APPROVAL SHEET

Title of Thesis: The Effectiveness of Street Sweeping and Bioretention in Reducing

Pollutants in Stormwater Name of Candidate: Catherine J. DiBlasi Master of Science in Civil Engineering, 2008 Thesis and Abstract Approved: ________________________________________ Dr. Upal Ghosh Associate Professor Department of Civil and Environmental Engineering Date Approved: ___________

CURRICULUM VITAE

Name: Catherine J. DiBlasi

Permanent Address: 2301 Birch Drive, Baltimore, Maryland 21207

Degree and Date to be Conferred: M.S., 2008

Date of Birth: May 19, 1980.

Place of Birth: Annapolis, Maryland.

Secondary Education:

Severna Park Senior High School, Severna Park, Maryland, May 1998

Collegiate Institutions Attended:

Columbia University, New York, New York. September 1998 to May 2002, B.S. May 2002.

Major: Environmental Biology. University of Maryland, Baltimore County, Baltimore, Maryland. September 2005 to April 2008, M.S. May 2008. Major: Civil Engineering – Focus in Environmental Engineering. Professional Publications:

DiBlasi, C., Law, N., Ghosh, U. (2007). Implementing a Sustainable Stormwater Management Program in an Urban Center – Baltimore, MD. Proceedings of the 2nd International Conference on Sustainability Engineering and Science, Auckland, New Zealand, Feb 21-23, 2007.

Professional Positions Held:

Maryland Department of Natural Resources Natural Resources Bioligist

1919 Lincoln Drive, Annapolis, Maryland, 21401. May 2002 to August 2005. Army Corps of Engineers Civil Engineer 10 South Howard Street, Baltimore, Maryland, 21201. April 2008 to Present.

ABSTRACT

Title of Thesis: The Effectiveness of Street Sweeping and Bioretention in Reducing Pollutants in Stormwater

Catherine J. DiBlasi, Master of Science, 2008

Thesis Directed by: Upal Ghosh, Associate Professor, Department of Civil and Environmental Engineering

Research has shown that a great majority of pollutants in urban stormwater are

strongly associated with particulate matter. Therefore, the effectiveness of a stormwater

best management practice (BMP) is largely dependent on its ability to reduce suspended

solids in stormwater. Both street sweeping and bioretention have the potential to

decrease stormwater suspended solid loads. Field investigations were performed in this

research to evaluate the effectiveness of these two BMPs. A paired-catchment study in

an urban watershed of Baltimore, Maryland was performed to physically and chemically

characterize street particulate matter (

THE EFFECTIVENESS OF STREET SWEEPING AND BIORETENTION IN REDUCING POLLUTANTS IN STORMWATER

By

Catherine J. DiBlasi

Thesis submitted to the Faculty of the Graduate School

of the University of Maryland Baltimore County in partial fulfillment of the requirements

for the degree of Master of Science in Civil Engineering

2008

1454791

1454791 2008

Copyright 2008 byDiBlasi, Catherine J. All rights reserved

© Copyright by

Catherine J. DiBlasi 2008

ii

ACKNOWLEDGEMENTS

I would like to sincerely thank the following people for their support: Dr. Upal Ghosh for his support and guidance, for always providing an example of excellence, and inspiring me to work harder. Dr. Neely Law for her enthusiasm and assistance throughout the many stages of this research project. Dr. Brian Reed, Dr. Claire Welty and Dr. Nagaraj Neerchal for their advice as members of my thesis committee and during my time here as a student. U.S. EPA Chesapeake Bay Program for funding this project. Dr. Allen Davis and Houng Li from the University of Maryland, College Park. Yan Zhuang for her assistance with all data organization and statistical analysis. Baltimore City DPW staff, especially Bill Stack, Matt Cherigo, Robert McAulay and Norm Seldon. U.S. Forest Service scientists, especially Rich Pouyat, Ken Belt and Ian Yesilonis. Baltimore County DEPRM and DPW staff, especially Steve Stewart, Megan Brosh and John Burnett. Fellow CEE students for their friendship and advice in the classroom, laboratory, office, and lunchroom. My family, especially my Mom and Dad, Rich, Beth, Glenn, Chris and Isaac for their encouragement, and for always listening. My fiancé, Russell, for his invaluable support, patience, and confidence in me.

iii

TABLE OF CONTENTS

ABSTRACT ACKNOWLEDGEMENTS ii TABLE OF CONTENTS iii LIST OF TABLES vi LIST OF FIGURES vii LIST OF ABBREVIATIONS xi 1. INTRODUCTION 1.1 Stormwater Runoff 1 1.2 Chesapeake Bay 4 1.2.1 A Holistic Approach to Stormwater Management 6 1.2.2 Baltimore, MD – Watershed 263 8 1.3 Evaluating the Effectiveness of Stormwater BMPs 10 1.4 Organization of Thesis and My Contribution 11 1.5 References 13 2. BACKGROUND 2.1 Street Particulate Matter 15 2.1.1 Sources of Street Particulate Matter, Associated Pollutants 16 2.1.2 Characterization of Street Particulate Matter 19 2.2 Street Sweeping as a Stormwater BMP 2.2.1 Literature Review of Street Sweeping as a BMP 24 2.2.2 Conceptual Model 27 2.3 Bioretention as a Stormwater BMP 2.3.1 Bioretention Concept 30 2.3.2 Literature Review of Bioretention as a BMP 33 2.3.3 Polycyclic Aromatic Hydrocarbons (PAHs) 36 2.4 References 39

iv

3. STUDY OBJECTIVES

3.1 Street Sweeping as a BMP for Solids, Nutrients and Metals 45 3.2 Bioretention as a Stormwater BMP for PAHs 46 4. EXPERIMENTAL METHODS 4.1 Street Sweeping Study 4.1.1 Description of Study Site – Watershed 263 47

4.1.2 Paired-Catchment Study Setup 48 4.1.3 Street Particulate Matter Sampling 53

4.1.3.1 Particle Type Analysis 57 4.1.3.2 Particle Size Analysis 57

4.1.3.3 Chemical Analysis 58 4.1.4 Stormwater Sampling 59 4.1.4.1 Water Quality Analysis 60 4.1.4.2 First Flush Inlet Samplers 62 4.1.4.3 Statistical Analysis of Water Quality Data 62 4.1.5 Precipitation Data 66 4.1.6 Flow Data and Runoff Coefficients 67 4.1.7 Annual Pollutant Load Calculations 68 4.2 Bioretention Cell Study 4.2.1 Description of Bioretention Cell Study Site 69 4.2.2 Stormwater Sampling 70 4.2.3 PAH Analysis 72 4.2.4 Bioretention Media Core Collection and Analysis 73 4.2.5 PAH Source Investigation 75 4.3 References 75 5. RESULTS AND DISCUSSION OF STREET SWEEPING STUDY

5.1 Street Particulate Matter 77

5.2 Water Quality Data 95 5.3 References 132

6. BIORETENTION STUDY PAPER 137 7. CONCLUSIONS AND RECOMMENDATIONS

7.1 Conclusions 166

v

7.2 Recommendations 169

8. APPENDICES 8.1 APPENDIX A 172

vi

LIST OF TABLES Table No. Title Page No. 1.1 Typical pollutant concentrations in stormwater 1 1.2 Nutrient sources to the Chesapeake Bay 5 2.1 US EPA’s 16 priority pollutant PAHs 37 4.1 Characteristics of catchments O and F 50 4.2 Street sweeping treatment periods in paired catchment study 51 4.3 Analytical parameters and methods for street particulate matter 59 4.4 Water quality parameters for stormwater analysis 61 4.5 The effect of sample dilution on laboratory detection limits 61 4.6 Distances from water quality sampling stations to rain gages 67 5.1 Collection date and type of street particulate matter samples 77 5.2 One-way ANOVA results for street particulate matter loadings 79 5.3 Gross pollutant sample weights by sample type 80 5.4 Street particulate matter particle size distribution 85 5.5 Chemical composition of street particulate matter 87 5.6 T-test results for particulate matter chemical composition 88 5.7 Pollutant load contribution of fine street particulate matter 92 5.8 Number of baseflow and stormwater samples collected 96 5.9 T-test results of baseflow vs. stormwater comparison 100 5.10 Percent of censored values in each water quality dataset 105 5.11 Effect of substitution methods for censored data on statistics 107 5.12 Median values of water quality parameters 110 5.13 T-test results of water quality in baseline and treatment periods 113 5.14 Runoff coefficients estimated from stormwater flow data 117 5.15 Stormwater concentrations in part 1 and part 2 of storms 119 5.16 Median concentrations in first flush samples 120 5.17 Comparing stormwater and first flush pollutant concentrations 121 6.1 Total PAH removal efficiency, TSS, rainfall and runoff depth 147 6.2 Log Kd values for phenanthrene and pyrene 155 S1 EPA’s 16 priority pollutant PAHs 164 A-1 Elgin whirlwind 4 MV specifications 172 A-2 Street particulate matter particle size analysis data 173 A-3 Street particulate matter chemical analysis data 174 A-4 Baltimore Street (Catchment O) water quality data – original 176 A-5 Lanvale Street (Catchment O) water quality data – original 182 A-6 Results of Shapiro-Wilk normality test 188 A-7 Example of ROS method and a substituted dataset 191

vii

LIST OF FIGURES Figure No. Title Page No.

