CRWR Online Report 05-10 
 
 
Particle Size Distribution of Highway Runoff and 
Modification Through Stormwater Treatment 
 
 
 
by 
 
Ana Marie Karamalegos, M.S.E. 
Michael E. Barrett, Ph.D. 
Desmond F. Lawler, Ph.D. 
Joseph F. Malina, Jr., Ph.D. 
 
 
December 2005 
 
 
 
 
Center for Research in Water Resources 
The University of Texas at Austin 
J.J. Pickle Research Campus 
Austin, TX 78712-4497 
 
 
This document is available online via the World Wide Web at 
http://www.crwr.utexas.edu/online.shtml 
Acknowledgements 
 
This research was funded by the Texas Department of Transportation (TxDOT) under 
grant number 0-4543, “Bridge Runoff Characterization”. The TxDOT Program 
Coordinator was David Stolpa, P.E., and the TxDOT Project Director was Melissa 
Gabriel.  The TxDOT Project Monitoring Committee consisted of David Zwerneman, 
Amy Ronnfeldt, Norm King, Kathleen Darnabry, and Mark Fisher. 
 
ii 
 
TABLE OF CONTENTS 
LIST OF TABLES...................................................................................................... V 
LIST OF FIGURES...................................................................................................VI 
CHAPTER 1    INTRODUCTION............................................................................. 1 
1.1 Overview...................................................................................................................1 
1.2 Regulatory Framework...........................................................................................1 
1.3 Project Objectives....................................................................................................2 
CHAPTER 2    LITERATURE REVIEW................................................................. 4 
2.1 Introduction .............................................................................................................4 
2.2 Particle Size Distribution ........................................................................................6 
2.3 Density of Particles in Stormwater Runoff............................................................9 
2.4 Particle Treatment.................................................................................................13 
2.5 Summary ................................................................................................................16 
CHAPTER 3    MATERIALS AND METHODS ................................................... 18 
3.1 Overview.................................................................................................................18 
3.2 Subdivision of Runoff Samples.............................................................................19 
3. 3 Solid Concentration Measurements....................................................................21 
3. 4 Coulter Counter Measurements..........................................................................25 
3.5 Density Calculations ..............................................................................................33 
CHAPTER 4    RESULTS AND ANALYSIS.......................................................... 34 
4.1 Suspended Sediment Concentration ....................................................................34 
iii 
4.2 Coulter Counter Measurements...........................................................................36 
4.3 Density Calculations ..............................................................................................38 
4.4 Relative Distribution by Mass ..............................................................................41 
4.5 Comparing Bridge Approach Highway Samples................................................42 
4.6 Before and After Treatment .................................................................................43 
4.7 Summary ................................................................................................................54 
CHAPTER 5 SUMMARY AND CONCLUSIONS ................................................ 56 
5.1 Conclusions.............................................................................................................56 
5.2 Recommendations for Future Research ..............................................................58 
APPENDIX A: SUSPENDED SEDIMENT CONCENTRATIONS AND        
RAINFALL DATA .......................................................................................... 59 
APPENDIX B: PARTICLE SIZE DISTRIBUTIONS OF THE RUNOFF       
SAMPLES ........................................................................................................ 64 
REFERENCES .......................................................................................................... 77 
 
iv 
 LIST OF TABLES 
 
Table 2.1 Summary of Reported Sediment Density Values..................................................... 12 
Table 3.1 Comparison of Known Concentrations of SIL-CO-SIL Suspensions and their 
Corresponding TSS and SSC............................................................................................ 22 
Table 3.2 Desired Count Range in Relationship to Aperture Size.......................................... 27 
Table 3.3 Comparing the Aperture Size Needed in Relationship to Particle Diameter........ 33 
Table 4.1 Suspended Sediment Concentration and Density of Particles less than 75 µm in 
Untreated Highway Runoff............................................................................................... 35 
Table 4.2 Average Particle Densities in Various Runoff Samples........................................... 39 
Table 4.3 SSCs for Samples Collected from the Storm Event on January 27, 2005 ............. 48 
Table 4.4 SSCs and VSS concentrations for the samples collected May 2005 ....................... 52 
Table 4.5 Comparison of BMP Treatment Efficiencies ........................................................... 53 
 
 
v 
                                         LIST OF FIGURES 
 
Figure 2.1 Typical Austin Sand Filter ....................................................................................... 16 
Figure 3.1 Map of Austin with Site Locations........................................................................... 19 
Figure 3.2 Photograph of a Dekaport Cone Sample Splitter................................................... 20 
Figure 3.3 Experimental Setup for Sieving the Samples.......................................................... 23 
Figure 3.4 Photograph of the Beckman Coulter Multisizer 3 ................................................. 26 
Figure 3.5 Electrolyte Filtering Set-up ...................................................................................... 29 
Figure 3.6 Original Particle Size Distribution Function for a SIL-CO-SIL Suspension ...... 30 
Figure 3.7 (A) Particle Size Distribution Function, (B) Number Distribution, and (C) 
Volume Distribution for the SIL-CO-SIL Suspension ................................................... 31 
Figure 4.1 (A) Particle Size Distribution Function and (B) Volume Distribution in the 
Bridge Approach Highway Runoff Sample from January 27, 2005 ............................. 37 
Figure 4.2 Comparison of the Relative Distribution by Mass ................................................. 41 
Figure 4.3 Particle Size Distribution Function of the Bridge Approach Highway Samples. 43 
Figure 4.4 (A) Particle Size Distribution Function and (B) Volume Distribution Before and 
After Treatment with an Austin Sand Filter for Samples Collected on January 27, 
2005 ..................................................................................................................................... 44 
Figure 4.5 (A) Particle Size Distribution Function and (B) Volume Distribution Before and 
After the Extended Detention Basin of Samples Collected on July 27 & 28, 2005 ...... 45 
Figure 4.6 (A) Particle Size Distribution Function and (B) Volume Distribution of the 
Samples Collected Along the Buffer Strip at Site 1 on January 27, 2005..................... 47 
Figure 4.7 (A) Particle Size Distribution Function and (B) Volume Distribution of the 
Samples Collected Along the Buffer Strip at Site 2 on May 28 and 29, 2005............... 49 
Figure 4.8 Rainfall Rate of the Storm Events on January 27
th
 and May 28
th
 of 2005........... 51 
 
 
vi 
CHAPTER 1    INTRODUCTION 
 
1.1 Overview 
A significant concern to regulatory agencies and to professionals in the environmental 
field is nonpoint source pollution.   This type of pollution includes the direct or scattered 
sources of pollution that enter a water system through runoff from agricultural fields as 
well as urban areas (USGS, 2005).  The continuing development of urban activities, i.e., 
construction and traffic flow, has increased nonpoint source pollution, which promotes 
the degradation of the quality of the receiving water.  The pollutants from these activities 
then adsorb onto the surface of the fine particles in the runoff.  Reducing the amount of 
fine particles in runoff through various Best Management Practices (BMPs) before the 
stormwater reaches the receiving water is a typical method for decreasing the adverse 
impact on water bodies due to nonpoint source pollution.  The treatment capabilities of 
BMPs can be improved by learning about the particle size distribution and particle 
density of these fine particles.           
 
1.2 Regulatory Framework 
Federal and state regulations dictate the handling of stormwater.  The Federal Water 
Pollution Control Act, also known as the Clean Water Act, has significant influence on 
various water quality concerns.  The United States Environmental Protection Agency 
(USEPA) created the act in 1972 that was amended in 1977.  This act formulated the 
regulatory structure for protecting the surface waters of the United States from pollutant 
discharge.  The regulations put in place water quality standards for all surface waters and 
1 
required permits to discharge pollutants from point sources.  Two sections of the 
Amendments to the Federal Water Pollution Control Act in 1977 are particularly relevant 
to highway runoff water quality issues: 303(d) and 404.  Section 303(d) of the Clean 
Water Act declares that, every other year, each state needs to submit to the EPA a list of 
water bodies within its jurisdiction that do not meet their designated use and/or that are 
impaired by contaminants.  EPA and United States Army Corps of Engineers have 
jurisdiction over section 404 of the Clean Water Act.  This section states that an 
individual, agency, or company must obtain a permit before placing fill materials in the 
water bodies (streams, ponds, lakes, and wetlands).  The permit is needed for the 
construction of roads and to lay pipes as well as the development of residential, 
commercial, and recreational sites.  The purpose of the permit is to balance the protection 
of the water bodies with the need to use filling materials.  There are three types of 404 
permits: Nationwide Permit, Regional General Permit, and Individual Permit.   
 
