Physical Alterations, SalinityStratification, and Hypoxia.” In Sustain-able Development in the SoutheasternCoastal Zone. F.J. Vernberg, W.B.Vernberg and T. Siewicki (eds.). Belle W.Baruch Library in Marine Science, No. 20.University of South Carolina Press,Columbia, SC.

Lopes, T. and S. Dionne. 1998. A Review ofSemi-Volatile and Volatile OrganicCompounds in Highway Runoff and UrbanStormwater. USGS Open file report 98-409.


References

Lopes, T., K. Fossum, J. Phillips and J.Marical. 1995. Statistical Summary ofSelected Physical, Chemical, and Micro-bial Contaminants and Estimates ofConstituent Loads in Urban Stormwater inMaricopa County, Arizona. USGS WaterResources Investigations Report 94-4240.

Luchetti, G. and R. Feurstenburg. 1993.Relative Fish Use in Urban and Non-urbanStreams Proceedings. Conference on WildSalmon. Vancouver, British Columbia.

MacCoy, D. and R. Black. 1998. OrganicCompounds and Trace Elements in Fresh-water Streambed Sediment and Fish fromthe Puget Sound Basin. USGS Fact Sheet105-98.

MacRae, C. 1996. “Experience From Morpho-logical Research on Canadian Streams: IsControl of the Two-year Frequency RunoffEvent the Best Basis for Stream ChannelProtection?” In Roesner, L.A. Editor.Effects of Watershed Development andManagement on Aquatic Ecosystems.Proceedings of the ASCE Conference.Snowbird, Utah.

MacRae, C. and M. DeAndrea. 1999. Assess-ing the Impact of Urbanization on ChannelMorphology. 2nd International Conferenceon Natural Channel Systems. Niagra Falls,OT.

MacRae, C. and A. Rowney. 1992. The Role ofModerate Flow Events and Bank Structurein the Determination of Channel Responseto Urbanization. 45th Annual Conference.Resolving Conflicts and Uncertainty inWater Management. Proceeding of theCanadian Water Resources Association,Kingston, Ontario.

Maiolo, J. and P. Tschetter. 1981. “RelatingPopulation Growth to Shellfish BedClosures: a Case Study from North Caro-lina.” Coastal Zone Management Journal9(1).

Mallin, M., E. Esham, K. Williams and J.Nearhoof. 1999. “Tidal Stage Variabilityof Fecal Coliform and Chlorophyll aConcentrations in Coastal Creeks.” MarinePollution Bulletin 38 (5):414-422.

Mallin, M., K. Williams, E. Esham and R.Lowe. 2000. “Effect of Human Develop-ment on Bacteriological Water Quality inCoastal Watersheds.” Ecological Applica-tions 10(4) 1047-1056.

Mallin, M., S. Ensign, M. McIver, G. Swankand P. Fowler. 2001. “Demographic,Landscape and Metrologic Factors Con-trolling the Microbial Pollution of CoastalWaters.” Hydrobiologia 460:185-193.

Malmqvist, P. 1978. “Atmospheric Fallout andStreet Cleaning- Effects on Urban SnowWater and Snow.” Progress in WaterTechnology 10(5/6):495-505.

Maryland-National Capital Park and PlanningCommission (MNCPPC). 2000. StreamCondition Cumulative Impact Models Forthe Potomac Subregion. Prepared for theMaryland- National Park and PlanningCommission, Silver Spring, MD.

Masterson, J. and R. Bannerman. 1994. Impactof Stormwater Runoff on Urban Streams inMilwaukee County, Wisconsin. WisconsinDepartment of Natural Resources. Madi-son, WI.

Maxted, J. 1999. “The Effectiveness of Reten-tion Basins to Protect Downstream AquaticLife in Three Regions of the UnitedStates.” In Conference Proceedings.Volume one. Comprehensive Stormwaterand Aquatic Ecosystem Management. FirstSouth Pacific Conference. 22-26 February,1999. Auckland Regional Council.Auckland, New Zealand pp. 215- 222.


References

Maxted, J. and E. Shaver. 1997. “The Use ofRetention Basins to Mitigate StormwaterImpacts on Aquatic Life.” In Roesner, L.A.Editor. Effects of Watershed Developmentand Management on Aquatic Ecosystems.Proceedings of the ASCE Conference.Snowbird, Utah.

May, C., R. Horner, J. Karr, B. Mar and E.Welch. 1997. “Effects of Urbanization onSmall Streams in the Puget Sound Low-land Ecoregion.” Watershed ProtectionTechniques 2(4): 483-494.

McCuen R. and G. Moglen. 1988.“Multicriterion Stormwater ManagementMethods.” Journal of Water ResourcesPlanning and Management 4 (114).

Menzie-Cura & Associates. 1995. Measure-ments and Loadings of Polycyclic Aro-matic Hydrocarbons (PAH) in Stormwater,Combined Sewer Overflows, Rivers, andPublically Owned Treatment Works(POTWs) Discharging to MassachusettsBays. Report to the Massachusetts BayProgram, August 1995, MBP-95-06.

Metropolitan Washington Council of Govern-ments (MWCOG). 1992. WatershedRestoration Sourcebook. Department ofEnvironmental Programs. MWCOG,Washington, DC.

MWCOG. 1983. Urban runoff in WashingtonMetropolitan Area- Final Report. Wash-ington. D.C Area Urban Runoff Program.Prepared for USEPA. WRPB.

Meyer, J. and C. Couch. 2000. Influences ofWatershed Land Use on Stream EcosystemStructure and Function. NCERQA GrantFinal Report.

Montgomery County Department of Environ-mental Protection (MCDEP). 2000. SpecialProtection Area Report.