2.1 Particle size distribution of street particulate matter 21 2.2 Conceptual model - street particulate matter sources and fate 29 2.3 Typical cross section of a bioretention cell 32 2.4 Image of a curbed parking lot median bioretention cell 33 4.1 Map of Baltimore, Maryland, and Watershed 263 48 4.2 Locations of catchments O and F within Watershed 263 49 4.3 Lanvale St. water quality monitoring station in Watershed 263 50 4.4 Street sweeper used in study – Elgin Whirlwind MV 52 4.5 Vacuum configuration for street particulate matter sampling 53 4.6 Locations of street particulate matter sampling streets 54 4.7 Diagram and photo of street particulate sampling methods 56 4.8 Picture of a first flush sampler installed in Watershed 263 62 4.9 Relationship between runoff coefficient and impervious area 68 4.10 Image of bioretention cell monitored in College Park, Maryland 70 4.11 Layout of monitored bioretention cell and sampling locations 74 5.1 Average street particulate matter loadings 78 5.2 Street particulate matter loadings throughout sampling period 79 5.3 Images of street particulate matter type fractions 83 5.4 Particle size distribution of street particulate matter samples 84 5.5 Comparing particle size distribution with literature values 86 5.6 Average chemical concentrations in street particulate matter 87 5.7 Chemical composition of particulate matter vs. literature values 90 5.8 Percent contribution of particle size fractions to pollutant loads 91 5.9 Relationship – particulate matter load and time since sweeping 94 5.10 Relationship – particulate load and time since sweeping or rain 95 5.11 Total metals in baseflow and stormwater at Baltimore St. 98 5.12 Total metals in baseflow and stormwater at Lanvale St. 99 5.13 Dissolved Zinc at Baltimore St. in baseflow and stormwater 102 5.14 E. coli at Baltimore St. in baseflow and stormwater 103 5.15 NO2-NO3 at Baltimore St. in baseflow and stormwater 104 5.16 Histograms - Effect of substitution method on censored data 108 5.17 Particulate metal concentrations during parts 1 and 2 of storms 119 5.18 First flush and stormwater concentrations of metals and solids 122 5.19 Lead concentrations in particulate matter and suspended solids 124 5.20 Estimated total phosphorus annual loads from Watershed 263 127 5.21 Estimated total nitrogen annual loads from Watershed 263 128 6.1 Image of monitored bioretention cell and sampling locations 145 6.2 Total PAH concentrations in stormwater influent and effluent 149 6.3 16 PAH concentrations in dissolved phase of stormwater 151

viii

6.4 16 PAH concentrations in particulate phase of stormwater 152 6.5 Comparison of PAH and TSS removal for five storms 153 6.6 Total PAH media profile in material from bioretention cell 158 S1 16 PAH concentrations in media profile 165 S2 16 PAH distribution profile of media and suspended solids 165

ix

LIST OF ABBREVIATIONS A – Accumulation Sample BaP – Benzo(a)pyrene BES – Baltimore Ecosystem Study BMP – Best Management Practice BOD-5 – Biological Oxygen Demand – 5 day BWI – Baltimore Washington International Airport NOAA weather station C – Control Sample CFS – Cubic Feet per Second CP – College Park CWA – Clean Water Act CWP – Center for Watershed Protection Cu – Copper (DisCu = Dissolved Copper, TotCu = Total Copper) DEPRM – Department of Environmental Resources and Management DHM – Downtown Baltimore Maryland NOAA weather station DPW – Department of Public Works E – Extractable E. coli - Escherichia coli EMC – Event Mean Concentration EPA – Environmental Protection Agency FecCol – Fecal Coliforms IA – Impervious Area Kd – partition coefficient between aqueous and solid phase LID – Low Impact Development MDE – Maryland Department of the Environment MPN – Most Probable Number MS4 – Municipal Separate Storm Sewer Systems NO2 – Nitrite NO3 – Nitrate NPDES – National Pollutant Discharge Elimination System PAH – Polycyclic Aromatic Hydrocarbons Pb – Lead (DisPb = Dissolved lead, TotPb = Total lead) PO4 - Phosphate ROS – Regression on Ordered Statistics Rv – Runoff coefficient, ratio of event runoff to event rainfall S – After Street Sweeping Sample SS – Suspended Solids, same as TSS (Total Suspended Solids) SO4 - Sulfate TKN – Total Kjeldahl Nitrogen TMDL – Total Maximum Daily Load TP – Total Phosphorus UMBC – University of Maryland, Baltimore County UMCP – University of Maryland College Park Zn – Zinc (DisZn = Dissolved Zn, TotZn = Total Zn)

- 1 -

1. INTRODUCTION

1.1 Stormwater Runoff

In the United States, the Environmental Protection Agency (EPA) estimates that

stormwater runoff is responsible for 21% of impaired lakes and 45% of impaired

estuaries (U.S. EPA 2007b). Stormwater runoff is defined as any rainwater that does not

evaporate or infiltrate into the ground (Chesapeake Bay Program 2006). As stormwater

flows over land it can pick up metals, nutrients, bacteria, pesticides and other

contaminants and transport them to receiving waters. Typical median pollutant

concentrations in urban stormwater runoff for residential, mixed, and commercial land

uses, compared to values in untreated municipal wastewater are shown in Table 1.1.

Table 1.1: Typical median event mean concentrations of pollutants in urban stormwater runoff for residential, mixed, and commercial land uses, compared to typical concentrations in untreated municipal wastewater.

Pollutant (units) Residentiala Mixeda Commerciala Municipal Wastewaterb

Biological Oxygen Demand (mg/L)

10 7.8 9.3 120-370

Chemical Oxygen Demand (mg/L)

73 65 57 260-900

Total Suspended Solids (mg/L)

101 67 69 120-370

Fecal coliform (MPN/100 ml)

103-104 (b)

105-107

Total Lead (�g/L) 144 114 104 Total Copper (�g/L) 33 27 29 0.1-10 (c,d) Total Zinc (�g/L) 135 154 226 Total Kjeldahl Nitrogen (mg/L)

1.9

1.3

1.2

20-45

Nitrate + Nitrite (mg/L) 0.73 0.56 0.57 0 Total Phosphorus (mg/L) 0.38 0.26 0.20 4-16 Sources: a Davis (2005), bMetcalf and Eddy (2003), cU.S. EPA (1981), dU.S. EPA (2003).

- 2 -

While biological oxygen demand and pathogens are more than an order of

magnitude lower in stormwater compared to untreated municipal wastewater, other

constituents like suspended solid and heavy metals are present at comparable or greater

concentrations in stormwater (Davis 2005, Metcalf and Eddy 2003, U.S. EPA 2003). In

addition to its pollutant load, the sheer volume of stormwater runoff can also cause

negative impacts, such as flooding and erosion. As stormwater flows over land, it may

erode soil and then redeposit the soil in receiving waters, decreasing water clarity and

degrading aquatic habitats. To reduce the negative impacts of stormwater runoff, it is

important to manage both stormwater quantity and quality.

Stormwater runoff occurs over a diffuse area and is typically considered a

nonpoint pollution source. Nonpoint source pollution is challenging to both assess and

regulate. Unlike many point sources, nonpoint sources are not constant, and inputs do

not follow a regular pattern. Additionally, the characteristics and pollution loads of

watersheds where nonpoint source pollution originates are constantly changing over time

(Burton and Pitt 2002). Urban nonpoint source pollution is usually covered by nonpoint

source pollution management programs developed by individual states, particularly

common in coastal environments, but the majority of stormwater management

regulations are focused on point sources (U.S. EPA 1993).

Much of the runoff that originates as a nonpoint source may eventually enter a

storm drain network, be channelized and become a point source. This is particularly

common in urban environments, where runoff which enters and is discharged through

conveyances such as municipal storm drain pipes are treated as point sources and subject

to permit requirements of the Clean Water Act (U.S. EPA 1993). Specifically, these non-

- 3 -

agricultural sources of stormwater discharges are regulated by the National Pollutant

Discharge Elimination System (NPDES) permit program, under section 402(p) of the

Clean Water Act.

The NPDES permit program is a cornerstone of the 1972 Federal Clean Water

Act (CWA), which requires that all pollutant discharges entering Unites States waters

from a point source must be authorized by a NPDES permit. A 1987 amendment to the

CWA greatly expanded the scope of the NPDES permitting program to include

stormwater discharges. Stormwater systems are currently one of the major categories of

point sources regulated under the NPDES program, and authorization is typically given

by state environmental agencies. There are two main categories in which NPDES

stormwater discharge permits are given, Phase I and Phase II. Phase I of the NPDES

program covers municipal separate storm sewer systems (MS4s) serving greater than

100,000 people and eleven categories of industrial activity including construction sites on

greater than five acres of land. Phase II of the program covers MS4s serving less than

100,000 people and construction activity affecting between 1 and 5 acres of land (U.S.

EPA 2006a). For regulated municipalities, the most important element of the NPDES

program is the requirement to develop and implement a stormwater management program.