Investigating how pollutants are transported in highway stormwater runoff and how 
efficiently various Best Management Practices (BMPs) operate in treating runoff is 
crucial to protect ground and surface water.   
 
1.3 Project Objectives 
The objective of this project was a documentation of the size distribution and density of 
particles in stormwater runoff.  The objective was achieved through the following 
approach: 
 
• Operating monitoring devices to collect samples of runoff from bridge 
approach highway and bridge deck as well as collecting runoff samples 
2 
from various BMPs (Austin sand filter, extended detention pond, and 
vegetated filter strip) 
   
• Developing and implementing a reliable method for correctly 
characterizing highway stormwater runoff 
 
• Calculating average density of particles in the stormwater runoff samples, 
and comparing the calculated densities to the density of sand 
 
• Analyzing the particle removal efficiency for the stormwater BMP 
treatment processes  
 
Chapter 2 examines the results of previous research dealing with particle size distribution 
and with the density of particles in stormwater.  The technique and process used in this 
research to characterize stormwater runoff are described in Chapter 3, and Chapter 4 
discusses the analysis of collected data.  Lastly, Chapter 5 presents conclusions and an 
overall discussion of the key findings.   
 
3 
CHAPTER 2    LITERATURE REVIEW 
 
2.1 Introduction  
Stormwater runoff is a major contributor to nonpoint source pollution that leads to the 
deterioration of receiving waters.  The particles in runoff and soluble contaminants that 
adsorb to surfaces of particulate material enter the water body.  Urban expansion may 
cause new highway development and consequent increased nonpoint source pollution 
from runoff.  Thus, urban development has led to concerns about potential declines in 
water quality and impacts to the health of residents (Barrett et al., 1998).  Therefore, 
research focused on the characteristics of particles in stormwater runoff as well as in the 
effluent of different types of Best Management Practices (BMPs) used to treat runoff is 
important to reduce the adverse effects of nonpoint source pollution.  
 
Storm events mobilize particles and enable them to be washed into receiving waters.  The 
density and the size distribution of particles affect the transport of the solids and 
associated pollutants (Characklis and Wiesner, 1997).  Larger particles in stormwater 
runoff settle out, but smaller particles remain suspended in stormwater runoff and travel 
greater distances.  In addition, decreasing particle size often is correlated with increasing 
surface area of the particles, allowing for more adsorption of dissolved constituents onto 
the surface of the particles.  Thus, examining the particle size distribution of the particles 
in runoff aids the understanding of the transportation time of pollutants, the magnitude of 
the area affected, and the treatability of these particles.  The properties of the particles 
also influence the type of treatment process that is appropriate for removal. 
 
The pollutants on roads and, therefore, also in highway runoff is harmful to receiving 
waters, so it is important to know the primary sources of this nonpoint source pollution.  
Young et al. (1996) enumerated these primary sources of contaminants.  A significant 
4 
amount of the particulate matter in runoff stems from the wear and tear of the pavement 
and vehicles both from normal traffic operations and from road maintenance.  Sediment 
disturbance and the atmosphere increase the amount of particulate matter, nitrogen, and 
phosphorus in highway runoff.   The use of fertilizer on highway right-of-ways also 
impacts nitrogen and phosphorous levels in runoff.  Lead and zinc found in runoff is 
generally produced by the following parts of automobiles: tire, lubricating oil and grease, 
and wheel bearings.  Copper constituents in runoff generally are generated from two 
sources: automobiles (metal plating, bearing wear, engine parts, and brake lining wear) 
and chemical substances (fungicides and insecticides).   
 
These harmful pollutants tend to adsorb onto fine sediment in highway runoff; therefore, 
further information about fine particles must be collected to improve the control measures 
and to advance the treatment and disposal of highway runoff.  Sansalone and Buchberger 
(1997) stated that the pollutant loads of zinc, lead, and copper with a diameter smaller 
than 100 µm for the road runoff samples attributed to more than 50% of the cumulative 
pollutant loads, but only 10% of the total weight (Furumai et al., 2002; Sansalone and 
Buchberger, 1997).  Numerous early studies examined the menace of sewer overflows on 
receiving waters in terms of biochemical oxygen demand (BOD), total suspended solids 
(TSS), and total coliform counts, and other commonly studied runoff constituents are 
chemical oxygen demand (COD), phosphorus, nitrogen, and heavy metals (Palmer 1950, 
1963; Characklis and Wiesner, 1997).   However, published information about the 
pollutants adsorbed to the fine particles in highway runoff is limited, so to truly improve 
stormwater treatment processes of highway runoff, the characteristics and density of the 
fine particles must be explored and recorded.    
 
5 
2.2 Particle Size Distribution  
2.2.1 Methods to Determine Particle Size Distribution 
 
Past research has used several different methods to examine the particle size distribution 
of stormwater runoff.  One common technique used to analyze particle size distribution 
of runoff is a sieving method, where the sediment is sieved through different sieve sizes 
to characterize the size distribution.  Two methods of sieving exist: wet and dry.  The 
liquid runoff samples are poured directly onto the sieve for the wet sieve technique, and 
for the dry sieve process, only the sediment from the runoff samples are poured onto the 
sieve.  Both methods produce reliable particle size distribution results and are easy to 
perform on runoff samples.  However, each method has its own disadvantages.  Dry 
sieving can be time intensive if the liquid in the runoff sample must be evaporate to get 
the road sediment.  Some of the fine particles can clump together and act as a larger 
particle during the evaporation step.  Wet sieving, on the other hand, does not have this 
issue; however during the wet sieve process for the smaller sieve sizes, a mucous layer 
can form that clogs the sieve holes.  Both methods are typically used to characterize the 
size distribution of the coarse particles, but they can not accurately measure the size 
distribution of the fine particles.    
 
Some researchers use the sieving process for the larger particles and then use a machine 
to determine the size distribution of the fine particles since it is difficult to characterize 
the size distribution of the fine particles when using the sieving method.  Different types 
of machines can be used for this measurement.  Two types of instruments include an 
electrical sensing zone (ESZ) instrument and a Coulter LS 130.  The ESZ instrument 
determines the particle size distribution by measuring the voltage flux between to 
electrodes caused by the particles, and the Coulter LS 130 quantify the diffraction of a 
6 
parallel beam of a monochrome laser due to the particles (Andral et al., 1999).  More 
research should be conducted to evaluate these various instruments to establish the most 
accurate method to determining the size distribution of the fine particles.    
 
Furumai et al. (2002) used solely the dry sieving technique to characterize the particle 
size distribution of collected highway runoff at the inlet of a retention pond in 
Winterthur, Switzerland.  The particle size distribution was determined by sieving the 
sample first with a 2 mm mesh sieve to remove the large debris and then with sieve sizes 
ranging from 50 µm to 800 µm.  TSS was also measured for each sample.  The data 
collected supported the notion that most of the suspended solids (SS) concentration was 
associated with the coarser size particles.   
 
Characklis and Wiesner (1997), however, used the electrical sensing zone instrument to 
explore the particle size distribution of grab samples from the Brays Bayou both before 
and after storm events.  The Brays Bayou is located in the Houston metropolitan area, 
and the Fort Bend and Harris counties’ stormwater drains into it.  Their research 
supported the idea that SS concentrations of runoff increases after a storm event, and they 
also found an increase in number concentration of particles with a particle diameter 
below 2.5 µm.   
 