Morisawa, M. and E. LaFlure. 1979. HydraulicGeometry, Stream Equalization andUrbanization. Proceedings of the TenthAnnual Geomorphology Symposia Series:Adjustments of the Fluvial System.Binghamton, New York.

Moring, J. and D. Rose. 1997. “Occurrence andConcentrations of Polycyclic AromaticHydrocarbon in Semipermeable MembraneDevices and Clams in Three UrbanStreams of the Dallas-Fort Worth Metro-politan Area, Texas.” Chemosphere 34(3):551-566.

Morse, C. 2001. The Response of First andSecond Order Streams to Urban Land-usein Maine, USA. Masters Thesis, TheUniversity of Maine, Orono, ME.

Moscript, A. and D. Montgomery. 1997.“Urbanization, Flood Frequency, andSalmon Abundance in Puget LowlandStreams.” Journal of the American WaterResources Association. 33:1289-1297.

Neller, R. 1989. “Induced Channel Enlarge-ment in Small Urban Catchments,Armidale, New South Wales.” Environ-mental Geology and Water Sciences 14(3):167-171.

Neller, R. 1988. “A Comparison of ChannelErosion in Small Urban and RuralCatchments, Armidale, New SouthWales.” Earth Surface Processes andLandforms 13:1-7.

New Mexico Surface Water Quality Bureau(NMSWQB). 1999. Total Maximum DailyLoad for Fecal Coliform on CanadianRiver Basin Six Mile, Cieneguilla andMoreno Creeks (Cimarron).

Northern Virginia Planning District Commis-sion (NVPDC). 1987. BMP Handbook forthe Occoquan Watershed. Annandale, VA.

NVPDC. 1979. Guidebook for EvaluatingUrban Nonpoint Source Strategies. Pre-pared for the Metropolitan WashingtonCouncil of Governments.


References

Northern Virginia Regional Commission(NVRC). 2001. The Effect of Urbanizationon the Natural Drainage Network in theFour Mile Run Watershed.

Novotny, V. and G. Chester. 1989. “Deliveryof Sediment and Pollutants from NonpointSources: a Water Quality Perspective.”Journal of Soil and Water Conservation44:568-576.

Novotny, V. and G. Chester. 1981. Handbookof Nonpoint Pollution: Sources andManagement. Van Nostrand ReinholdCompany. NY.

Novotny, V., D. W. Smith, D. A. Kuemmel, J.Mastriano and A. Bartosova. 1999. Urbanand Highway Snowmelt: Minimizing theImpact on Receiving Water. Water Envi-ronment Research Foundation. Alexandria,VA.

Obert, G. 1999. “Return to Lake McCarrons:Does the Performance of Wetlands Holdup Over Time?” Watershed ProtectionTechniques 3(1):597-600.

Oberts, G. 1994. “Influence of SnowmeltDynamics on Stormwater Runoff Quality.”Watershed Protection Techniques 1(2):55-61.

Oberts, G., P. Wotzka and J. Hartsoe.1989. TheWater Quality Performance of SelectUrban Runoff Treatment Systems. Metro-politan Council, St. Paul, MN. Publ. No.590-89-062a.

Occoquan Watershed Monitoring Lab(OWML). 1983. Washington Area NURPReport VPISU: Final Report. MetropolitanWashington Council of Governments.Manassas, VA.

Ohrel, R. 1995. “Rating Deicer Agents – RoadSalts Stand Firm.” Watershed ProtectionTechniques 1(4):217-220.

Oliver, G., P. Milene and N. La Barre. 1974.“Chloride and Lead in Urban Snow.”Journal Water Pollution Control Federa-tion 46(4):766-771.

Overton, D. and M. Meadows. 1976. StormWater Modeling. Academic Press. NewYork, NY.

Paul, M., D. Leigh and C. Lo. 2001. Urbaniza-tion in the Etowwah River Basin: Effectson Stream Temperature and Chemistry.Proceedings of the 2001 Georgia WaterResources Conference. University ofGeorgia, Athens, GA.

Pedersen, E. and M. Perkins. 1986. “The Useof Benthic Invertebrate Data for Evaluat-ing Impacts of Urban Runoff.”Hydrobiologia 139: 13-22.

Pierstorff, B. and P. Bishop. 1980. “ WaterPollution From Snow Removal Opera-tion.” Journal of Environmental Engineer-ing Division 106 (EE2):377-388.

Pitt, R. 1998. “Epidemiology and StormwaterManagment.” In Stormwater QualityManagement. CRC/Lewis publishers. NewYork, NY.

Pitt, R. and M. Bozeman. 1982. “Sources ofUrban Runoff Pollution and Its Effects onan Urban Creek.” EPA-600/52-82-090.U.S. Environmental Protection Agency.Cincinnati, OH.

Pitt, R. and J. Voorhees. 1989. Source Loadand Management Model (SLAMM) – AnUrban Nonpoint Source Water QualityModel. Wisconsin Department of NaturalResources, v. I-III, PUBL-WR-218-89.

Pizzuto, J., W. Hession and M. McBride. 2000.“Comparing Gravel-Bed Rivers in PairedUrban and Rural Catchments of Southeast-ern Pennsylvannia.” Geology 28(1):79-82.


References

Poff, N. and J. Ward. 1989. “Implications ofStream Flow Variability and Predictabilityfor Lotic Community Structure: A Re-gional Analysis of Streamflow Patterns.”Canadian Journal of Fisheries and AquaticScience 46:1805-1818.

Porter, D.E., D. Edwards, G. Scott, B. Jonesand S. Street. 1997. “Assessing the Im-pacts of Anthropogenic and PhysiographicInfluences on Grass Shrimp in LocalizedSalt Marsh Estuaries: a Multi-DisciplinaryApproach.” Aquatic Botany 58:289-306.