Stormwater management programs typically include measures to: identify major

outfalls and pollutions sources, detect and eliminate non-stormwater discharges to the

system, reduce pollutants in runoff from all existing land uses and control runoff from

new development or redevelopment areas (U.S. EPA 2007c). A critical element of

stormwater management programs is the use of practices (called Best Management

Practices), which reduce or prevent the discharge of pollutants into receiving waters.

- 4 -

Best Management Practices (BMPs) are essentially any structural or non-structural

practice that reduces the quantity and/or improves the quality of stormwater in a cost-

effective manner. A closer look at a specific stormwater management program in section

1.2 of this paper will provide better understanding of urban stormwater management in

the context of Baltimore City and the Chesapeake Bay. Examining urban runoff in the

Chesapeake Bay watershed can provide insight to issues facing much of the nation. As

population and development trends indicate that more than half of the United States will

live in coastal towns and cities by the year 2010, urban stormwater runoff will continue to

be an increasingly important issue (U.S. EPA 2006b).

1.2 Chesapeake Bay

Estuaries have immense commercial and recreational value, providing habitat for

more than 75% of the United State’s commercial fish catch and between 80 to 90 percent

of recreational fish catch (U.S. EPA 2007a). However, more than 60 percent of coastal

rivers and bays are categorized as moderately to severely degraded by nutrient pollution,

and this problem is particularly acute in the mid-Atlantic states (Clement et al. 2001). In

the Chesapeake Bay, eutrophication resulting from excess nutrient loading is the main

cause of poor water quality and aquatic habitat loss, and reducing nutrient inputs to the

Bay is a critical element of restoration efforts. The three major contributors of nutrient

pollution to the Chesapeake Bay include effluent from wastewater treatment plants,

agricultural runoff, and urban stormwater (Chesapeake Bay Program 2006). Major

sources of nutrients to the Chesapeake Bay are included in Table 1.2, along with potential

management options to reduce these contributions.

- 5 -

Table 1.2: Common sources of nutrients to the Chesapeake Bay and potential management options to reduce their contributions. Nutrient Source Type Management Options Wastewater Treatment Plant Effluent

Point - Limit concentration in effluent - Recycle water - Public education to reduce water consumption

Agricultural Runoff

Nonpoint - Riparian buffers - Limit manure/fertilizer application - Regulate animal waste disposal practices

Urban Stormwater Runoff

Nonpoint - Reduce impervious cover - Practice low impact development - Source control - Best Management Practices (BMPs)

Atmospheric Deposition Nonpoint - Restrict discharges from fossil fuel burning facilities

Groundwater Discharge Nonpoint - Septic tank restrictions

To date, the majority of action has been focused on reducing inputs from

wastewater treatment plants by upgrading plants with technology to reduce nitrogen and

phosphorus concentrations in their effluent. However, it is estimated that point sources

contribute only about 20% of nitrogen delivered to Chesapeake Bay, while nonpoint

sources contribute the remaining 80% (U.S. Environmental Protection Agency 2002).

Clearly, to improve the health of the Bay, nonpoint sources like agricultural and urban

runoff must be aggressively addressed. Although agricultural runoff is considered the

single greatest source of nutrients to the Bay, contributing about 40% of nitrogen and

50% of phosphorous loads, it is particularly challenging to regulate (Chesapeake Bay

Foundation 2003). Therefore, many municipalities in the Chesapeake Bay watershed are

primarily focused on urban stormwater management.

Urbanized areas have higher impervious cover, such as streets, rooftops,

sidewalks and parking lots. Pollutants settle and accumulate on the impervious areas

until wash off by rain into the storm drain system and eventually into receiving waters.

- 6 -

Common pollutants which build up on impervious surfaces in urban environments

include: sediment, litter, pesticides, fertilizers, oils, road salt, and other debris.

The Chesapeake Bay Program (2006) estimates that urban runoff is responsible

for 16% of the phosphorus, 11% of the nitrogen, and 9% of the sediment loads entering

the Chesapeake Bay. To reduce the negative effects urban landscapes have on water

quality, the Chesapeake Bay Program (2006) recommends reducing impervious cover

and its impact through low impact development practices, source reduction, and best

management practices (BMPs). Examples of traditional BMPs include detention basins,

grass swales and vegetated buffers. However, in an urban environment, implementation

of stormwater BMPs is often limited by space and budget constraints, and staffing

shortages. Nevertheless, there are alternative stormwater management options available

which urban municipalities can implement to both improve water quality and positively

impact public health and attitude. To achieve these multiple objectives, municipalities

must adopt a holistic approach and develop a sustainable stormwater management

program.

1.2.1 A Holistic Approach to Stormwater Management

A holistic approach to stormwater management implies focusing not only on the

immediate problem of polluted stormwater runoff but on all the potential sources and

their underlying causes in a specific watershed. For instance, municipalities should work

beyond simply implementing a few individual BMPs to fulfill permit requirements, and

instead select practices with an understanding of their interactions and how they will

perform in synergy to impact the watershed. History has shown that the traditional

- 7 -

method of rapidly conveying stormwater to receiving waters leads to environmental

degradation and sustainable or holistic stormwater management is largely a response to

this traditional method. Sustainable management of urban stormwater reflects values of

water conservation, pollution prevention, and ecological restoration (Brown 2005) while

continuing to utilize BMPs.

In addition to realizing the interactions among practices, municipal operators

should acknowledge the interconnectedness of their local environment and the local

community. For example, an urban stormwater management program should address the

fact that nearly all street litter is intentionally left there by humans. Therefore, a

cornerstone of any effective urban stormwater management program should be public

outreach and education, along with public provisions that encourage alternate behavior

(i.e. using trash receptacles). Overall, gaining community support and adopting a holistic

approach to stormwater management will make it possible for municipalities to design

and implement a sustainable, long term program. A sustainable stormwater management

program incorporates environmental, social, and economic concerns in the decision

making process. Additionally, a sustainable stormwater management program should

control and reduce the impact of current runoff and plan for future challenges with

respect to population growth and landscape changes. In Baltimore, Maryland, a unique

and collaborative project is currently underway to develop a sustainable stormwater

management program which aims to ultimately restore an impoverished watershed.

- 8 -

1.2.2 Baltimore, Maryland – Watershed 263

The project site is a 376 hectare (930 acre) area called Watershed 263 (outfall

number) located in southwest Baltimore City. Several unique characteristics of the

research site make it a challenging and exciting study location. First, impervious cover in

Watershed 263 averages close to 75%, which is significantly higher than the 40% city-

wide average in Baltimore (Richardson 2006). Also, Watershed 263 is absent of any

flowing surface waters - all area streams were piped and buried about 100 years ago,

creating 69 km (43 miles) of pipes which serve as the main components of the storm

drain system (Center for Watershed Protection 2006). Within the storm drain system,

there is a substantial dry-weather baseflow, which is likely due to sewage, drinking water,

and groundwater entering the system through leaky pipes (Richardson 2006). Nearly all

of the neighborhoods within Watershed 263 have suffered moderate to severe economic

decline due to suburbanization and the loss of industrial development. Economic decline

in Watershed 263 has led to a large concentration of vacant houses and lots, a high

unemployment rate and a significant portion of the population living below the poverty

level (Center for Watershed protection 2006).

Watershed 263 has received significant research attention in the past several years

from a number of partners including the federal government (U.S. Forest Service), the

state (University of Maryland, Baltimore County), Baltimore City (Department of Public

Works), and non-profits (Center for Watershed Protection, Parks & People Foundation).

However, what sets this urban environmental research project apart from many others is a

commitment to community involvement and participation. The key to community

involvement is a partnership between the Baltimore City Department of Public Works

- 9 -

and the Parks & People foundation, an organization dedicated to improving quality of life

for Baltimore residents. The Parks & People foundation worked to raise public

awareness in the watershed by holding over 40 community meetings to explain the

restoration plan, along with separate community forums to hear resident feedback. Since

the initial meetings, an advisory group of about 20 concerned residents has been formed

to represent each geographic neighborhood in planning decisions (Richardson 2006).

There are a number of innovative stormwater management projects being

implemented in Watershed 263, and two of the most notable are the schoolyard greening

initiative and the Clean & Green program. The schoolyard greening project has removed

more than 1.5 hectares of unused asphalt from several public schools and replaced it with

green lawns and gardens. This program uses critical area and stormwater management

credits to reduce impervious area in the watershed and improve aesthetics, while also

offering environmental education opportunities for elementary and middle school

students. These schools involve the students in the redesign of their schoolyards and

incorporate the planting of trees and gardens into the school curriculum. The Clean &

Green program is a partnership between two community outreach organizations in

Baltimore City developed to improve the hundreds of vacant lots across southwest

Baltimore. To date, this program has converted over 330 vacant lots (about 3.3 hectares)

to green space by planting grass and more than 500 trees (Center for Watershed

Protection 2006).

In addition to these high profile beautification projects, Watershed 263 is also the

site of an intensive field study designed to determine the effectiveness of stormwater

management practices. Developed as a paired-catchment study, this research is

- 10 -

examining the effects of various stormwater practices on the water quality of two urban

catchments within Watershed 263. Details about the paired-catchment study, which

constitutes a large portion of this thesis research will be discussed further in section 4.1 of

this paper. The diverse work taking place in Watershed 263 provides examples of

challenges that many urban municipalities are facing and illustrates the breadth and

complexity of urban stormwater management. While sustainable stormwater

management programs vary by site, all municipalities requires the same essential

scientific information - data on the effectiveness of stormwater BMPs.