Andral et al. (1999) used both the sieving method along with the use of the Coulter LS 
130 to determine the particle size distribution for sediment captured in a catchment area 
off a motorway in Kerault Region, France.  The most notable finding from this research 
was that the particles with a diameter larger than 100 µm settle out of the suspension 
easily; however, the particles with a diameter less than 100 µm remain in suspension.  
Thus, the investigators concluded that particles with a diameter less than 50 µm must be 
studied in order to effectively treat runoff.   
 
7 
Furumai et al. (2002), Characklis and Wiesner (1997), and Andral et al. (1999) were all 
able to determine the size distribution of particles in runoff; however, they used different 
techniques.  Since different techniques can yield the same type of data results, it is 
necessary to investigate the advantages and disadvantages of each method.  This method 
comparison should be recorded to help guide future researchers create a protocol for 
accurately determining particle size distributions.   
 
2.2.2 Association of Pollutants with Particle Size 
 
Researchers have been examining which particle size range needs to be targeted when 
designing stormwater treatment systems since the 1970’s, so determining the particle size 
distribution of sediment in highway runoff will enable researchers to then go a step 
further by associating pollutants with particular particle sizes.  Numerous research studies 
have shown that the treatment systems must be able to effectively remove fine particles in 
runoff to significantly reduce the pollutant loads.  Research on this subject has supported 
the idea that the urban dust and dirt in the small particle size range is correlated with the 
higher concentrations of pollutants, i.e., heavy metals (Pitt and Amy, 1973; Woodward-
Clyde, 1994; Vaze and Chiew, 2004).  The data showed that 75% heavy metal found on 
road sediment were associated with a particle diameter below 500 µm, and approximately 
half of the heavy metal found on road sediment was adsorbed on particles with a diameter 
between 60 µm to 200 µm.  High concentrations of copper, zinc, and phosphorus were 
found on sediment with a particle diameter between 74 µm and 250 µm (Dempsey et al., 
1993; Vaze and Chiew, 2004).  Characklis and Wiesner (1997) found that particles with a 
particle diameter below 2.5 µm did not account for a large portion of the total mass; 
however, it impacted the total surface area, allowing for more pollutants to adsorb onto 
the surface of the particles.  Thus, the concentration of metal, zinc for example, in terms 
of total percent of stormwater solid mass increased as the particle size decreases.  Metals, 
like zinc and iron, may not be effectively treated with sedimentation, which only removes 
larger particles efficiently.   
8 
 
Vaze and Chiew (2004) concluded that treatment systems must reduce fine particles in 
runoff, but they stated that it is necessary to remove particles that are larger than 53 µm, 
but preferably 11 µm, in order to decrease the amount of particulate matter that has total 
phosphorus (TP) and total nitrogen (TN) adsorbed onto it.  This conclusion was based on 
their research project in Melbourne, Australia.  Dry solids as well as stormwater grab 
samples were collected from a street that had an average traffic volume of 
3,000 vehicles/day.  The perpendicular street had an average traffic volume of 
30,000 vehicles/day.  Once dry and stormwater samples were collected, their nutrient 
loads were analyzed.  More TN was adsorbed to the fine sediment in the stormwater 
samples compared to the wet sieved samples.  Twenty to fifty percent of TN was 
dissolved in the stormwater grab samples.  However, the amount of TP dissolved in the 
stormwater and wet sieved samples were similar.  More than 60% of the TP was attached 
to sediment with a diameter between 11 µm and 150 µm, and 40-50% of the attached TP 
was adsorbed onto particles with a diameter between 11 µm and 53 µm.  Similar to TP, 
the majority of the TN was attached to particles in the size range of 11 µm to 150 µm.  
However, only a small amount of TP and TN adsorbed onto particles with a diameter 
between 4.5 µm and 11 µm.  Thus, these investigators concluded that treatment facilities 
should be designed to remove particles with a diameter greater than 11 µm.  Overall, 
previous studies support the idea that it is crucial to remove the fine particles in 
stormwater runoff in order to reduce the pollutant concentrations.   
 
2.3 Density of Particles in Stormwater Runoff                                                                        
Knowing the density of particles in stormwater runoff is critical because the density 
impacts the water quality.  The particle density also influences the behavior in advective 
transport, sedimentation, filtration, coagulation/flocculation, and reentrainment.  
Frequently, the density of particles is considered to be equivalent to the density of sand, 
which is 2.65 g/cm
3
, and the actual particle density is seldom determined in terms of a 
9 
function of particle diameter (Allen, 1990; Cristina et al., 2001).  Many treatment 
designs, such as those for highway runoff settling basins, are developed by using the 
concept of minimum trapping efficiency.  This trapping efficiency is related to the 
settling velocities of the particles, which are strongly influenced by particle density 
(Cristina et al., 2001). 
 
Cristina et al. (2001) investigated the particle size distribution and density of 
anthropogenic particulate matter carried in highway snow and snowmelt in Cincinnati, 
Ohio.  The particle size distribution was determined by using the common method of dry 
sieving.    The sieve sizes ranged from 25-4750 µm.  A hydrometer was then employed to 
analyze the particles with a diameter less than 75 µm.  The density was measured using 
an inert gas pycnometer.  The inert gas was an ultra-high pure He.  The mean density 
values of the coarse and fine particulate matter were 2.86 g/cm
3
 and 2.75 g/cm
3
, 
respectively.   
 
Sansalone and Triboullard (1999) also studied sediment on Cincinnati roads using similar 
particle size distribution methods and density analysis as Cristina et al. (2001); however, 
their research examined the particulate matter accumulated on the pavement of highways 
instead of snowmelt runoff.  This particulate matter would be mobile once a storm event 
occurred.  The sediment was collected by using a conventional wet-dry vacuuming 
technique.  The large particles were easier recovered compared to the fine particles 
because the small particles got caught in the cracks and joints of the pavement.  The 
densities ranged from 2.70 to 3.01 g/cm
3
 for the sediment found for all gradations, and 
the larger densities were associated with the particles in the size range of 850 to 1400 µm.  
The data suggested that the fine particles, such as tire material, were deposited beyond 
the pavement and shoulder areas because the abraded tires possess a density between 1.5-
1.7 g/cm
3
 with a particle diameter less than 20 µm (Kobriger and Geinopolos, 1984; 
Sansalone and Triboullard, 1999).   Sansalone and Triboullard (1999) also stated that 
abraded vehicular matter has a larger range in density and particle diameter.  The density 
10 
and particle diameter range are 1.6-4.0 g/cm
3
 and 1-104 µm, respectively.  Thus, the 
densities of the road sediment seemed to be affected by the abraded vehicular matter 
when looking at the magnitude of the density values.   
 
Kobriger and Geinopolos (1984) stated that vehicular-infrastructure abrasion is the 
primary source of particulate matter, which was supported by Sansalone and Triboullard 
(1999).  Vehicular-infrastructure abrasion includes tire-pavement interaction as well as 
abrasion between metallic vehicular parts.  The abraded pavement accounts for 40-50% 
of the total particulate mass, and abraded tires account for 20-30% of the total particulate 
mass.  Thus, previous research implied that the particulate matter on highways was 
primarily inorganic due to the fact that organic materials have a lower density values 
compared to the values that were recorded.     
 
Exploring how the density of road sediment and snowmelt compares to the density of 
particles in stormwater runoff would be a significant advancement to the stormwater 
treatment field, since these data could improve the design of the treatment systems.   
Evaluating the data variance between the usage of an inert gas pycnometer and a Coulter 
Counter for density measurements would also add insight to the field.  Currently, a 
limited amount of data exists on the density of particles in stormwater runoff and how it 
varies between storm events.  A summary of the recorded density values of highway 
particulate matter is displayed in Table 2.1.  
 