Rabanal, F. and T. Grizzard. 1995.“Concentra-tions of Selected Constituents in Runofffrom Impervious Surfaces in UrbanCatchments of Different Land Use.” InProceedings of the 4th Biennial Conferenceon Stormwater Research. Oct 18-20.Clearwater, Florida. Southwest FloridaWater Management District. pp. 42-52.

Reice, S. 2000. “Regulating Sedimentation andErosion Control into Streams: What ReallyWorks and Why?” In Proceedings of theNational Conference on Tools for UrbanWater Resource Management & Protec-tion. Published by the US EnvironmentalProtection Agency, Office of Research andDevelopment, Washington, D.C.

Richards, C., L. Johnson and G. Host. 1993.Landscape Influence on Habitat, WaterChemistry, and Macroinvertebrate Assem-blages in Midwestern Stream Ecosystems.Center for Water and the Environment.Natural Resources Research Institute(NRRI) Technical Report TR-93-109.

Richey, J. 1982. Effects of Urbanization on aLowland Stream in Urban Washington.PhD Dissertation. University of Washing-ton.

Roth, N., M. Southerland, D. Stebel and A.Brindley. 1998. Landscape Model ofCumulative Impacts: Phase I Report.Maryland Department of Natural Re-sources.

Roth, N., J. David and D. Erickson. 1996.“Landscape Influences on Stream BioticIntegrity Assessed at Multiple SpatialScales.” Landscape Ecology 11(3):141-156.

Rutkowski, C., W. Burnett, R. Iverson and J.Chanton. 1999. “The Effect of Groundwa-ter Seepage on Nutrient Delivery andSeagrass Distribution in the NortheasternGulf of Mexico.” Estuaries 22(4):1033-1040.

S.R. Hanson and Associates. 1995. FinalReport: Identification and Control ofToxicity in Stormwater Discharges toUrban Creeks. Prepared for AlamedaCounty Urban Runoff Clean Water Pro-gram.

Samadpour, M. and N. Checkowitz. 1998.“Little Soos Creek Microbial SourceTracking.” Washington Water Resource.(Spring) University of Washington UrbanWater Resources Center.

Salt Institute. 2001. Data on U.S. Salt Sales.Available on-line: www.saltinstitute.org

Sanger, D., F. Holland and G. Scott. 1999.“Tidal Creek and Salt Marsh Sediment inSouth Carolina Coastal Estuaries: I.Distribution of Trace Metals.” Archive ofEnvironmental Contamination and Toxi-cology (37):445-457.

Saravanapavan, T. 2002. Personal communica-tion.

Sauer, V., T. Stricker and K. Wilson. 1983.Flood Characteristics of Urban Water-sheds in the United States. US GeologicalSurvey Water Supply Paper 2207.

Scanlin, J. and A. Feng. 1997. Characteriza-tion of the Presence and Sources ofDiazinon in the Castro Valley CreekWatershed. Alameda Countywide CleanWater Program and Alameda CountyFlood Control and Water ConservationDistrict, Oakland, CA.


References

Schiff, K. 1996. Review of Existing StormwaterMonitoring Programs for EstimatingBight-Wide Mass Emissions from UrbanRunoff.

Schrimpff, E. and R. Herrman. 1979. “Re-gional Patterns of Contaminants (PAH,Pesticides and Trace metals) in Snow ofNortheast Bavaria and their Relationship toHuman Influence and OrogeographicEffects.” Water, Air and Soil Pollution11:481-497.

Schueler, T. 2001. “The Environmental Impactof Stormwater Ponds.” The Practice ofWatershed Protection. T. Schueler and H.Holland (Eds). Center for WatershedProtection. Ellicott City, MD.

Schueler, T. 2000. “The Compaction of UrbanSoils.” Watershed Protection Techniques3(3):661-665.

Schueler, T. 1999. “Microbes and UrbanWatersheds.” Watershed ProtectionTechniques 3(1): 551-596.

Schueler, T. 1994a. “The Importance ofImperviousness.” Watershed ProtectionTechniques 2(4): 100-111.

Schueler, T. 1994b. “Pollutant Dynamics ofPond Muck.” Watershed ProtectionTechniques 1(2).

Schueler, T. 1987. Controlling Urban Runoff:a Practical Manual for Planning andDesigning Urban Best ManagementPractices. Metropolitan WashingtonCouncil of Governments. Washington,D.C.

Schueler, T. and D. Caraco. 2001. “TheProspects for Low Impact Land Develop-ment at the Watershed Level.” In LinkingStormwater BMP Designs and Perfor-mance to Receiving Water Impacts Mitiga-tion. United Engineering Foundation.Snowmass, CO.

Schueler, T. and J. Galli. 1992. “Environmen-tal Impacts of Stormwater Ponds.” Water-shed Restoration Sourcebook. AnacostiaRestoration Team Metropolitan Washing-ton Council Government. Washington D.C.

Schueler, T. and H. Holland. 2000. The Prac-tice of Watershed Protection- Techniquesfor Protecting Our Nations, Streams,Rivers, Lakes and Estuaries. Center forWatershed Protection. Ellicott City, MD.

Schueler, T. and J. Lugbill. 1990. Performanceof Current Sediment Control Measures atMaryland Construction Sites. MetropolitanWashington Council of Governments(MWCOG).

Schueler, T. and D. Shepp. 1993. The Quantityof Trapped Sediments in Pool WaterWithin Oil Grit Separators in SuburbanMD. Metropolitan Washington Council ofGovernments (MWCOG).

Scott, J., C. Steward and Q. Stober. 1986.“Effects of Urban Development on FishPopulation Dynamics in Kelsey Creek,Washington.” Transactions of the Ameri-can Fisheries Society. 115:555-567.