1.3 Evaluating the Effectiveness of Stormwater BMPs

When municipal operators design and implement their stormwater management

programs, it is vital that they select the most appropriate BMPs for the sites within their

community. Ideally, when a study on BMP effectiveness is performed, it will provide

useful information for that individual site as well as for other similar and different types

of BMPs at other locations. However, due to variation in study methods and lack of

information about specific design and reporting protocols, the majority of BMP studies

result in data which is difficult to use in comparing effectiveness and selecting individual

BMP design types (Strecker et al. 2001). Strecker and others (2001) reported on a U.S.

EPA funded cooperative research program with the ASCE to develop a more useful set of

data on BMP effectiveness in reducing pollutants in stormwater. One of the major

recommendations of this study was that effluent quality is a much more robust measure

of BMP efficiency rather than “percent removal” which is typically reported.

- 11 -

Monitoring the input and output of an individual BMP is the typical approach

used, but control watersheds and before/after studies on a watershed are also performed.

Larger watershed-scale studies, such as Watershed 263, are valuable as long as all other

potential factors that may be contributing to changes are identified and accounted for.

Overall, large field studies can provide excellent information on whether the

implementation of BMPs cause a significant difference across the watershed (Strecker et

al. 2001).

This thesis focuses on the effectiveness of two BMPs in reducing pollutants in

stormwater. The BMPs selected include one of the oldest and most commonly used

practices, street sweeping, and one more recently developed practice, bioretention. Both

of these practices have the potential to be of particular value to urban stormwater

management programs. The impacts of street sweeping and bioretention on stormwater

quality were investigated in two separate, intensive field studies. Before discussing the

details and objectives of these studies, it is important to provide background information

on these practices and their use as stormwater BMPs. Street sweeping and bioretention

are discussed in detail in sections 2.1 and 2.2, respectively, including an overview of the

concepts, a history of their uses, and a literature review of previous research on their

effectiveness.

1.4 Organization of Thesis and My Contribution

This thesis is organized into seven chapters. The chapters, described in more

detail below include: 1) Introduction, 2) Background, 3) Study Objectives, 4) Materials

- 12 -

and Methods, 5) Street Sweeping Study Results and Discussion, 6) Bioretention Study

Paper, and 7) Conclusions and Recommendations.

Chapter 1 includes an introduction to stormwater runoff and stormwater

management, followed by a discussion of how stormwater impacts the Chesapeake Bay.

The need for a holistic approach to stormwater management is discussed, using Baltimore

City Watershed 263 as an example. Chapter 2 provides background or literature review

information on three general topics: street particulate matter as a pollution source

(physical and chemical characterization), street sweeping as a stormwater BMP, and

bioretention as a BMP with an overview of PAHs included. Chapter 3 summarizes the

study objectives of this research in two sections entitled: street sweeping as a stormwater

BMP for suspended solids, nutrients, and metals, and bioretention as a stormwater BMP

for PAHs. Chapter 4 includes detailed materials and methods for all aspects of this

research project. Chapter 5 is presented as a report of the results and discussion section

of the street sweeping study, divided into two main sections: the results of the street

particulate matter monitoring and the results of the stormwater quality monitoring.

Chapter 6 covers the results of the bioretention study, and is presented as a stand alone

paper suitable for submission to a journal for publication. Chapter 7 summarizes the

major results of this research and provides recommendations for future research on street

sweeping and bioretention as stormwater BMPs. An appendix follows with detailed

datasheets that could not be included in the main body of the thesis.

Both the street sweeping and bioretention studies were collaborative projects and

it is important to clarify my role in these research efforts. In the street sweeping study,

my personal contribution included: a 6 month literature review of street sweeping as a

- 13 -

stormwater management practice, field coordination and collection of all street

particulate matter samples with Baltimore City DPW, particle type analysis of street

particulate matter, and analysis of all water quality and street particulate matter data. All

major laboratory analyses of street particulate matter and stormwater were performed by

either Baltimore County or Microbac Laboratories. All statistical analysis of data was

performed in partnership with two statistician consultants, Dr. Nagaraj Neerchal and Yan

Zhuang. For the bioretention study, my role included: laboratory analysis of stormwater

influent and effluent samples for polycyclic aromatic hydrocarbons (PAHs) for five storm

events. Stormwater sampling was performed by Houng Li and Dr. Allen Davis at

University of Maryland, College Park. Additionally, I performed bioretention media core

sampling, and analysis of core segments for PAHs.

1.5 References

Burton, G.A. Jr. and Pitt, R.E. (2002). Stormwater Effects Handbook, A Toolbox for Watershed Managers, Scientists, and Engineers. Boca Raton: Lewis Publishers of CRC Press LLC.

Center for Watershed Protection (2006). Monitoring Plan – Deriving Reliable Pollutant

Removal Rates for Municipal Street Sweeping and Storm Drain Cleanout Programs in the Chesapeake Bay Basin. Updated 01/23/2006 by Neely Law.

Chesapeake Bay Program (2006). Urban Storm Water Fact Sheet. Last modified

01/11/2006. http://www.chesapeakebay.net/stormwater/htm. Chesapeake Bay Foundation (2003). Water Pollution in the Chesapeake Bay Fact Sheet.

Last modified 07/2003. http://www.cbf.org/site/PageServer?pagename=resources_facts_water_pollution

Clement, C., Bricker, S.B., Pirhalla, D.E. (2001). Eutrophic Conditions in Estuarine

Waters. In: NOAA’s State of the Coast Report. Silver Spring, MD: NOAA. Davis, A. P. (2005). Green Engineering Principles Promote Low-Impact Development.

Environmental Science & Technology A Pages, Vol. 39, pp. 338A – 344A.

- 14 -

Metcalf and Eddy (2003). Wastewater Engineering: Treatment and Reuse. Fourth Edition, Revised by: Tchobanoglous, G., Burton, F. L., and Stensel, H. D. McGraw Hill, New York, NY.

Richardson, D.C. (2006). Watershed 263: A Resource Uncovered. Stormwater. Volume 7, No. 6., September 2006.

http://www.gradingandexcavating.com/sw_0609_watershed.html Strecker, E. W., Quiqley, M. M., Urbonas, B. R., Jones, J. E., and Clary, J. K. (2001).

Determining Urban Storm Water BMP Effectiveness. Journal of Water Resources Planning and Management, May/June 2001, pp. 144 – 149.

U.S. Environmental Protection Agency (1981). Process Design Manual - Land

Treatment of Municipal Wastewater. EPA 625/1-81-013. October 1981. U.S. Environmental Protection Agency (1993). Guidance Specifying Management

Measures for Sources of Nonpoint Pollution in Coastal Waters. EPA 840-B-92-001 January 1993.

U.S. Environmental Protection Agency (2002). The State of the Chesapeake Bay

Watershed: U.S. Environmental Protection Agency Report 903-R-02-002. U.S. Environmental Protection Agency (2003). Relative Risk Assessment of

Management Options for Treated Wastewater in South Florida. EPA 816-R-03-010.

U.S. Environmental Protection Agency (2006a). Phases of the NPDES Stormwater

Program. Last Updated 03/10/2006 http://cfpub.epa.gov/npdes/stormwater/swphases.cfm

U.S. Environmental Protection Agency (2006b). Managing Urban Runoff, Pointer No. 7.

EPA841-F-96-004G. http://www.epa.gov/owow/NPS/facts/point7.htm Last Updated November 2006.

U.S. Environmental Protection Agency (2007a). The National Estuary Program Coastal

Condition Report – Fact Sheet. EPA-842-F-06-001. http://www.epa.gov/owow/oceans/nepccr/

U.S. Environmental Protection Agency (2007b). Stormwater Overview in Mid-Atlantic

Stormwater Quick Finder. http://www.epa.gov/reg3wapd/stormwater/ Last updated October 10, 2007.

U.S. Environmental Protection Agency (2007c). Permit Application Requirements for Medium and Large MS4s. http://cfpub.epa.gov/npdes/stormwater/lgpermit.cfm Last updated April 09, 2007.

- 15 -

2. BACKGROUND

2.1 Street Particulate Matter

Street sweeping ranks among the oldest practices used to control storm water

pollution, yet very limited and sometimes conflicting data has been published in regard to

its performance in removing nutrients and other pollutants (Burton and Pitt 2002, U.S.

EPA 1983, Mineart and Singh 1994, Sutherland and Jelen 1997). Historically, street

sweeping programs focused on aesthetics and maintaining sanitary conditions rather than

pollution reduction and stormwater management. However, as environmental awareness

increased in the 1960s and 1970s, the goals and motivations of street sweeping programs

began to shift. The knowledge that stormwater runoff contains substantial quantities of

contaminants emerged, and interest in the sources of stormwater contaminants led to

investigations into the materials which commonly reside on street surfaces.