11 
Table 2.1 Summary of Reported Sediment Density Values 
Density (g/cm
3
) Reference Location Source 
Cristina et al., 2001 Cincinnati, OH
Snowmelt: 
Coarse Fraction 2.86 
Cristina et al., 2001 Cincinnati, OH
Snowmelt: 
Fine Fraction 2.75 
Jacopin et al., 1999 
Bordeax, 
France 
Detention Basin: 
stormwater sewer 2.20 & 2.27 
Jacopin et al., 1999 
Bordeax, 
France 
Detention Basin: 
combined sewer 2.24 
Bachoc, 1992; 
Referenced by 
Jacopin et al., 1999 France Stormwater sewers 2.19-2.56 
Bulter et al., 1992 
London, 
England Street Surface Sediment 2.10-2.51 
Zanders, 2004 
Hamilton, 
New Zealand Road Sediment 2.14 (d
p
<32µm)
Zanders, 2004 
Hamilton, 
New Zealand Road Sediment 
2.15 
(32<d
p
<63µm) 
Arthur and Ashley, 
1998 
Dundee, 
Scotland 
Inorganic Particles: 
combined sewer 
1.000-1.998 
(9<d
p
<1100µm)
Fan et al., 2003 - Sewer Sediment 
2.4-2.6 
(40<d
p
<900µm)
Li et. al., in press, a 
Los Angeles, 
California Highway Runoff 
1.30-1.42
*
(d
p
<0.45µm) 
Allen, 1990; 
Cristina et al., 2001 - 
Sand (Typically 
assumed runoff density) 2.65 
Cristina et al., 2001 - 
Typical Highly Organic 
Solids 1.1 
*
 Wet Specific Gravity 
 
The data in the table illustrate that the particle density in runoff ranged from 1.00-2.86 
g/cm
3
.  The snowmelt density was slightly larger than the density of road sediment, street 
12 
surface sediment, and detention basin samples.  Also, the smaller particle range had a 
lower density compared to the samples that incorporated the small, medium, and large 
particle size ranges in the density measurements. 
 
2.4 Particle Treatment 
Particle treatment plays a vital role in removing pollutants from highway runoff.  Various 
Best Management Practices are commonly used to treat stormwater runoff; therefore, 
optimizing their treatment efficiency is crucial.  These treatment systems include 
detention basins, sand filters, ponds, wetlands, and vegetated controls.  Kang et al. (2005) 
stated that it is important that BMPs operate efficiently with minimal supervision.  BMPs 
must also be able to manage a variation in pollutant concentration and flow rate.  Rainfall 
characteristics, daily traffic flow, and period of dry days affect the characteristics of the 
runoff.  Details about extended detention ponds, vegetated filter strips, and Austin sand 
filters are presented in this section.  
 
2.4.1 Extended Detention Pond 
 
Extended detention ponds are a widely used type of Best Management Practice for 
pollution reduction.  They treat the stormwater by a sedimentation process; thus the 
particles settle to the bottom of the pond over time.  The detention pond temporarily 
holds the water for a storm event with a minimum detention time of 24 hr.  Longer 
detention times increase the amount of particulate matter that settles to the bed of the 
pond, which enhances the particle removal efficiency.  They are also easy to integrate 
into multi-chamber stormwater treatment systems (Connecticut Stormwater Quality 
Manual, 2004). 
 
Jacopin et al. (1999) examined the particle removal efficiencies of detention tanks in 
Bordeaux, France as a method to effectively treat runoff.  This research studied two 
13 
detention ponds: Perinot (underground, off-line, combined sewer system) and Bourgailh 
(grassed sides and bottom, separate stormwater sewer system).  The combined sewer 
system (Perinot) captured a larger amount of organic particulate matter compared to the 
separate stormwater sewer system (Bourgailh).  In addition, the percentage of fine 
particles increased when the basin was filled more frequently; however, the size 
distribution of the collected particles from the detention pond traps still varied.  The 
number percent of particles with a diameter below 100 µm varied between 58% and 91%, 
and the median diameter range was 22 µm to 75 µm.  The size of the particles was 
correlated to two factors: the characteristic of the runoff and the trap submersion depth 
when the basin was filled during a storm event.  The solids that settled out of the runoff 
formed a deposit layer on the bottom of the basin; fortunately, the migration of the 
pollutants, including hydrocarbons and heavy metals, into the soil was only limited to the 
superficial layer of soil (0 cm to 10 cm).  Jacopin et al. concluded that the detention 
basins have the potential to treat fine particles if they are operated and designed properly.   
 
2.4.2 Vegetated Filter Strips 
 
Vegetative controls are effective treatment processes for removing pollutants from 
highway runoff.  They are also adaptable to site conditions and are one of the least 
expensive techniques for handling highway runoff (Dorman et al., 1996).   One 
commonly used type of vegetative control is vegetated filter strips, also known as buffer 
strips.  The stormwater is treated as the filter strips receives the highway runoff as 
overland sheet flow (Barrett et al., 2004).  The primary mechanisms that vegetated filter 
strips use to treat highway runoff are sedimentation, adsorption, infiltration into the soil, 
and biological/chemical activity of the grass and soil media.  This type of BMP is 
designed with relatively smooth and dense vegetation areas with moderate slope, which is 
typically less than 5% (Young et al., 1996; Barrett et al., 2004).   
 
14 
Vegetated filter strips, like all BMPs, have both advantages and disadvantages.  Zanders 
(2005) stated that vegetated filter strips decreased the amount of pollutant accumulation 
that can lead to pollutant loads in stormwater compared to kerb and gutter systems 
(Zanders, 2005).  Barrett et al. (2004) also verified that this type of BMP was effective at 
pollutant removal for highway runoff.  They concluded that vegetated filter strips were 
consistently effective at decreasing the suspended solid and total metal concentrations in 
highway runoff.  This BMP also reduced the dissolved metal concentration when the 
edge of pavement concentration was high.  However, the buffer strips did not treat the 
nitrogen and phosphorus in the runoff, and the concentration of organic carbon and of 
dissolved solids actually increased as the stormwater flowed along the grassy strips.   If 
the vegetation coverage of the buffer strips was below 80%, the performance of the BMP 
reduced drastically (Barrett et al., 2004).  Another debatable issue deals with small 
particle removal efficiencies because fine particles tended to have densities below 2.2 
g/cm
3
, which reduces the predicted trapping efficiencies.  In addition, these fine sediment 
particles may enter the BMP through aerial deposition instead of stormwater runoff 
(Zanders, 2005).  Both the advantages and disadvantages of vegetated filter strips should 
be considered before implementing this type of BMP on site; therefore, gaining more 
information about buffer strips ability to treat fine particles in runoff allows the designer 
to make a more informed decision.   
 
2.4.3 Austin Sand Filters 
 
Austin sand filters are another example of a commonly used BMP to treat highway 
runoff.  This type of BMP consists of two to three different chambers.  The first chamber 
is a sedimentation basin, also known as an extended detention basin.  The heavy sediment 
settles out of the stormwater in this section.  The sedimentation basin is followed by a 
filtration chamber, where the stormwater percolates through a sand bed.  Next, the water 
flows through a discharge chamber to either a storm drainage system or surface water.  
Figure 2.1 displays the typical design of an Austin sand filter.   
15 
 
Figure 2.1 Typical Austin Sand Filter (Shaver, 1991; USEPA, 1999) 
 
Austin sand filters are beneficial because they can be installed in highly developed sites.  
They also effectively reduce sediment, biochemical oxygen demand, and fecal coliform 
bacteria concentrations (USEPA, 1999).  A study by Barrett (2003) showed that the 
effluent concentration of the discharge from the Austin sand filter was independent of the 
influent runoff concentration.  Therefore, the system consistently decreased the TSS of 
the runoff to 7.8 mg/L.   Unfortunately, Austin sand filters can be limited to climate 
condition because it has not been proven yet that this BMP can efficiently operate in cold 
weather where freezing may occur.  Austin sand filters can only handle small drainage 
area, and other BMPs may cost less than an Austin sand filter (USEPA, 1999).  The 
Austin sand filter can effectively treat stormwater runoff; however, it is important to 
weigh the advantages and disadvantages of each BMP when choosing which one to 
implement at a particular site.   
 