Scott, W. and N. Wylie. 1980. “The Environ-mental Effects of Snow Dumping: ALiterature Review.” Journal of Environ-mental Management 10:219-240.

Shaver, E., J. Maxted, G. Curtis and D. Carter.1995. “Watershed Protection Using anIntegrated Approach.” In B. Urbonas andL. Roesner Editors. Stormwater NPDES-related Monitoring Needs. Proceedings ofan Engineering Foundation Conference.Crested Butte, CO.

Sherman, K. 1998. Severn Sound UrbanStormwater Pollution Control PlanningReport. Ontario, Canada.

Short, F. and S. Wyllie-Echeverria. 1996.“Natural and Human-Induced Disturbanceof Seagrasses.” Environmental Conserva-tion 23(1): 17-27.


References

Sills, R. and P. Blakeslee. 1992. “Chapter 11:The Environmental Impact of Deicers inAirport Stormwater Runoff.” ChemicalDeicers and the Environment. LewisPublishers. Ann Arbor, MI.

Silverman, G., M. Stenstrom and S. Fam.1988. “Land Use Considerations in Reduc-ing Oil and Grease in Urban StormwaterRunoff.” Journal of Environmental Sys-tems 18(1): 31-46.

Simmons, D and R. Reynolds. 1982. “Effectsof Urbanization on Baseflow of SelectedSouth-Shore Streams, Long Island, NY.”Water Resources Bulletin 18(5): 797-805.

Smullen, J. and K. Cave. 1998. Updating theU.S. Nationwide Urban Runoff QualityDatabase. 3rd International Conference onDiffuse Pollution. Scottish EnvironmentProtection Agency, Edinburg Scotland.1998.

Spence, B., G. Lomnicky, R. Hughes and R.Novitzki. 1996. An Ecosystem Approach toSalmonid Conservation. TR-401-96-6057.ManTech Environmental Research Ser-vices Corporation, Corvallis, OR. (Avail-able on the NMFS-NWR website:

Spinello, A.G. and D.L. Simmons. 1992.Baseflow of 10 South Shore Streams, LongIsland, New York 1976-85 and the Effectsof Urbanization on Baseflow and FlowDuration. USGS. Water Resources Investi-gation Report 90-4205.

Squillace, P., D. Pope and C.V. Price. 1995.Occurrence of the Gasoline AdditiveMTBE in Shallow Groundwater in Urbanand Agricultural Areas. USGS Fact Sheet114-95.

States, S., K. Stadterman, L. Ammon, P.Vogel, J. Baldizar, D. Wright, L. Conleyand J. Sykora. 1997. “Protozoa in RiverWater: Sources, Occurrence and Treat-ment” Journal of the American WaterWorks Association 89(9):74-83.

Steedman, R. J. 1988. “Modification andAssessment of an Index of Biotic Integrityto Quantify Stream Quality in SouthernOntario.” Canadian Journal of Fisheriesand Aquatic Sciences 45:492-501.

Stern, D. 1996. “Initial Investigation of theSources and Sinks of Cryptosporidium andGiardia Within the Watersheds of the NewYork City Water Supply System.” InMcDonnel et al. Editors. New York CityWater Supply Studies. Proceedings of anAmerican Water Resources AssociationSymposium. Herndon, VA.

Steuer, J., W. Selbig, N. Hornewer and JeffreyPrey. 1997. Sources of Contamination inan Urban Basin in Marquette, Michiganand an Analysis of Concentrations, Loads,and Data Quality. U.S. Geological Survey,Water-Resources Investigations Report 97-4242.

Steward, C. 1983. Salmonoid Populations inan Urban Environment—Kelsey Creek.,Washington. Masters Thesis. University ofWashington.

Stormwater Assessment Monitoring Perfor-mance (SWAMP). 2000a. PerformanceAssessment of a Highway StormwaterQuality Retention Pond - Rouge River,Toronto, Ontario. SWAMP Program.Ontario Ministry of the Environment.Toronto and Region Conservation Author-ity. Toronto, Canada.

SWAMP. 2000b. Performance Assessment of aStormwater Retrofit Pond - Harding Park,Richmond Hill, Ontario. SWAMP Pro-gram. Ontario Ministry of the Environ-ment. Town of Richmond Hill. Torontoand Region Conservation Authority.Toronto, Canada.

Stranko, S. and W. Rodney. 2001. HabitatQuality and Biological Integrity Assess-ment of Freshwater Streams in the SaintMary’s River Watershed. MarylandDepartment of Natural Resources. CBWP-MANTA-EA-01-2.


References

Stribling, J., E. Leppo, J. Cummins, J. Galli, S.Meigs, L. Coffman and M.Cheng. 2001.“Relating Instream Biological Conditionsto BMP Activities in Streams and Water-sheds.” In Linking Stormwater BMPDesigns and Performance to ReceivingWater Impacts Mitigation. United Engi-neering Foundation. Snowmass, CO.

Sturm, T. and R. Kirby. 1991. SedimentReduction in Urban Stormwater Runofffrom Construction Sites. Georgia Instituteof Technology. Atlanta, GA.

Swann, C. 2001. “The Influence of SepticSystems at the Watershed Level.” Water-shed Protection Techniques 3(4):821-834.

Swietlik, W. 2001. “Urban Aquatic Life Uses -a Regulatory Perspective.” In LinkingStormwater BMP Designs and Perfor-mance to Receiving Water Impacts Mitiga-tion. United Engineering Foundation.Snowmass, CO.

Taylor, B.L. 1993. The Influences of Wetlandand Watershed Morphological Character-istics and Relationships to Wetland Veg-etation Communities. Master’s Thesis.Dept. of Civil Engineering. University ofWashington, Seattle, WA.