There are many pollution sources within a catchment that contribute to the street

surface material load, including runon from adjacent land areas, vehicle emissions and

wear, atmospheric deposition, deterioration of the street surface, and direct deposits such

as sanding and littering. The combined material generated from these sources may be

defined as the total street load, and often contains sediment, paper and plastic litter, glass,

and vegetation. The street surface material smaller than 5 mm in size, here called street

particulate matter, is the focus of this discussion. For the purpose of this study, street

particulate matter is defined as any sediment, soil, and other particulate matter as well as

any vegetation or debris smaller than 5 mm and any associated pollutants, that resides on

the street surface and can be washed off by a storm event of sufficient intensity (at least

- 16 -

2.54 mm (0.1 inches)). Although the amount that each source contributes to the total

street particulate matter load and the composition of street particulate matter can vary

greatly from site to site, it is established that street surfaces are significant sources of

urban stormwater pollutants (Pitt and Amy 1973, Sartor and Gaboury 1984, Pitt 1985,

Waschbusch et al. 1999).

Bannerman and others (1993) found that streets were the “single most important

source area” for pollutants in urban runoff among five urban source areas (lawns,

driveways, rooftops, parking lots, and streets). Streets contributed the largest

concentrations of suspended solids, bacteria, and metals; with street inputs measuring

about four to eight times more than the other sources. A more recent study by

Waschbusch and others (1999) also reported that streets are the major source of

suspended solids in urban runoff, contributing about 70-80% of the total. Streets are also

typically considered the second most important source of nutrients in stormwater, with

lawns being the largest source (Pitt 1985, Bannerman et al. 1993, Waschbusch et al.

1999). Both Pitt (1985) and Waschbusch (1999) found that streets contribute about 20-

30% of the nitrogen and phosphorus load, with streets and lawns together contributing

70-80%. Focusing specifically on streets as a stormwater pollutant source, it is important

to discuss the specific sources of street particulate matter and its associated pollutants, as

well as the physical and chemical characterization of street particulate matter.

2.1.1 Sources of Street Particulate Matter and Associated Pollutants

This section provides a discussion of five major sources of street particulate

matter and its associated pollutants: runon from adjacent land areas, vehicle emissions

- 17 -

and wear, deterioration of the street surface, atmospheric deposition, and anti-skid

compounds. Street particulate matter contributed from adjacent land areas includes both

runon from impervious areas, and the erosion of local soils from pervious areas. Pitt

(1985) found that for very small rains, impervious areas contributed the majority of

pollutants, but when rain increased (to more than 2.5 mm (0.098 inches)) the contribution

of pervious areas became more important. Runon from impervious areas, such as

rooftops, sidewalks, driveways, and parking lots, can transport particulate matter, as well

as oils, grease, and other toxic substances that have accumulated during the interim storm

period. Erosion of local soils can be a result of rain or wind, and is typically one of the

largest sources of street surface particulates (Sartor and Boyd 1972, Pitt 1979). In areas

where soil erosion is a major source of street particulate matter, the type of parent

material found within the drainage basin may impact the grain size of the street

particulate matter. Therefore, local geology can play a large role in the makeup of street

surface contaminants. For example, in Florida where sandy soil is common, much of the

street particulate matter is coarse grained (Brinkmann and Tobin 2001). However, in

cold weather climates, the use of sand as a de-icing material will contribute to the particle

size distribution of street particulate matter, which may mask effects of local geology.

Street particulate matter load originating directly from vehicle emissions and wear

may contribute only a small percentage (by weight) of the total load, but these small

amounts are often very toxic (Sartor and Boyd 1972, Pitt et al. 1997). Aside from spills

and leaks of vehicle fluids, the normal operation and wear of vehicles is responsible for

significant amounts of contaminants on the street, particularly metals. A study of

nonpoint sources in the San Francisco Bay area found that vehicles contribute more than

- 18 -

50% of the copper, cadmium and zinc entering the Bay (Santa Clara Valley NSCP 1992).

The majority of the cadmium and zinc load can be attributed specifically to automobile

tire wear, and copper is mainly due to the wear of brake pads. Vehicle emissions,

particularly diesel engines, are often considered the primary source of lead, and also a

significant source of chromium, silver, mercury, copper and zinc (Santa Clara Valley

NSCP 1992).

The contaminant contribution made by wear or deterioration of the street surface

is strongly dependent on the condition of the road. Pitt (1979) found that the

contributions from the wear of smooth streets in good condition are insignificant. In

general, the rougher the road surface, the more sediment will erode from the street

surface. Sediment on rough streets may also contain more large particles than the

sediment of a smooth street. During a 1983 Milwaukee, Wisconsin field project (WI

DNR 1983), a single asphalt site was divided into two sections, one smooth in good

condition, and one rough in poor condition. Street loads at the rough site were three to

four times greater than the smooth site, and the smooth section had a greater percentage

of particles smaller than 250 microns than the rough section (WI DNR 1983).

Another source of street particulate matter is atmospheric deposition, which can

deliver finer sized particles by either dry or wet deposition. Some of the sources

discussed above, such as vehicle emissions and soil erosion due to wind, are often

grouped into the larger source category of atmospheric deposition. Nutrients, metals and

other pollutants are often found in both wetfall and dryfall samples and may vary by land

use type (NC DNRCM 1983, WI DNR 1983).

- 19 -

In some regions of the United States, anti-skid compounds such as salts (NaCl and

CaCl2), sand, and ash are frequently applied to roadways to melt ice and increase traction

during cold weather. Aside from adding to the total street load, fluctuating

concentrations of NaCl entering receiving waters can have detrimental effects on the

local ecosystems (Hvitvet-Jacobsen and Yousef 1991, Kaushal et al. 2005). The

accumulation of chloride or other anti-skid materials on streets is likely traffic-dependent

because the application of these compounds is focused on well-traveled streets (Pitt 1979).

2.1.2 Physical and Chemical Characterization of Street Particulate Matter

The amount of material on the street surface varies greatly by land use, street

condition and material, and environmental conditions. A typical amount of street

particulate matter load ranges from 250-300 g/curb-meter (887 to 1064 lbs/curb-mile) as

reported by Sartor and Gaboury (1984). In general, the quantity of material on a street

surface depends largely on the length of time since the last cleaning, either by sweeping

or rainfall. However, past research has also revealed a few other trends in street

particulate matter loadings.

Higher loadings on uneven, cracked streets is due to both the degradation of the

street surface itself and the street particulate matter getting trapped in the cracks or

potholes, where it is less likely to be removed by wind, rain, or street sweeping (Sartor

and Boyd 1972, Pitt 1979, WI DNR 1983). Shaheen (1975) also found that loadings

increase with curb height and that at most sites there is a strong positive relationship

between traffic volume and street particulate matter loadings.

- 20 -

Another important trend is that street particulate matter is not distributed

uniformly across a street surface. Sartor and Boyd (1972) concluded that a typical cross-

section of a street will have a small amount of material at the center or crown, very little

in the traffic lanes, a substantial amount in the parking lane, and the largest amount in the

gutter area. This uneven distribution is due to a higher elevation at the crown and due to

wind and traffic blowing the street particulate matter out of the vehicle lanes and into the

gutter area where it is contained by the curbs and accumulates. Sartor and Boyd (1972)

estimated that about 70-80% of all street debris lies within six inches of the curb, and

90% lies within 12 inches. Less extreme across-street loadings were observed by Pitt

(1979) and the NC DNRCM (1983), but they still found that between 40-80 percent of

the total street particulate matter lies within one foot of the curb.

Particle size distribution of street particulate matter varies among individual sites,

but typically the sand fraction (250-1000 �m) contains the highest percentage of particles

by weight. Particle size distribution data from four previous studies on street sweeping

are shown in figure 2.1.

In three of the four studies, sand particles were the largest fraction, making up 30-

50% of the total street particulate matter load, by weight. The fine sand fraction (63-250

�m) fraction also contributed a significant amount of material, about 20-25% of the total

load. While the clay and silt fraction, particles less than 63 �m in size, typically makeup

less than 10% of the total mass, studies have shown that this fraction contains a

disproportionate amount of contaminants. Considerable quantities of metals, nutrients,

and organics are transported from the street surface to receiving waters as sediment-

bound contaminants. This brief discussion of the chemical characterization of street

- 21 -

particulate matter will focus on nutrients and heavy metals with respect to particle size

and source.

0

10

20

30

40

50

60

>1 0.25-1 0.063 - 0.25

- 22 -

Some forms of nutrients are more strongly associated with street surface solids

than other nutrients, and nitrogen is an excellent example. Terstriep and others (IL

DENR 1982) measured the dissolved fraction for a number of constituents in urban

runoff and found that nearly 100 percent of ammonia-N and nitrite-nitrate N is dissolved

in runoff, meaning they have no clear relationship with solids. Supporting this finding,

Sartor and Boyd (1972) also noted that relative to the other contaminants measured,

nitrates were found in very small amounts in street particulate matter. The dissolved

fraction of Kjeldahl nitrogen was found to be smaller than ammonia and nitrite-nitrate,

measured at 69%, and therefore this parameter is more strongly associated with street

particulate matter (IL DENR 1982).