2.5 Summary 
Additional work needs to be conducted to gain information about the treatability of the 
fine particles in stormwater runoff by various types of Best Management Practices.  The 
16 
fine particles possess a large surface area, which enables pollutants to adsorb onto them.  
Learning more about the size distribution of particles in runoff helps researchers to 
understand the characteristics of highway runoff.  Thus, correlating size distribution of 
particles with suspended sediment concentration would lead to more information about 
the density of the particles in highway runoff.  Estimating the density of particles in 
highway stormwater runoff is crucial when designing and operating BMPs because most 
BMPs incorporate the sedimentation process, which is influenced by particle density.   
These density values need to be published and recorded to make advancements in 
designing stormwater treatment processes.   
17 
CHAPTER 3    MATERIALS AND METHODS 
 
3.1 Overview  
This study examined the characteristics of stormwater runoff.  The particle size 
distribution and suspended sediment concentration were measured for each collected 
runoff sample, and then the average particle density was calculated.  The runoff samples 
were collected from five sites throughout Austin, TX. 
 
The locations of all of the sites are shown in Figure 3.1.  Four of the sites were located on 
Loop 360, which is a 14-mile state highway located on the western side of Austin; it is 
also know as the Capital of Texas Highway.  It extends from the south side of Austin near 
the Barton Creek/Mopac area to the north side at Highway 183.  Two adjacent filter strip 
sites were located on the northbound shoulder at 1905 South Loop 360, which is north of 
the Loop 360 and Loop 1 intersection.   The bridge deck site was located one mile east of 
Loop 1 and west of South Lamar Blvd. on Loop 360.  The bridge approach highway site 
was the stretch of the Loop 360 highway from the bridge deck site to the peak of the 
South Lamar overpass.  A full sedimentation sand filter was in place at this site.  Another 
sand filter was located on Anderson Mill Road near Parmer Lane, but because of design 
and construction flaws it is operated as an extended detention basin.     
18 
 
Extended Detention Basin
Vegetated 
Filter Strip 
Sites
Bridge Deck &
Bridge Approach 
Highway Sites 
Figure 3.1 Map of Austin with Site Locations (Austin City Connection, 2005) 
 
3.2 Subdivision of Runoff Samples  
The stormwater samples were collected and brought to the laboratory within 24 hours 
after a storm event.   These samples were immediately placed into a 4
◦
C refrigerator until 
19 
they were subdivided.  Each sample was split by using a Dekaport Cone Sample Splitter 
(Rickly Hydrological Company, http://www.rickly.com/sai/dekaport.htm), which is 
displayed in Figure 3.2. 
 
 
Figure 3.2 Photograph of a Dekaport Cone Sample Splitter                                                    
(Rickly Hydrological Company, 2004) 
 
The sample splitter is designed to overcome the issue of dividing the main sample into 
representative sub-samples.  It is difficult to split the main sample because it contains 
sand, grit, and other large particles which are all highly settleable.  This splitter has a 
cone that is 26 in deep with a diameter of 4 in.  Filling the cone to the top with the water 
sample is important to maintain a substantial head above the exit, and it allows the head 
loss to be equal in all of the tubes, which enables the sample to be divided evenly.  The 
problem with the rapid settling particles is overcome when the entire sample is poured 
through the splitter.  It is vital to note that proper operation of the sample splitter is 
required so that the main sample is divided into representative sub-samples.   
 
The sample splitter was used to divide the main stormwater sample into ten sub-samples 
with relatively equal volumes and concentration of suspended particles.  If the sample 
was between 5 L and 10 L, each tube of the cone splitter was connected to a 1 L bottle; 
however, if the sample was less than 5 L, two adjacent of the tubes were connected to 
each of five 1 L bottles.  In the rare instance that the sample was larger than 10 L, the first 
20 
and last 5 L of the sample were poured through the cone splitter.  In all cases, two of the 
sub-divided samples were then poured back into the cone splitter with each tube 
connected to a 250 mL bottle.  Double splitting the samples was necessary because the 
SSC work needed to use runoff samples that were approximately 200 mL.  The sub-
divided samples were then placed back into the 4
◦
C refrigerator until they were needed 
for analysis. 
 
3. 3 Solid Concentration Measurements 
3.3.1 Comparing TSS and SSC 
 
The first step in creating an experimental protocol to calculate the particle density in 
runoff was to determine the best technique for measuring the suspended solids 
concentration: total suspended solids (TSS) or suspended sediment concentration (SSC).  
These two measures are performed quite similarly.  In both cases, a know volume of the 
sample is passed through a standard filter paper.  This filter is dried at 105
◦
C, brought to 
room temperature in a desiccator, and weighed both before and after the sample is 
filtered.  The difference in the weight of the two measurements is the mass of sediment 
captured on the filter.  The difference in the two measurements is in how the sample to be 
measured is obtained.  The sub-sample is taken from the suspension with a pipette and 
then is dispensed onto filter paper to gather the sediment for TSS (Standard Methods, 
1995).  For SSC, the whole sample is poured onto the filter paper to catch the sediment 
(Gray, 2000).  The reason that the SSC technique is considered by many to be the 
preferred method for runoff samples is that it is difficult to get a representative sample 
into a pipette.   The stormwater runoff samples have particles with high settling 
velocities.  These particles settle rapidly to the bottom of the container, so it is difficult 
for them to be pulled up by the pipette.     
 
21 
Experiments were performed using SIL-CO-SIL suspensions to establish which 
technique would best determine the sediment concentration of the stormwater runoff.  
SIL-CO-SIL is ground pure silica.  It is an inert, white, low moisture substance that is at 
least 99.4% SiO
2
.  SIL-CO-SIL 49 was used in these experiments, where the number 
means that approximately 98% by weight of the particles are less than 49 µm in diameter 
(U.S. Silica, 1998).  Known concentrations of SIL-CO-SIL suspensions were made; and 
then, both TSS and SSC measurements were made on these suspensions.  Table 3.1 
displays the TSS and SCC for the known concentrations of the SIL-CO-SIL suspensions.   
 
Table 3.1 Comparison of Known Concentrations of SIL-CO-SIL Suspensions and 
their Corresponding TSS and SSC  
Concentration 
(mg/L) 
SSC 
(mg/L) 
TSS 
(mg/L) 
500 494 283
*
100 79 36
*
*
 represents the average TSS  
 
The SSC for the 500 mg/L SIL-CO-SIL suspension was close to the actual concentration 
of the suspension.   However, the SSC for the 100 mg/L suspension was lower than the 
expected concentration of the suspension.  It is important to note that the 100 mg/L 
suspension was made by diluting a 500 mg/L suspension using a pipette method.  If the 
pipette method is biased against the large particles, which seems to be a valid conclusion, 
the actual concentration of the 100 mg/L SIL-CO-SIL suspension could have been lower 
than desired.  The TSSs were all lower than the actual concentration of the suspensions.  
The TSS results supported the notion that the pipette process, which was used in TSS, 
was partial against the large particles in a suspension because these particles had 
significant settling velocities.  It was difficult for these large particles to turn 180º to be 
pulled into the pipette.  The data collected in this experiment supported U.S. Geological 
Survey’s (USGS) belief that SSC is a more accurate measurement of suspended sediment 
concentration than TSS (Gray, 2000).  For this reason, the SSC technique was used to 
measure the suspended solids concentration in this research.  
22 
3.3.2 Sieving Sub-samples 
 
The runoff sub-samples were wet-sieved before SSC measurements were made to 
determine the mass of particles between certain diameter ranges.  For each runoff sample 
collected, one sub-sample was sieved through a 75 µm sieve, and another sub-sample was 
poured through a 125 µm sieve.  For some of the runoff samples, a sub-sample was also 
sieved through a 105 µm sieve, but the dry particle mass difference between the samples 
sieved through a 105 µm sieve and a 75 µm was small.  There also was only a small 
difference in the solids content of samples sieved through 105 µm and 125 µm sieves, so, 
the 105 µm sieve was eliminated from the process.  The wet-sieving set-up is illustrated 
in Figure 3.3. 
 