Taylor, B., K. Ludwa and R. Horner. 1995.Third Puget Sound Research MeetingUrbanization Effects on Wetland Hydrol-ogy and Water Quality. Proceedings of thePuget Sound Water Quality AuthorityMeeting. Olympia, WA.

Thomas, P. and S. McClelland. 1995. “NPDESMonitoring - Atlanta, Georgia Region.” InStormwater NPDES Related MonitoringNeeds. Proceedings of an EngineeringFoundation Conference. Edited by HarryTorno. New York, NY.

Trimble, S. 1997. “Contribution of StreamChannel Erosion to Sediment Yield froman Urbanizing Watershed.” Science 278:1442-1444.

U.S. Department of Agriculture (USDA).1992. Computer Program for ProjectFormulation-Hydrology (TR-20). Hydrol-ogy Unit. Natural Resources ConservationService. Washington, D.C.

USDA. 1986. Urban Hydrology for SmallWatersheds. Technical Release 55. (TR-55). Soil Conservation Service Engineer-ing Division. Washington, D.C.

United States Environmental ProtectionAgency (USEPA). 2000. Mid-AtlanticIntegrated Assessment (MAIA) ProjectSummary: Birds Indicate EcologicalCondition of the Mid-Atlantic Highlands.U.S.EPA, Office of Research and Develop-ment, Washington, DC.

USEPA. 1998. The Quality of Our Nation’sWaters: 1996. U.S.EPA, Office of Water,Washington, DC. EPA-841-S-97-001.

USEPA. 1993. Guidance Specifying Manage-ment Measures for Sources of Non-pointPollution in Coastal Waters. U.S.EPA,Office of Water, Washington, DC. 840-B-92-002.

USEPA. 1988. Dissolved Solids. Water QualityStandards Criteria Summaries: A Compi-lation of State/Federal Criteria. Office ofWater, Regulations, and Standards, Wash-ington, DC.

USEPA. 1983. Results of the NationwideUrban Runoff Project: Final Report.U.S.EPA, Office of Water, Washington,DC.

USEPA. 1982. Sources of Urban RunoffPollution and Its Effects on an UrbanCreek. U.S.EPA, Washington, DC. -600/S2-82-090.

United States Geological Survey (USGS).2001a. Selected Findings and CurrentPerspectives on Urban and AgriculturalWater Quality. National Water-QualityAssessment Program. USGS Fact Sheet.FS-047-01.


References

USGS. 2001b. The Quality of Our Nation’sWaters: Nutrients and Pesticides. USGS.FS-047-01.

USGS. 1999a. Pesticides and Bacteria in anUrban Stream - Gills Creek, Columbia,South Carolina. USGS. Fact Sheet FS-131-98.

USGS. 1999b. Pesticides Detected in UrbanStreams During Rainstorms and Relationsto Retail Sales of Pesticides in KingCounty, Washington. USGS. Fact Sheet097-99.

USGS. 1998. Pesticides in Surface Waters ofthe Santee River Basin and CoastalDrainages, North and South Carolina.USGS Fact Sheet. FS-007-98.

USGS. 1996. Effects of Increased Urbanizatonfrom 1970s to 1990s on Storm RunoffCharacteristics in Perris Valley, Califor-nia. USGS Water Resources InvestigationsReport 95-4273.

University of California at Davis (UC Davis).1998. UC Report: MTBE Fact Sheet.Available online: http://tsrtp.ucdavis.edu/mtberpt

Valiela, I., J. McClelland, J. Hauxwell, P.Behr, D. Hersh and K. Foreman. 1997.“Macroalgal Blooms in Shallow Estuaries:Controls and Ecophysiological and Eco-system Consequences.” Limonology andOceanography 42(5, part 2): 1105-1118.

Valiela, I. and J. Costa. 1988. “Eutrophicationof Buttermilk Bay, a Cape CodEmbaymnet: Concentrations of Nutrientsand Watershed Nutrient Budgets.” Envi-ronmental Management 12(4):539-553.

Van Loon, J. 1972. “The Snow RemovalControversy.” Water Pollution Control110:16-20.

Varner, 1995. Characterization and SourceControl of Urban Stormwater Quality. Cityof Bellevue Utilities Department. City ofBellevue, Washington.

Velinsky, D. and J.Cummins. 1994. Distribu-tion of Chemical Contaminants in WildFish Species in the Washington, D.C. Area.Interstate Commission on the PotomacRiver Basin, ICPRB., Rockville, MD.Report No. 94-1.

Vernberg, W., G. Scott, S. Stroizer, J. Bemissand J. Daugomah. 1996a. “The Effects ofCoastal Development on WatershedHydrography and Transport of OrganicCarbon.” In Sustainable Development inthe Southeastern Coastal Zone. F.J.Vernberg, W.B. Vernberg and T. Siewicki(eds.). Belle W. Baruch Library in MarineScience, No. 20. University of SouthCarolina Press, Columbia, SC.

Vernberg, F., W. Vernberg and T. Siewicki.1996b. Sustainable Development in theSoutheastern Coastal Zone. Editors. BelleW. Baruch Library in Marine Science. No.20. University of South Carolina Press.Columbia, SC.

Vernberg, F.J., W.B. Vernberg, E. Blood, A.Fortner, M. Fulton, H. McKellar, W.Michener, G. Scott, T. Siewicki and K. El-Figi. 1992. “Impact of Urbanization onHigh-Salinity Estuaries in the SoutheasternUnited States.” Netherlands Journal of SeaResearch 30:239-248.

Walling, D. and J.Woodward. 1995. “TracingSources of Suspended Sediment in RiverBasins: A Case Study of the River Culm,Devon, UK.” Marine and FreshwaterResearch 46: 324-336.

Wahl, M., H. McKellar and T. Williams. 1997.“Patterns of Nutirent Loading in Forestedand Urbanized Coastal Streams.” Journalof Experimental Marine Biology andEcology 213:111-131.