Heavy metals found in street particulate matter can be largely attributed to

automobile emissions and wear and commonly include: Pb, Fe, Zn, Ca, Cd, Cr, Cu, Hg,

Ni, Mn and Fe (Sartor and Boyd 1972, Shaheen 1975, Wilbur and Hunter 1979,

Fergusson and Ryan 1984). Concentrations of heavy metals in street particulate matter

can vary widely from site to site, and also with respect to land use. Fergusson and Ryan

(1984) measured metals in street particulate matter in five cities in the U.S. and Europe

and found that lead concentrations ranged from a few hundred ppm to over 10,000 ppm.

When looking at the three traditional land use types (residential, industrial and

commercial), Sartor and Boyd (1972) found that commercial sites had the highest

concentrations of zinc, lead, copper, and mercury. This was likely due to the higher

traffic volume that often occurs at commercial shopping areas. It is widely accepted that

heavy metals are strongly associated with the fine particle size fractions in street

particulate matter, and concentrations generally increase with decreasing particle size

- 23 -

(Shaheen 1975, Fergusson and Ryan 1984, Hvitved-Jacobsen and Yousef 1991, Schorer

1997). Sartor and Boyd (1972) found that more than 50% of all heavy metals are

associated with sediment particles smaller than 246 microns. Similar distributions were

observed by Shaheen (1975) who found that about 54% of zinc and 62% of lead are

associated with particles smaller than 250 microns.

However, for any contaminant, the physical composition of the street particulate

matter at each site must be taken into consideration as well, and depending on the particle

size distribution at a site, the high concentrations associated with the smallest size

fractions may be negligible. Therefore, the concentration of each size class must be

multiplied by the total solids load contributed by that size class to calculate the percent of

the total pollutant load in each class (Concentration x Mass = total pollutant load). In

general, the particle size fraction that contributes the greatest portion of the mass often

also contributes the greatest portion of the contaminant load.

In conclusion, material on the street surface is highly contaminated and

considered a major pollution source to urban stormwater. This knowledge emerged in the

early 1970s through a number of EPA studies (Sartor and Boyd 1972, Pitt and Amy 1973,

Shaheen 1975), along with the suggestion that street sweeping could possibly reduce

street particulate matter as a pollutant source. As a response to these findings, a number

of intensive field monitoring studies on stormwater runoff and street sweeping as a

stormwater pollution control practice were initiated.

- 24 -

2.2 Street Sweeping as a Stormwater BMP

2.2.1 Literature Review of Street Sweeping as a Stormwater BMP

In the late 1970s and early 1980s, the U.S. EPA conducted the Nationwide Urban

Runoff Program (NURP) at 28 cities across the country. While the 28 individual NURP

projects were different, they all focused on characterizing pollutant types, their effect on

water quality, and evaluating various practices for the control of stormwater pollution.

Five of the 28 NURP studies (Bellevue, WA; Champaign-Urbana, IL; Castro Valley, CA;

Winston Salem, NC; Milwaukee, WI) studied the effectiveness of street sweeping in

reducing pollutants in stormwater runoff (U.S. EPA 1983). These five NURP studies

were designed using a paired basin approach, where two adjacent or nearby basins were

monitored during a control (often unswept) period to establish baseline data and then

street sweeping was implemented or intensified in one treatment basin, while the other

remained as a control. Stormwater quality monitoring was performed throughout the

baseline and treatment periods to determine any impacts of street sweeping on

stormwater quality.

All five of the NURP studies found that street sweeping was ineffective in

reducing mean concentrations of pollutants in urban storm runoff, and therefore

ineffective as a stormwater management practice (Pitt and Shawley 1981, IL DENR 1982,

WI DNR 1983, NC DNRCM 1983, Pitt 1985). These studies reported that while street

sweeping was effective at removing litter and larger particles, the NURP-era street

sweepers were unable to pick up the finer grained sediment fraction (

- 25 -

not translate to significant reductions in stormwater pollutants. In the final report of the

NURP results (U.S. EPA 1983), the issue of variation in stormwater Event Mean

Concentrations (EMCs) was also a major concern. For each study site, median EMC data

for five parameters (TSS, COD, TP, TKN and Pb) was based on between 10 to 60 storm

events, with 30 events typical. The U.S. EPA (1983) reported that no reduction of

contaminant EMCs observed was greater than 50% and concluded that any benefits of

street sweeping that did occur were masked by the large variability of the EMCs, and

therefore a larger database is required. The NURP results led many street sweeping

manufacturers to focus on improving technology to pickup fine particles.

Ten years after the NURP-era studies, in the mid 1990s, interest in street

sweeping as a best management practice was renewed. Improvements in sweeping

technology with the development of vacuum sweepers and regenerative air sweepers

were thought to have improved their ability to pick up fine particulates. Further, new

stormwater regulations created the need for best management practices that were

relatively inexpensive and easy to implement in urban watersheds and street sweeping

provided an option.

More recent street sweeping studies (completed in the past 15 years) have been

more narrowly focused in project and geographic scope (e.g., Brinkmann and Tobin 2001,

Kuhns et al. 2003). In general, these studies involved sampling street particulate matter

or stormwater that largely compared the effectiveness of different types of sweepers, and

had less extensive field monitoring components compared to the NURP-era studies.

Researchers have also utilized modeling as a tool to assess the impacts of street sweeping

on stormwater quality (Sutherland and Jelen 1996, Sutherland and Jelen 1997, Zariello et

- 26 -

al. 2002). Another current trend in street sweeping research is the evaluation of sweeping

for the removal of total suspended particles (TSP) and particles less than 10 micrometers

(PM10) to improve air quality (Chang et al. 2005, Kuhns et al. 2003). The U.S. EPA is

particularly concerned with inhalable particles 10 �m in diameter or smaller (PM10)

because they can enter the body where they can affect the heart and lungs and cause

serious health effects (U.S. EPA 2006b).

As street sweeping studies have continued to focus on quantifying the amount of

material removed little is known about how these practices affect water quality on a

catchment or watershed scale. As a consequence there remains limited understanding on

how street sweeping can be used as a best management practice to effectively reduce

pollutant loadings to improve and maintain water quality. Although street sweeper

pickup performance and variability in stormwater quality sampling are important, there

are many other factors that can influence the effectiveness of street sweeping in

improving stormwater quality.

To determine the stormwater pollutant load reduction that may be achieved by

street sweeping, one needs to understand the factors and processes that effect street

sweeping performance. Pollutant load reduction is defined as the loading rate (e.g.

deposition) of street particulate matter less the rate at which material is washed off, plus

material that may be considered permanent storage in road cracks, etc. (Pitt et al. 1997).

The deposition and removal of street particulate matter depends on multiple factors that

are location specific and include the physical characteristics of the catchment,

environmental conditions such as weather, the design and operation of the street

sweeping program and the particle size distribution of pollutants.

- 27 -

In addition to these factors, streets are only one source of stormwater pollutants.

As previously discussed, potential sources of stormwater pollutants include streets,

parking lots, sidewalks, rooftops, driveways, and pervious areas. Therefore, reducing the

street particulate matter load through street sweeping is only reducing one of several

sources. Street sweeping has the highest potential to reduce stormwater concentrations of

contaminants for which streets are the major source (such as suspended solids) and which

are largely associated with solids (such as lead).

To date, although there is limited evidence to show how reductions in street

particulate matter can affect stormwater quality, most studies and street sweeping

programs continue to operate under the premise that if street particulate matter is

removed, then an improvement in water quality should follow.

2.2.2 Conceptual Model

To better understand the impact of street particulate matter and street sweeping on

stormwater quality, it is valuable to look at a conceptual model (Figure 2.2) of the inputs,

outputs, and fate of street particulate matter in an urban catchment. The conceptual

model incorporates three major components, the street surface, the storm drain or inlet,

and the storm drain pipe network. As discussed previously, there are many inputs from

many source areas of street particulate matter to the street surface including: littering,

runon, degradation of the street surface, vehicle emissions and wear, sand/salt, wet

deposition and dry deposition. Once the material is on the street surface, there are four

major removal mechanisms or outputs: street sweeping, loss by wind and traffic, washoff

to the storm drain, and runoff (not to an inlet). It is important to remember that the entire

- 28 -

load of street particulate matter is not available to be removed by these mechanisms. For

example, street sweeping cannot remove materials underneath parked cars and washoff

may not remove materials trapped in cracks or potholes of the street surface. Street

particulate matter that remains on the street surface may then be deposited in the storm

drain inlet by a rain event (washoff) or by wind, traffic or direct littering. Some material

is typically trapped at the bottom of the inlet, and inlet trapping efficiency is a function of

the type and capacity of the inlet.

- 29 -

Figure 2.2: Conceptual model of the fate of street particulate matter in an urban catchment.

Sediment Pool Water

Curb Street Surface

Inputs

Littering Runon Degradation of street surface Vehicle emissions and wear Sanding/Salt Wet deposition Dry deposition

Outputs Street Sweeping Runoff (not to inlet) Loss by wind and traffic

Dry Inputs by wind, traffic littering Washoff

Inputs to Storm Drain

Municipal Storm Drain Cleanout

Outputs

Bedload (settled solids)

Storm Drain Network

Receiving Waters

Stormwater or Baseflow

Flush out by large storm

- 30 -

Inlets will store this material until it is removed by municipal storm drain cleanout

or flushed out into the storm drain network by a storm event of adequate size. Once

material is flushed inside the storm drain network of pipes, larger particles and debris will

settle to the bottom of the storm drain pipe (called bedload) while other material is

carried in the stormwater to local receiving waters. Bedload may remain on the bottom

of a storm drain pipe until a storm of sufficient intensity removes it and transports it to

receiving waters as well. This conceptual model illustrates the interconnectedness of

streets, storm drain inlets and storm drain systems and the many factors which impact

stormwater quality in an urban catchment.