**
** 
* * * * 
* * 
Sample
Sieve
Graduated 
Cylinder
Funnel 
 
Figure 3.3 Experimental Setup for Sieving the Samples 
 
The sieve was placed on top of a plastic funnel, which was attached to a graduated 
cylinder.  The graduated cylinder was tested to make sure that its volume was accurate 
before it was used.  The sample was mixed in its container, and then the entire sample 
was poured onto the sieve.  The water/sediment suspension that passed through the sieve 
was captured in the graduated cylinder, where its volume was measured.  Once the 
volume was recorded, Millipore water was used to rinse any remaining particles out of 
the container onto the sieve.  Millipore water was also poured on the sieve to make sure 
23 
that none of the particles with a diameter smaller than the sieve size remained on the 
sieve.  The sample was then ready for the suspended solids work. 
 
3.3.3 SSC Measurements 
 
SSC measurements were made on the collected runoff samples.  The SSCs were 
determined for particles with a diameter less than 75 µm, less than 125 µm, and for all of 
the particles in sample.  The procedures followed were similar to the procedures 
described in ASTM’s Method D 3977-97 (ASTM, 2002): 
 
1. Place a vacuum filter apparatus on a flask with a side nozzle, which is attached to 
the laboratory vacuum by a plastic tube 
2. Insert a Whatman 934-AH microfiber filter paper into the vacuum filter apparatus 
3. Rinse the filter paper with approximately 200 mL of Millipore water to remove 
dissolved solids on the paper; discard the water 
4. Place the filter paper in an aluminum weighing container and set it in a desiccator 
5. Transfer the filter paper and container to a 105
◦
C oven for 60 min (For volatile 
suspended solids (VSS), ignite for 15 mins in a 550
◦
C oven) 
6. Set the filter paper and container in a desiccator for 30 min, while the items cool 
to room temperature 
7. Measure the mass of the filter paper and the container 
8. Reinsert the clean filter paper into the vacuum apparatus 
9. Shake and pour an entire runoff sample with a known volume onto that filter 
paper 
10. Rinse the sample container with Millipore water and pour the water onto the filter 
paper  
11. Place the filter paper and collected sediment back onto the aluminum container in 
a desiccator 
24 
12. Put the filter paper/container back in a 105
◦
C oven for 60 min (For VSS, ignite for 
20 mins in a 550
◦
C oven) 
13. Remove these items and set them in a desiccator for 30 min 
14. Weigh the items 
 
The initial mass of the filter paper, the mass of the filter paper/sediment, and the volume 
of the sample were needed to solve for SSC.  Equation 3.1 is the formula used to solve 
for this value. 
 
() ()[]
()mLV
L
mL
g
mg
gXgX
L
mg
SSC
s
fprfp ⎥
⎦
⎤
⎢
⎣
⎡
⎟
⎠
⎞
⎜
⎝
⎛
∗
⎥
⎦
⎤
⎢
⎣
⎡
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
∗−
=
⎟
⎠
⎞
⎜
⎝
⎛
+
33
1010
        Equation 3.1 
 
In the equation, X
fp+r
 is the mass of the filter paper plus residue, X
fp
 is the initial mass of 
the filter paper, and V
s
 is the volume of the sub-sample. 
 
3. 4 Coulter Counter Measurements 
3.4.1 Coulter Counter Description 
 
A Beckman Coulter Multisizer 3 was used to measure the particle size and the particle 
count distributions of the runoff samples for this research; a picture of the instrument is in 
Figure 3.4, along with a schematic of the central measuring zone.   This technology is 
based on the electrical sensing zone principle (ESZ).  A steady electrical current is 
maintained between two electrodes, one inside an aperture tube and the other immersed 
in the sample being measured.  The suspension is pulled through an aperture at a steady 
rate.  As each particle is pulled through the small aperture, it increases the resistance 
between the electrodes.  This resistance changes the voltage between the electrodes.  The 
voltage fluctuation is proportional to the volume of the particle.  The sorting of the 
25 
voltage pulses into size bins yields the number concentration of particles in a large 
number of size increments, i.e., the particle size distribution.  The Beckman Coulter 
Multisizer 3 can size and count particles in the size range of 0.4 to 1200 µm.   
 
 
Figure 3.4 Photograph of the Beckman Coulter Multisizer 3 
 
The Coulter Counter is operated through a PC by running the Beckman Coulter’s 
Multisizer 3 software.  The raw data were displayed immediately on the computer screen 
since the technology used digital pulse measurements.  The data were saved to a disk for 
further processing.  The collected data were then exported to an Excel file to convert the 
results into the number distribution, volume distribution, and particle size distribution 
function.  Refer to the Beckman Coulter’s Multisizer 3 operation’s manual for more 
detailed information about the Coulter Counter (Multisizer 3 Operator’s Manual, 2002). 
 
 
26 
3.4.2 Sample Preparation for the Coulter Counter 
 
Diluting the runoff sample with an electrolyte solution was necessary when measuring 
the particles in a runoff sample on the Coulter Counter.  The Coulter Counter could not 
measure the sample directly because the stormwater samples were too concentrated for 
the machine.  At the same time, it is necessary to have enough of the stormwater sample 
measured by the Coulter Counter because high conductance solution is needed for the 
electrical sensing to work.  Thus, mixing the appropriate proportion of sample and 
electrolyte solution to get a particle count in the count range for the aperture tube used 
was essential.  The desired count range for each aperture size is shown in Table 3.2.   
 
Table 3.2 Desired Count Range in Relationship to Aperture Size 
Particle Size
*
Aperture Size 
(µm) (µm) 
Desired Count Range 
 (#/µL) 
30 d
p
 < 30 µm 200 - 360 
100 d
p
 < 100 µm 55 - 80 
400 d
p
 < 400 µm 0.15 - 0.23 
 
*
 It is important to note that the each aperture can technically have particles with a diameter below the 
aperture size pass through the aperture hole; however, if the particle diameter is above 75 µm, the settling 
velocity will be so great that the particle will settle to the bottom of the container instead of passing through 
the aperture. 
 
Making sure that the mixture of electrolyte solution and runoff sample was within the 
desired count range was an integral part of the Coulter Counter work.   If the suspension 
was too concentrated, the Coulter Counter could have counted several little particles as 
one large particle or the aperture hole would have clogged.  If there were too few 
particles, the certainty that the data were a representative measurement of all of the 
particles in the runoff sample was low.  Typically, for the 30 µm and 100 µm apertures, 
40 mL of electrolyte solution was mixed with 0.6-15 mL of runoff sample, which 
depended on the concentration of the stormwater sample.  The 30 µm aperture needed 
27 
more of the runoff sample compared to the 100 µm aperture.  The 400 µm aperture was 
generally run with a suspension of 80-100 mL of electrolyte solution and 1-20 mL of 
runoff sample.  Compared to the other aperture sizes, the 400 µm aperture consumed a 
significant amount of the suspension.  Since the 400 µm aperture drew in such a large 
volume of suspension, it used the vacuum meter to pull in the suspension, instead of the 
meter pump that the 30 µm and 100 µm apertures used.  The 400 µm aperture was also 
more sensitive to substantial count jumps.  These jumps come from a large clump 
temporally clogging the aperture, and this clump creates a lot of false electrical signals.  
When this occurs, the data are not reflective of the real particles going through the 
aperture, so the data are flawed.  If the values had sudden, large count peaks, the run was 
redone.   
 
The electrolyte solution used for experimentation was either 2% NaCl
2
 + 0.1% NaN
3
 or 
0.01 M CaCl
2
 + 2% NaCl
2
 + 0.05% NaN
3
, which was dependent on the needs of the other 
researchers in the laboratory.  The electrolyte solution was filtered successively through 
0.22 µm and 0.05 µm filter papers continuously for 48 hours before it was used for the 
Coulter Counter work.  Figure 3.5 shows the apparatus set-up for the filtering of the 
electrolyte solution. 
 