Wang, L., J. Lyons, P. Kanehl and R.Bannerman. 2001. “Impacts of Urbaniza-tion on Stream Habitat and Fish AcrossMultiple Spatial Scales.” EnvironmentalManagement. 28(2):255-266.


References

Wang, L., J. Lyons, P. Kanehl and R. Gatti.1997. “Influences of Watershed Land Useon Habitat Quality and Biotic Integrity inWisconsin Streams.” Fisheries 22(6): 6-11.

Ward, J. and J. Stanford. 1979. The Ecology ofRegulated Streams. Plenum Press. NewYork, NY.

Waschbusch R., W. Selbig and R. Bannerman.2000. “Sources of Phosphorus in Stormwa-ter and Street Dirt from Two UrbanResidential Basins in Madison, Wisconsin,1994-1995.” In: National Conference onTools for Urban Water Resource Manage-ment and Protection. US EPA February2000: pp. 15-55.

Washington Department of Fish and Wildlife(WDFW). 1997. Final EnvironmentalImpact Statement for the Wild SalmonidPolicy. Olympia, Washington.

Water Environment Research Foundation(WERF). 1999. Chapter 4: Accumulationof Pollutants in Snowpack. Urban andHighway Snowmelt: Minimizing the Impacton Receiving Water. Alexandria, VA.

Water, B. 1999. Ambient Water QualityGuidelines for Organic Carbon. WaterMangement Branch Environment andResource Management. Ministry ofEnvironment, Lands and Parks.

Watters, G. 1996. Reasons for Mussel Declineand Threats to Continued Existence.Academy of Natural Sciences, Philadel-phia. Available at http://coa.acnatsci.org/conchnet/uniowhat.html

Weaver, L. 1991. Low-Intensity WatershedAlteration Effects on Fish AssemblageStructure and Function in a VirginiaPiedmont Stream. Masters Thesis. VirginiaCommonwealth University. VA.

Weaver, L. and G. Garman. 1994. “Urbaniza-tion of a Watershed and HistoricalChanges in Stream Fish Assemblage.”Transactions of the American FisheriesSociety 123: 162-172.

Weiskel, P., B. Howes and G.. Heufelder.1996. “Coliform Contamination of aCoastal Embayment: Sources and Trans-port Pathways.” Environmental Scienceand Technology 30:1871-1881.

Wernick, B.G., K.E. Cook, and H. Schreier.1998. “Land Use and Streamwater Nitrate-N Dynamics in an Urban-rural FringeWatershed.” Journal of the AmericanWater Resources Association 34(3): 639-650.

Williams, J., S. Fuller and R. Grace. 1992.“Effects of Impoundment on FreshwaterMussels (Mollusca: Bivalvia: Unionidae)in the Main Channel of the Black Warriorand Tombigbee Rivers in Western Ala-bama.” Bulletin of the Alabama Museum ofNatural History 13: 1-10.

Williams, J., M. Warren, Jr., K. Cummings,J.Harris and R. Neves. 1993. “Conserva-tion Status of Freshwater Mussels of theUnited States and Canada.” Fisheries18(9): 6-22.

Winer, R. 2000. National Pollutant RemovalPerformance Database for StormwaterTreatment Practices, 2nd Edition. Centerfor Watershed Protection. Ellicott City,MD.

Wotzka, P., J. Lee, P. Capel and M. Lin.1994.Pesticide Concentrations and Fluxes in anUrban Watershed. Proceedings AWRA1994 National Symposium on WaterQuality.

Wood, P. and P. Armitage. 1997. “BiologicalEffects of Fine Sediment in the LoticEnvironment.” Environmental Manage-ment 21(2):203-217.

Woodward-Clyde Consultants. 1992. SourceIdentification and Control Report. Pre-pared for the Santa Clara Valley NonpointSource Control Program. Oakland, Califor-nia.


References

Xiao, L., A. Singh, J. Limor, T. Graczyk, S.Gradus and A. Lal. 2001. “MolecularCharacterization of CryptosporidiumOocysts in Samples of Raw Surface Waterand Wastewater.” Applied and Environ-mental Microbiology 67(3):1097-1101.

Yoder, C. 1991. “The Integrated Biosurvey Asa Tool for Evaluation of Aquatic Life UseAttainment and Impairment in OhioSurface Waters.” In Biological Criteria:Research and Regulation, Proceedings of aSymposium, 12-13 December 1990,Arlington, VA, U.S. EPA, Office of Water,Washington, DC, EPA-440/5-91-005:110.

Yoder, C. and R. Miltner. 2000. “UsingBiological Criteria to Assess and ClassifyUrban Streams and Develop ImprovedLandscape Indicators.” In Proceedings ofthe National Conference on Tools forUrban Water Resource Management &Protection: Published by the US Environ-mental Protection Agency, Office ofResearch and Development, Washington,DC.

Yoder, C., R. Miltner and D.White. 1999.“Assessing the Status of Aquatic LifeDsignated Uses in urban and SuburbanWatersheds.” In Everson et al. Editors.National Conference on Retrofit Opportu-nities for Water Resource Protection inUrban Environments, Chicago, IL. EPA/625/R-99/002.

York, T. H. and W. J. Herb. 1978. ”Effects ofUrbanization and Streamflow on SedimentTransport in the Rock Creek and AnacostiaRiver Basins. Montgomery County, MD1972-1974.” USGS Professional Paper1003.

Young, K. and E. Thackston. 1999. “HousingDensity and Bacterial Loading in UrbanStreams.” Journal of EnvironmentalEngineering December:1177-1180.

Zapf-Gilje, R., S. Russell and D. Mavinic.1986. “Concentration of Impurities DuringMelting Snow Made From SecondarySewage Effluent.” Waterscience Technol-ogy 18:151-156.