While street sweeping has the potential to remove street particulate matter before

it is picked up by stormwater runoff, another practice, bioretention, has the potential to

remove suspended solids and associated pollutants in stormwater runoff.

2.3 Bioretention as a Stormwater BMP

2.3.1 Bioretention Concept

Bioretention is a relatively new practice, which gets its name from the concept of

using biomass to retain pollutants in stormwater runoff. Developed in the early 1990s in

Prince George’s County, Maryland, bioretention is a natural-based BMP system also

known as a rain garden (Glass and Bissouma 2005). The practice of bioretention is

considered part of the larger concept of low impact development (LID), which integrates

environmental concerns with land development. In contrast to traditional management of

urban stormwater, which attempts to convey runoff away from a developed area as

quickly as possible, LID aims to manage stormwater runoff at the source (Davis 2005).

- 31 -

LID is considered a philosophy for development that focuses on minimizing

adverse environmental impacts and practicing sustainable water management. One of the

ultimate goals of LID designs is to replicate pre-development hydrologic conditions as

closely as possible. The practice of LID begins with site design and includes practices

such as: leaving wooded areas on lots, minimizing impervious cover by narrowing streets,

and using vegetated swales and filter strips in place of traditional curb and gutter systems

(Davis 2005). Obviously, some impervious area cannot be avoided and excess runoff can

be managed with onsite vegetated infiltration practices such as bioretention.

Vegetated bioretention areas or rain gardens, are management practices designed

to interrupt the flow of runoff from impervious surfaces to storm drain systems or

receiving waters. Bioretention is typically designed as individual cells (sized at

approximately 4-5% of the drainage area), which include soil, sand, organic matter and

vegetation engineered to store, infiltrate and treat stormwater runoff (Davis et al. 2003,

Hsieh et al. 2007). Many mechanisms can occur within a bioretention cell to improve

water quality, including: adsorption, precipitation, filtration, evapotranspiration and bio-

transformation processes (Davis et al. 2003, Hsieh and Davis 2005a). Water which

infiltrates a bioretention facility is either allowed to continue for groundwater recharge or

collected through a subsurface perforated pipe (underdrain) and conveyed to receiving

waters or traditional storm drain systems (Davis et al. 2001). Although designs vary, a

typical bioretention cell cross-section is shown in figure 2.3. As runoff enters a

bioretention area, typically through a cutout in the curb length, it will flow through

several cell components including (from top to bottom): native vegetation, a 7-15 cm (3-6

inch) ponding area, a 5-8 cm (2-3 inch) mulch layer, a soil layer (composed of 20-30%

- 32 -

top soil, 20-30% leaf compost, and 50% sand), and an underdrain surrounded by gravel

(Prince George’s County 2002).

Figure 2.3: Typical cross section of a bioretention cell. Source: Modified from Prince George’s County (2002).

There are many different designs and ways to implement bioretention as a

stormwater best management practices in commercial, industrial and residential areas.

Some examples illustrated in a Bioretention Manual published by Prince George’s

County in 2002 include: parking lot (curbed or uncurbed) perimeter bioretention, parking

lot (curbed or uncurbed) island and median bioretention, parking swale bioretention,

rooftop bioretention, residential on-lot bioretention, and tree and shrub pit bioretention.

Ponding Area

Mulch Layer

Soil Medium

Vegetation

1.5-3 ft., typically: 20-30% topsoil 20-30% leaf compost 50+% sand

1.5-2” Stone

2-3” provides excellent solids and metals removal

3-6”

Underdrain and Gravel Filter

- 33 -

A picture of a typical curbed parking lot island and median bioretention area is shown in

figure 2.4. In this photograph, the flow entrance cutout in the curb surrounding the

bioretention cell can clearly be seen.

Figure 2.4: Typical curbed parking lot island and median bioretention. Source: Prince George’s County (2002). Although bioretention is a relatively new practice, there have been several studies

evaluating the effectiveness of bioretention in removing stormwater pollutants both in the

lab and in the field.

2.3.2 Literature Review of Bioretention as a Stormwater BMP

The majority of bioretention research to date has focused on removal of

suspended solids, oil and grease, metals and nutrients in stormwater runoff. Bioretention

studies both in the lab and in the field have shown removal efficiencies of 91-98% for

suspended solids (Hsieh and Davis 2005a, Hsieh and Davis 2005b, Glass and Bissouma

2005). A laboratory study by Hsieh and Davis (2005a) used a bioretention test column to

- 34 -

treat a synthetic runoff event once a week for twelve weeks. During this time, the authors

examined the removal of total suspended solids and the long term effects of these filtered

solids on stormwater infiltration rates. For all but one storm, more than 90% of the total

suspended solids (TSS) was removed by the bioretention media, and the authors observed

that most of the TSS was filtered by the top mulch layer. Additionally, the filtered solids

had no effect on runoff infiltration rate, which remained constant at 0.35 cm/min

throughout the study (Hsieh and Davis 2005a). When visually examining the influent

and effluent stormwater at field bioretention facilities, color differences indicated that

most of the suspended solids in the effluent are materials from the bioretention media

rather than part of the influent TSS (Hsieh and Davis 2005b).

Oil and grease are also removed with excellent efficiencies (greater than 96%) in

both laboratory column studies and field studies at existing bioretention sites (Hsieh and

Davis 2005a, Hsieh and Davis 2005b). The removal of heavy metals (copper, lead and

zinc) through bioretention is also typically high (greater than 90%) and is often correlated

to the removal of TSS (Davis et al. 2003, Hsieh and Davis 2005b). However, these

values were reported for laboratory scale studies or well maintained field sites using

synthetic runoff. A study of 15 storm events at an existing bioretention facility by Glass

and Bissouma (2005) showed lower removals for Cu (75%), Pb (71%) and Zn (81%) and

the authors attributed this to lack of maintenance of the bioretention mulch layer. While

the mulch layer acts as a filter for suspended solids, Davis and others (2001) determined

that metal removal also occurs via adsorption to the mulch layer and soil media as water

flows though the cell.

- 35 -

Although data vary with regard to nutrients, bioretention cells are typically more

effective at removing phosphorus from stormwater than nitrogen. Total phosphorus (TP)

and TKN removal efficiencies are moderate and quite variable, ranging from about 50-

85% (Davis et al. 2006, Hsieh and Davis 2005a, Hsieh and Davis 2005b). Nitrate

reduction is consistently poor (less than 20%), and studies report that nitrate removal may

be improved by adding an engineered denitrification layer (Kim et al. 2003, Hsieh and

Davis 2005b).

A study comparing different types and configurations of bioretention media found

that a layered medium with a permeable sand/soil mixture layer provided the best overall

pollutant removal efficiencies for bioretention (Hsieh and Davis 2005b). However,

bioretention cells should be designed for individual sites depending on infiltration needs,

pollutant removal goals and any other site concerns such as groundwater contamination.

As with any stormwater BMP, it is important to consider both the maintenance

costs and long-term efficiency of the practice. At sites where pollutant loadings in urban

stormwater are average or lower, long bioretention lifetimes (up to 20 years) are possible

(Davis et al. 2003). During this time, plant matter should be somewhat maintained, and

replacing the mulch layer may also extend the lifetime of a cell (Davis et al. 2001).

However, at sites where pollutant loadings are high, pollutants will accumulate more

quickly in the bioretention cell and maintenance and eventually disposal of bioretention

cell media are more complicated. Davis and others (2001) suggest that another option is

to use media which captures metals or other pollutants and renders them biologically

unavailable.

- 36 -

In addition to improving stormwater quality, bioretention cells, and particularly

the plant life they support, have several other benefits. A properly designed bioretention

cell planting scheme can improve aesthetics, provide shade, shield wind, support wildlife,

adsorb noise and increase the value of a site (Prince George’s County 2002).

In conclusion, bioretention is a promising stormwater BMP which deserves

further research attention, particularly in regard to other types of pollutants. Since

bioretention is very effective at removing suspended solids in stormwater runoff, this

practice has the potential to remove many contaminants which are strongly associated

with solids, such as hydrophobic organic compounds. Polycyclic aromatic hydrocarbons

(PAHs) are a growing concern in stormwater runoff, and the fate of PAHs in stormwater

treated by bioretention is the focus of this study. Before describing the study objectives,

it is helpful to describe PAHs, their characteristics, sources and inputs to stormwater

runoff.

2.3.3 Polycyclic Aromatic Hydrocarbons (PAHs)

Polycyclic aromatic hydrocarbons (PAHs) represent the largest class of suspected

carcinogens (Bjorseth and Ramhahl 1985) and they are widely distributed in the air,

water and sediments of urban environments (Van Metre et al. 2000, Larsen and Baker

2003, Stout et al. 2004, Stein et al. 2006). There are more than 100 different PAH

compounds and although many are naturally occurring, the majority of PAHs in the

environment are from anthropogenic sources (Agency for Toxic Substances and Disease

Registry 1996). Anthropogenic sources of PAHs include the release of petroleum

- 37 -

products (petrogenic source) and the combustion of organic matter such as petroleum,

coal, wood and oil (pyrogenic source) (Stein et al. 2006).