28 
 
Figure 3.5 Electrolyte Filtering Set-up 
 
This filtering setup has a pump which is on the top shelf.  The pump pulls electrolyte 
solution from the bottom 4 L jar and pushes it through the filters which discharge to the 
top 4 L jar.  When the top jar reaches a certain height, a siphon is automatically started, 
and the collected electrolyte solution is drained to the bottom jar. 
3.4.2 Particle Size Distributions 
 
The data for each run from the Coulter Counter were imported into a calibrated excel 
spreadsheet for the particular electrolyte solution being used.  The data from the three 
aperture sizes were placed in the same excel file so that the data overlaid each other.  
Figure 3.6 displays the particle size distribution function for a 100 mg/L SIL-CO-SIL 
suspension.   
29 
0
1
2
3
4
5
6
7
-0.5 0 0.5 1 1.5 2
30 μm aperture
100 μm aperture
400 μm aperture
Pa
r
t
ic
l
e
 Size Di
st
r
i
b
ut
i
on
 
Fu
ncti
on
log
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
cm
-3
μ
m
-1
) 
log d
p
 (d
p
 in μm)
   
  Figure 3.6 Original Particle Size Distribution Function for a                                      
SIL-CO-SIL Suspension 
 
These data were then cleaned up so that there was a smooth transition in particle sizes 
between the different aperture sizes.  It was important that the data did not overlap so that 
double counting of particles would not occur.   If this occurred, the volume concentration 
determined, which was used to calculate the average particle density, would be larger 
than the actual volume concentration.  The volume concentration was calculated by 
summing the volume concentrations for each aperture size.  The area under the volume 
distribution curve was also equal to the cumulative volume concentration for the 
suspension.  Figure 3.7 shows the particle size distribution function, number distribution, 
and volume distribution for the SIL-CO-SIL suspension after the data were adjusted to 
prohibit overlapping.   
30 
0
1 10
7
2 10
7
3 10
7
-0.5 0 0.5 1 1.5 2
Vo
lum
e
 Di
st
r
i
but
io
n
Δ
V/
Δ
 l
og d
p
 (
V
 
in
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
0
1
2
3
4
5
6
7
30 μm aperture
100 μm aperture
400 μm aperture
Pa
r
t
ic
le Si
z
e
 D
i
str
i
b
u
t
i
on 
Fu
nct
i
o
n
lo
g
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
c
m
-3
μ
m
-1
) 
0
1 10
6
2 10
6
3 10
6
4 10
6
Nu
m
b
e
r
 Dis
tr
i
b
u
ti
o
n
Δ
N/
Δ 
log
 d
p
 (
N
 
in 
cm
-3
) 
 
A.
B.
C.
      
    Figure 3.7 (A) Particle Size Distribution Function, (B) Number Distribution, and 
(C) Volume Distribution for the SIL-CO-SIL Suspension 
 
31 
The particle size distributions were all normalized with a logarithmic scale to give better 
resolution to the data.  Figure 3.7 A is the particle size distribution function.  The y-
coordinate for this figure is the logarithm of the arithmetic change in number of particles 
for a small increment divided by the arithmetic change in diameter size.  This distribution 
is typically used for flocculated suspension (Lawler, 2004).  Figure 3.7 B is the number 
distribution.  It divides the change in number of particles in a small increment that is 
correlated to the logarithmic increment in particle diameter.  The area under the curve 
yields the number concentration.  This distribution is also useful because it visually 
displays the size range that contains the most particles and shows the spread of the 
particle size distribution (Lawler, 2004).  Figure 3.7 C is the volume distribution.  This 
distribution graphs the volume concentration for a small increment divided by the 
logarithmic increment in particle diameter.  The total area under the curve is equaled to 
the total volume concentration, and this value is needed to solve for the particle density of 
the runoff.    
 
Figure 3.7 also shows that each aperture size measures a different size range of particles.  
The 30 µm aperture could not count particles with a large diameter because these 
particles could have not fit through the aperture hole.  In addition, the 400 µm aperture 
was not as sensitive to the smaller particles that could have been measured by the 30 µm 
aperture.  Table 3.3 displays the typical size range of the particles measured by each 
aperture size. 
 
32 
Table 3.3 Comparing the Aperture Size Needed in Relationship to Particle Diameter 
Aperture Size (µm) Particle Diameter Range (µm) 
30 0.8 – 2 
100 2 – 10 
400 10 – 75 
 
3.5 Density Calculations 
The density of the particles in the runoff samples was calculated after the SSCs and 
volume concentrations from the Coulter Counter work were collected.  The SSCs used 
were the SSC measurements for the samples that were sieved through the 75 µm sieve, 
because it was assumed that the Coulter Counter was only counting the particles that 
were smaller than 75 µm.  The particles that were larger than 75 µm had such a high 
settling velocity that they fell to the bottom of the container before they were pulled 
through the aperture tube.  The average particle density was calculated using Equation 3-
2. 
 
   
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
⎟
⎠
⎞
⎜
⎝
⎛
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
==
⎟
⎠
⎞
⎜
⎝
⎛
g
mg
1000
L
mL
1000
mL
m
,
cm
m
10)mg/L,(
cm
g
Density 
3
3
3
12
3
μ
μ
ρ
p
p
V
SSC
  Equation 3.2 
 
The densities for the collected samples were then compared to the density of sand, to the 
density of highly organic materials, and to the densities found in the literature review to 
add insight about the runoff characteristics.       
33 
CHAPTER 4 RESULTS AND ANALYSIS 
The data collected in this research are presented and interpreted in this chapter.  The 
suspended sediment concentrations, particle size distributions, and densities of the 
particles in the runoff samples are compared for the different storm events and site 
locations.  The treatment capability of Austin sand filters and of vegetated filter strips are 
discussed in the light of the data.    
 
4.1 Suspended Sediment Concentration  
Suspended sediment concentration measurements were taken for all of the gathered 
runoff samples.  In preparation for this measurement, the collected samples were 
subdivided.  Sub-samples from each storm event were sieved through different sieve 
sizes.  Initially, the samples were sieved with a 75 µm and 105 µm sieves, but after 
analysis of the SSCs for the first few storm events, the future samples were also sieved 
with a 125 µm sieve in order to better characterize the particle size distribution.  
However, the mass difference between the samples that were sieved through a 105 µm 
sieve and the other two sieve sizes were small, so the 105 µm sieve size was removed 
from the sieving process.  Table 4.1 illustrates the SSCs for the untreated runoff at each 
location: bridge approach highway, bridge, vegetated filter strip, and extended detention 
basin.  
   
34 
Table 4.1 Suspended Sediment Concentration and Density of Particles less than 75 
µm in Untreated Highway Runoff 
  
SSC (mg/L) % of Total Mass 
   
Date Location
*
Total 
d
p
 
<125µm 
d
p
 
<105µm 
d
p
 
<75µm 
d
p 
<125µm 
d
p
 
<105µm 
d
p
 
<75µm 
11/1/2004 Bridge 26 - 12 11 - 45 43 
11/15/2004 Bridge 27 - 17 13 - 63 48 
1/27/2005 Bridge 76 - 32 30 - 42 39 
1/4/2005 BHA 104 - 87 77 - 84 74 
1/27/2005 BHA 702 - 268 249 - 38 35 
3/26/2005 BHA 2803 141 - 129 5 - 5 
5/8/2005 BHA 1007 129 - 111 13 - 11 
1/27/2005 VFS1 289 - - 176 - - 61 
5/28/2005 VFS2 193 113 - 76 59 - 39 
7/27/2005 EDB 114 58 - 42 51 - 37 
8/5/2005 EDB 605 28 - 16 5 - 3 
8/8/2005 EDB 60 26 - 15 43 - 25 
* 
BHA = Bridge Approach Highway; VFS 1 = Vegetated Filter Strip Site 1; VFS 2 = Vegetated Filter Strip      
Site 2; EDB = Extended Detention Basin 
 
The suspended sediment concentration for a runoff sample was dependent on the 
particular storm event; in addition, the percent of total mass under a certain particle size 
varied for each storm, suggesting that the intensity of the storm caused differences in the 
relative size distributions.  The percent of total mass with a particle diameter below 125 
µm ranged from 5% to approximately 59%, and the percent of total mass with a particle 
diameter below 75 µm ranged from 5% to 61%.  For the two samples collected on 
January 27, 2005, the bridge deck sample had a far lower SSC compared to the bridge 
approach highway sample.  This occurrence could have been due to the fact that the 
bridge approach highway runoff flowed a greater distance along the road, which enabled 
it to pick up more sediment, compared to the bridge deck runoff.   
 