Zielinski, J. 2001. Watershed VulnerabilityAnalysis. Prepared for Wake County (NC).Center for Watershed Protection. EllicottCity, MD.

Zielinski, J. 2000. “The Benefits of Better SiteDesign in Residential Subdivision andCommercial Development.” WatershedProtection Techniques 3(2): 633-656.


References


Glossary

Glossary

1st order stream: The smallest perennial stream. A stream that carries water throughout theyear and does not have permanently flowing tributaries.

2nd order stream: Stream formed by the confluence of two 1st order streams.

3rd order stream: Stream formed by the confluence of two 2nd order streams.

Acute toxicity: Designates exposure to a dangerous substance or chemical with sufficientdosage to precipitate a severe reaction, such as death.

Alluvial: Pertaining to processes or materials associated with transportation or deposition byrunning water.

Anadromous: Organisms that spawn in freshwater streams but live most of their lives in theocean.

Annual Pollutant Load: The total mass of a pollutant delivered to a receiving water body in ayear.

Bankfull: The condition where streamflow just fills a stream channel up to the top of the bankand at a point where the water begins to overflow onto a floodplain.

Baseflow: Stream discharge derived from ground water that supports flow in dry weather.

Bedload: Material that moves along the stream bottom surface, as opposed to suspendedparticles.

Benthic Community: Community of organisms living in or on bottom substrates in aquatichabitats, such as streams.

Biological Indicators: A living organism that denotes the presence of a specific environmen-tal condition.

Biological Oxygen Demand (BOD): An indirect measure of the concentration of biologi-cally degradable material present in organic wastes. It usually reflects the amount ofoxygen consumed in five days by bacterial processes breaking down organic waste.

Carcinogen: A cancer-causing substance or agent.

Catchment: The smallest watershed management unit. Defined as the area of a developmentsite to its first intersection with a stream, usually as a pipe or open channel outfall.

Chemical Oxygen Demand (COD): A chemical measure of the amount of organic sub-stances in water or wastewater. Non-biodegradable and slowly degrading compounds thatare not detected by BOD are included.

Chronic Toxicity: Showing effects only over a long period of time.

Combined Sewer Overflow (CSO): Excess flow (combined wastewater and stormwaterrunoff) discharged to a receiving water body from a combined sewer network when thecapacity of the sewer network and/or treatment plant is exceeded, typically during stormevents.


Glossary

Combined Indices (C-IBI or CSPS): Combined indices that use both fish and aquatic insectmetrics and a variety of specific habitat scores to classify streams.

Cryptosporidium parvum: A parasite often found in the intestines of livestock which con-taminates water when animal feces interacts with a water source.

Deicer: A compound, such as ethylene glycol, used to melt or prevent the formation of ice.

Dissolved Metals: The amount of trace metals dissolved in water.

Dissolved Phosphorus: The amount of phosphorus dissolved in water.

Diversity: A numerical expression of the evenness and distribution of organisms.

Ecoregion: A continuous geographic area over which the climate is uniform to permit thedevelopment of similar ecosystems on sites with similar geophysical properties.

Embeddedness: Packing of pebbles or cobbles with fine-grained silts and clays.

EPT Index: A count of the number of families of each of the three generally pollution-sensitiveorders: Ephemeroptera (mayflies), Plecoptera (stoneflies), and Trichoptera (caddisflies).

Escherichia coli (E. coli): A bacteria that inhabits the intestinal tract of humans and otherwarm-blooded animals. Although it poses no threat to human health, its presence indrinking water does indicate the presence of other, more dangerous bacteria.

Eutrophication: The process of over-enrichment of water bodies by nutrients, often typified bythe presence of algal blooms.

Fecal coliform: Applied to E. coli and similar bacteria that are found in the intestinal tract ofhumans and animals. Coliform bacteria are commonly used as indicators of the presenceof pathogenic organisms. Their presence in water indicates fecal pollution and potentialcontamination by pathogens.

Fecal streptococci: Bacteria found in the intestine of warm-blooded animals. Their presencein water is considered to verify fecal pollution.

Fish Blockages: Infrastructures associated with urbanization, such as bridges, dams, andculverts, that affect the ability of fish to move freely upstream and downstream inwatersheds. Can prevent re-colonization of resident fish and block the migration ofanadromous fish.

Flashiness: Percent of flows exceeding the mean flow for the year. A flashy hydrograph wouldhave larger, shorter-duration hydrograph peaks.

Geomorphic: The general characteristic of a land surface and the changes that take place in theevolution of land forms.

Giardia lamblia: A flagellate protozoan that causes severe gastrointestinal illness when itcontaminates drinking water.

Herbicide: Chemicals developed to control or eradicate plants.

Hotspot: Area where land use or activities generate highly contaminated runoff, with concentra-tions of pollutants in excess of those typically found in stormwater.

Hydrograph: A graph showing variation in stage (depth) or discharge of a stream of water overa period of time.

Illicit discharge: Any discharge to a municipal separate storm sewer system that is not com-posed entirely of storm water, except for discharges allowed under an NPDES permit.


Glossary

Impervious Cover: Any surface in the urban landscape that cannot effectively absorb orinfiltrate rainfall.

Impervious Cover Model (ICM): A general watershed planning model that uses percentwatershed impervious cover to predict various stream quality indicators. It predictsexpected stream quality declines when watershed IC exceeds 10% and severe degrada-tion beyond 25% IC.

Incision: Stream down-cuts and the channel expands in the vertical direction.

Index of Biological Integrity (IBI): Tool for assessing the effects of runoff on the quality ofthe aquatic ecosystem by comparing the condition of multiple groups of organisms ortaxa against the levels expected in a healthy stream.