PAH compounds generally have low solubilities (Table 2.1) and are typically

associated with the solids rather than liquids in the environment. It is also notable that

PAH compounds range from two rings to six rings, and the solubility of PAHs tends to

decrease as the number of rings increase (Bjorseth 1983).

Table 2.1. EPA’s 16 priority pollutant PAHs.

Source: Modified from Stein et al. (2006) and Peters et al. (1999).

In urban environments, studies have found that pyrogenic sources dominate (Hoffman

1985, Menzie et al. 2002, Stein et al. 2006). Specifically, urban sources may include:

vehicle exhaust, home heating through coal and wood burning stoves, trash burning,

power plants and other industrial processes, and the leaching of PAHs in sealant used to

coat parking lots and driveways. Because PAHs are ubiquitous in urban environments

and largely associated with solid particles, they are also a very accessible contaminant to

16 PAHs Mol. Wt.

No. Rings

Aqueous Solubility at 25ºC (mg/L)

Naphthalene 128 2 31 Acenaphthylene 152 3 3.9 Acenaphthene 154 2 3.8 Fluorene 166 3 1.9 Phenanthrene 178 3 1.1 Anthracene 178 3 0.05 Fluoranthene 202 4 0.26 Pyrene 202 4 0.13 Benz(a)anthracene 228 4 0.011 Chrysene 228 5 0.002 Benzo(b)fluoranthene 252 5 0.0015 Benzo(k)fluoranthene 252 5 0.0008 Benzo(a)pyrene 252 5 0.004 Indeno(1,2,3-cd)pyrene 276 6 0.062 Dibenz(a,h)anthracene 278 5 0.0005 Benzo(g,h,i)perylene 276 6 0.0003

- 38 -

be picked up and transported by stormwater runoff. Urban stormwater runoff is

considered an important source of PAHs to aquatic environments. Studies have found the

pattern of PAHs observed in near shore sediments closely reflects the pattern of PAHs

found in urban stormwater runoff. It is estimated that urban runoff contributes about 14-

36% of the total PAH load to aquatic ecosystems (Hoffman et al. 1984, Menzie et al.

2002).

A 1985 study in Narragansett Bay (Hoffman 1985) and a study by Menzie and

others (2002) both found that fluoranthene, phenanthrene, pyrene and chrysene are the

dominant PAH compounds in stormwater. All four of these compounds are categorized

by the U.S. EPA as category D – not classifiable as to human carcinogenicity, based on

no human data and inadequate data from animal assays. However, the U.S. EPA

classifies the PAH, Benzo(a)pyrene (BaP) as a probable human carcinogen (category B2),

and states that PAHs similar to BaP can potentially cause adverse health affects due to

both acute and chronic exposure (U.S. EPA 2006a). Possible exposure pathways to

humans include the inhalation of contaminated air, smoking of cigarettes, and

consumption of contaminated food (particularly grilled and smoked foods) and water

(U.S. EPA 2006c). Stormwater is a source of BaP and other PAHs to receiving waters,

where they often accumulate in sediments and their toxicity can affect sediment and

aquatic life.

A recently discovered source of PAHs in urban environments, which is

particularly relevant to stormwater runoff, is sealers used to coat parking lots and

driveways. These sealers, used to improve the appearance of asphalt pavement are most

frequently made with a coal-tar base, however asphalt-emulsion based sealers are also

- 39 -

used (Mahler et al. 2004). A 2003 study in Austin, Texas by Mahler and others (Mahler

et al. 2004) investigated the concentrations and loads of PAHs in simulated runoff from

various parking lot surfaces and determined to what degree these sealers are a source of

PAHs in the urban environment. The authors reported that stormwater suspended solids

from parking lots with coal-tar emulsion sealcoat had mean PAH concentrations of 3500

mg/kg, 65 times higher than the mean PAH concentration in runoff from unsealed asphalt

and cement parking lots (Mahler et al. 2005).

Although the fate of PAHs in stormwater is a significant concern, the concept of

treating PAHs in urban stormwater has received little attention from researchers and

municipal stormwater operators. Bioretention has the potential to reduce PAH loads

from urban stormwater runoff before it enters the storm drain network and enters

receiving waters. The following section describes the objectives for both research

components included in this thesis, the street sweeping study and the bioretention cell

study.

2.4 References

Bannerman, R., Owens, R., D., Dodds, R., and Hornewer, N. (1993). Sources of Pollutants in Wisconsin Stormwater. Water Science and Technology, 28:3-5, 241-259.

Brinkman, R. and Tobin, G. A. (2001). Urban Sediment Removal: The Science, Policy,

and Management of Street Sweeping. Boston: Kluwer Academic Presses. Bjorseth, A., Ed. (1983) Handbook of Polycyclic Aromatic Hydrocarbons;

Marcel Decker: New York, 1983. Bjorseth, A., Ramdahl, T., Eds. (1985) Handbook of Polycyclic Aromatic Hydrocarbons;

Marcel Decker: New York, 1985.

- 40 -

Burton, G.A. Jr. and Pitt, R.E. (2002). Stormwater Effects Handbook, A Toolbox for Watershed Managers, Scientists, and Engineers. Boca Raton: Lewis Publishers of CRC Press LLC.

Center for Watershed Protection (2006). Technical Memorandum 1 – Literature

Review: Research in Support of an Interim Pollutant Removal Rate for Street Sweeping and Storm Drain Cleanout Activities. Prepared for the U.S. Chesapeake Bay Program, October 2006.

Chang, Y.M., Chou, C.M., Su, K.T., and Tseng, C.H. (2005). Effectiveness of Street

Sweeping and Washing for Controlling Ambient TSP. Atmospheric Environment 39: 1891 – 1902.

Davis, A. P., Shokouhian, M., Sharma, H., Minami, C. (2001). Laboratory Study of

Biological Retention for Urban Stormwater Management. Water Environment Research, Volume 73, Number 1, pp. 5-14.

Davis, A. P., Shokouhian, M., Sharma, H., Minami, C., Winogradoff, D. (2003). Water

Quality Improvement through Bioretention: Lead, Copper, and Zinc Removal. Water Environment Research, Volume 75, Number 1, pp. 73 – 82.

Davis, A. P. (2005). Green Engineering Principles Promote Low-Impact Development.

Environmental Science & Technology A Pages, Vol. 39, pp. 338A – 344A. Davis, A. P., Shokouhian, M., Sharma, H., Minami, C. (2006). Water Quality

Improvement through Bioretention Media: Nitrogen and Phosphorus Removal. Water Environment Research, Volume78, Number 3, pp. 284-293.

Fergusson, J.E. and Ryan, D.E. (1984). The elemental composition of street dust from

large and small urban areas related to city type, source and particle size. Science of the Total Environment, 34: 101-116.

Glass, C. and Bissouma, S. (2005). Evaluation of Parking Lot Bioretention Cell for

Removal of Stormwater Pollutants. WIT Transactions on Ecology and the Environment, Vol. 81, 699 – 708.

Hsieh, C. and Davis, A. P. (2005a). Multiple-event Study of Bioretention for Treatment

of Urban Storm Water Runoff. Water Science & Technology, Vol. 51, No. 3 – 4, pp 177 – 181.

Hsieh, C. and Davis, A. P. (2005b). Evaluation and Optimization of Bioretention Media

for Treatment of Urban Storm Water Runoff. Journal of Environmental Engineering, November 2005.

- 41 -

Hsieh, C., Davis, A. P., Needelman, B. A. (2007). Bioretention Column Studies of Phosphorus Removal from Urban Stormwater Runoff. Water Environment Research, Vol. 79, No. 2, pp. 177 – 184.

Hoffman, E.J., Mills, G.L., Latimer, J.S., Quinn, J.G. (1984). Urban Runoff as a source

of polycyclic aromatic hydrocarbons to coastal waters. Environmental Science & Technology, Vol. 18, pp. 580 – 587.

Hoffman, E. J. (1985). Urban Runoff Pollutant Inputs to Narragansett Bay: Comparison

to Point Sources, p. 159 – 164. In Proceedings from the Conference on Perspectives on Nonpoint Source Pollution. Report No. 440/5-85-001. U.S. Environmental Protection Agency, Kansas City, Missouri.

Hvitvet-Jacobsen, T. and Yousef, Y.A. (1991). Road runoff quality, Environmental

Impacts and Control in Road Pollution. In: Hamilton, R.S. and Harrison, R.N. (eds.), Highway Pollution, p.165-209. London: Elsevier.

Illinois Department of Energy and Natural Resources (IL DENR) (1982). Nationwide Urban Runoff Project, Champaign, Illinois: Evaluation of the Effectiveness of Municipal Street Sweeping in the Control of Urban Storm Runoff Pollution. IL Environmental Protection Agency and U.S. Environmental Protection Agency, PB83-209890.

Kaushal, S. S., Groffman, P. M., Likens, G. E., Belt, K. T., Stack, W. P., Kelly, V. R., Band, L. E., and Fisher, G. T. (2005). Increased Salinizat