Infiltration: The downward movement of water from the surface to the subsoil. The infiltrationcapacity is expressed in terms of inches per hour.

Insecticide: Chemicals developed to control or eradicate insects.

Large Woody Debris (LWD): Fundamental to stream habitat structure. Can form dams andpools; trap sediment and detritus; provide stabilization to stream channels; dissipate flowenergy and promote habitat complexity.

Mannings N: A commonly used roughness coefficient; actor in velocity and discharge formulasrepresenting the effect of channel roughness on energy losses in flowing water.

Methyl Tertiary-Butyl Ether: An oxygenate and gasoline additive used to improve the effi-ciency of combustion engines in order to enhance air quality and meet air pollutionstandards. MTBE has been found to mix and move more easily in water than many otherfuel components, thereby making it harder to control, particularly once it has enteredsurface or ground waters.

Microbe: Short for microorganism. Small organisms that can be seen only with the aid of amicroscope. Most frequently used to refer to bacteria. Microbes are important in thedegradation and decomposition of organic materials.

Nitrate: A chemical compound having the formula NO

3. Excess nitrate in surface waters can

lead to excessive growth of aquatic plants.

Organic Matter: Plant and animal residues, or substances made by living organisms. All arebased upon carbon compounds.

Organic Nitrogen: Nitrogen that is bound to carbon-containing compounds. This form ofnitrogen must be subjected to mineralization or decomposition before it can be used bythe plant community.

Overbank Flow: Water flow over the top of the bankfull channel and onto the floodplain.

Oxygenate: To treat, combine, or infuse with oxygen.

Peak Discharge: The maximum instantaneous rate of flow during a storm, usually in referenceto a specific design storm event.

Pesticides: Any chemical agent used to control specific organisms, for example, insecticides,herbicides, fungicides and rodenticides.

Piedmont: Any plain, zone or feature located at the foot of a mountain. In the United States, thePiedmont (region) is a plateau extending from New Jersey to Alabama and lying east ofthe Appalachian Mountains.


Glossary

Pool: A stream feature where there is a region of deeper, slow-moving water with fine bottommaterials. Pools are the slowest and least turbulent of the riffle/run/pool category.

Protozoan: Any of a group of single-celled organisms.

Rapid Bioassessment Protocols (RBP): An integrated assessment, comparing habitat, waterquality and biological measures with empirically defined reference conditions.

Receiving Waters: Rivers, lakes, oceans, or other bodies of water that receive water fromanother source.

Riffle: Shallow rocky banks in streams where water flows over and around rocks disturbing thewater surface; often associated with whitewater. Riffles often support diverse biologicalcommunities due to their habitat niches and increased oxygen levels created by the waterdisturbance. Riffles are the most swift and turbulent in the riffle/run/pool category.

Roughness: A measurement of the resistance that streambed materials, vegetation, and otherphysical components contribute to the flow of water in the stream channel and flood-plain. It is commonly measured as the Manning’s roughness coefficient (Manning’s N).

Run: Stream feature characterized by water flow that is moderately swift flow, yet not particu-larly turbulent. Runs are considered intermediate in the riffle/run/pool category.

Runoff Coefficient: A value derived from a site impervious cover value that is applied to agiven rainfall volume to yield a corresponding runoff volume.

Salmonid: Belonging to the family Salmonidae, which includes trout and salmon.

Sanitary Sewer Overflow (SSO): Excess flow of wastewater (sewage) discharged to areceiving water body when the capacity of the sewer network and/or treatment plant isexceeded, typically during storm events.

Semi-arid: Characterized by a small amount of annual precipitation, generally between 10 and20 inches.

Simple Method: Technique used to estimate pollutant loads based on the amount of IC foundin a catchment or subwatershed.

Sinuosity: A measure of channel curvature, usually quantified as the ratio of the length of thechannel to the length of a straight line along the valley axis. It is, in essence, a ratio of thestream’s actual running length to its down-gradient length.

Soluble Phosphorus: The amount of phosphorus available for uptake by plants and animals.

Stormwater: The water produced as a result of a storm.

Subwatershed: A smaller geographic section of a larger watershed unit with a drainage area ofbetween two to 15 square miles and whose boundaries include all the land area drainingto a point where two 2nd order streams combine to form a 3rd order stream.

Total Dissolved Solids (TDS): A measure of the amount of material dissolved in water (mostlyinorganic salts).

Total Kjeldhal Nitrogen (TKN): The total concentration of nitrogen in a sample present asammonia or bound in organic compounds.

Total Recoverable Metals: The amount of a metal that is in solution after a representativesuspended sediment sample has been digested by a method (usually using a dilute acidsolution) that results in dissolution of only readily soluble substances).


Glossary

Total Maximum Daily Load (TMDL): The maximum quantity of a particular water pollutantthat can be discharged into a body of water without violating a water quality standard.

Total Nitrogen (Total N): A measure of the total amount of nitrate, nitrite and ammoniaconcentrations in a body of water.

Total Organic Carbon (TOC): A measure of the amount of organic material suspended ordissolved in water.

Total Phosphorous (Total P): A measure of the concentration of phosphorus contained in abody of water.

Total Suspended Solids (TSS): The total amount of particulate matter suspended in the watercolumn.

Trophic Level: The position of an organism in a food chain or food pyramid.

Turbidity: A measure of the reduced transparency of water due to suspended material whichcarries water quality and aesthetic implications. Applied to waters containing suspendedmatter that interferes with the passage of light through the water or in which visual depthis restricted.

Volatile Organic Compounds (VOC): Chemical compounds which are easily transportedinto air and water. Most are industrial chemicals and solvents. Due to their low watersolubility they are commonly found in soil and water.


Glossary

Documents

Watershed Protection Research Monograph No. 1 IMPACTS