35 
While the data set is small, it appears that storm events that cause very high total SSCs 
have only a small fraction of the mass in the small particles.  Even though the total 
suspended sediment concentrations was dramatically different for the bridge approach 
highway samples collected on March 26
th
 and May 8
th
 of 2005, a similar amount of mass 
was caught between the 75 µm and 125 µm sieves for these samples. 
 
The total SSC for the extended detention basin also varies significantly.  The largest SSC 
was 605 mg/L from the storm event on August 8, 2005, and the lowest SSC was 60 mg/L 
from the storm event on August 5, 2005.  The SSC for the particles with a diameter 
below 75 µm were relatively similar, even though the total SSCs varied notably.  The 
storm event on July 27, 2005 has a total SSC of 114 mg/L; however, it had a much larger 
SSC for the particles with a diameter below 75 µm, which was 42 mg/L, compared to the 
other extended detention basin runoff samples. 
 
4.2 Coulter Counter Measurements  
The Coulter Counter was used to find the size distribution of the particles in the runoff 
samples.  Figure 4.1 displays the particle size distribution function and volume 
distribution of the particles in the bridge approach highway sample collected on January 
27, 2005.  The other individual particle size distribution functions and volume 
distributions for the bridge approach highway and bridge deck samples are shown in 
Appendix B.  They are also combined and discussed subsequently in the chapter.   
36 
0
1 10
8
2 10
8
3 10
8
-0.5 0 0.5 1 1.5 2
30 um aperture
100 um aperture
400 um aperture
Vo
l
u
m
e
 D
i
s
t
ri
bu
ti
o
n
Δ
V/
Δ
 l
og 
d
p
 (V i
n
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
0
1
2
3
4
5
6
7
30 um aperture
100 um aperture
400 um aperture
Pa
r
t
i
c
l
e
 Size 
Di
s
t
r
i
b
u
t
i
on 
F
u
nction
log
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
cm
-3
μ
m
-1
) 
30 μm aperture
100 μm aperture
400 μm aperture
30 μm aperture
10  pert
400 μm aperture
A.
B.
 
Figure 4.1 (A) Particle Size Distribution Function and (B) Volume Distribution in 
the Bridge Approach Highway Runoff Sample from January 27, 2005 
 
The graph shows that the 30 µm aperture measured the particles with a diameter between 
0.8 µm to 3 µm (-0.1 < log d
p
 < 0.5).  The 100 µm aperture assessed the particles with a 
diameter between 3 µm to 16 µm (0.5 < log d
p
 < 1.2), and the 400 µm aperture analyzed 
the particles with a diameter between 16 µm to 75 µm (1.2 < log d
p
 < 1.9).  The particle 
size distribution function, shown in Figure 4.1 A, illustrated that, on a number basis, 
more particles were in the size range measured by the 30 µm aperture compared to the 
37 
400 µm aperture.  However, the volume distribution (Figure 4.1 B) showed that the 
majority of the total particle volume concentration came from the large particles, and the 
small particles only contributed a small amount to the overall volume.  When plotted as 
shown, the area under the volume distribution is the total particle volume concentration; 
this value was used in the density calculations.   
 
4.3 Density Calculations 
The average density of the particles in the stormwater runoff samples was calculated 
from the Coulter Counter and SSC data.  The bridge approach highway sample from 
January 27, 2005 is analyzed in detail here to explain the density calculations.   The SSC 
of the sample was 249 mg/L, and the volume concentration of the sample was 
1.46*10
8
 µm
3
/mL, or 
V
36
3
312
3
3
3
6
ppm 146
m10
m
146
m10
cm
cm
mL
mL
m
10146 ==
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
⎟
⎠
⎞
⎜
⎝
⎛
∗
μ
μ
μ
μ
.     
 
3
3
8
3
3
12
cm
g
71.1
g
mg
1000
L
mL
1000
mL
m
1046.1
cm
m
10
L
mg
249
Density =
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
⎟
⎠
⎞
⎜
⎝
⎛
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
∗
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
⎟
⎠
⎞
⎜
⎝
⎛
=
μ
μ
        Equation 4.1  
 
These values yielded a particle density of 1.71 g/cm
3
.  Table 4.2 displays the SSCs, 
volume concentrations, and densities for the particles in the bridge approach highway, 
bridge, and vegetated filter strip samples. 
 
38 
Table 4.2 Average Particle Densities in Various Runoff Samples 
SSC 
(mg/L) 
Volume 
(ppm
v
)
b
Location
a
Date 
Density 
(g/cm
3
) 
1/27/2005 Bridge Deck 30 23 1.31 
1/4/2005 BHA 77 67 1.17 
1/27/2005 BHA 249 164 1.71 
3/26/2005 BHA 129 64 2.00 
5/8/2005 BHA 111 46 2.42 
1/27/2005 VFS1 0m 176 125 1.41 
5/28/2005 VFS2 0m 76 64 1.19 
5/28/2005 VFS2 2m 56 73 0.76 
5/28/2005 VFS2 4m 30 45 0.67 
5/28/2005 VFS2 8m 22 27 0.81 
7/27/2005 EDB inlet 42 15 2.80 
8/5/2005 EDB inlet 16 12 1.33 
8/8/2005 EDB inlet 15 11 1.36 
a
 BHA= Bridge Approach Highway; VFS1 = Vegetated Filter Strip Site 1; VFS2 = Vegetated Filter Strip            
Site 2; EDB = Extended Detention Basin                                                                                                          
b
 1 ppm
v
 = 10
6
 µm
3
/mL 
 
These densities are the densities of the particles with a diameter less than 75 µm.  The 
particles with a diameter larger than 75 µm were not measured by the Coulter Counter 
due to their high settling velocities.  These particles dropped to the bottom of the 
container, which made it impossible for them to be pulled through the aperture tube.  The 
calculated density of the particles in the runoff samples ranged from 0.81 g/cm
3
 
to 2.80 
g/cm
3
.  Clearly, the density of the particles is not likely to be below 1 g/cm
3
.  The low 
densities could have been measured lower than in reality due to the fact that the density 
calculation, which includes two mass measurements and three Coulter Counter 
measurements (three separate apertures), had compounded error.  Many investigations 
have assumed that the density of the particles in stormwater runoff is that of sand, 
39 
2.65 g/cm
3
; however, almost all of the calculated average densities were below the 
density of sand.   
 
Since typical organic materials possess a density of approximately 1.1 g/cm
3
, the 
presence of organic material may have impacted the density of the particles in the runoff 
samples, which caused the average particle density of the samples to be lower than the 
density of sand (Cristina et al., 2001).  Thus, the low densities found for the bridge deck 
site, for vegetated filter strip sites, and for two of three inlet samples at the extended 
detention basin site suggested that these samples contained a significant amount of 
organic material.  The vegetated filter strip samples from site 2 had visible grass and 
insects in the runoff, which is organic material.  The runoff samples could have also 
collected rubber from tires and asphalt from the road.  These materials have densities of 
1.5-1.7 g/cm
3
, which would lower the average particle density of the runoff (Kobriger 
and Geinopolos, 1984; Sansalone and Triboullard, 1999).  The calculated densities still 
supported the idea that the particles in highway runoff do not have a density equivalent to 
sand even though error existed in the values.   
 
The bridge approach highway sample collected on 1/4/05 had a noticeably lower density 
compared to the other bridge approach highway samples; thus it is possible that the level 
of organic material in the runoff may vary notably between storm events.  The bridge 
approach highway samples collected on 3/26/05 and 5/8/05 had a density in a similar 
range compared to the literature review values for detention basins, stormwater sewers, 
and street surface sediment.  The larger densities seemed to be correlated with the larger 
SSCs, which suggested that the fraction of inorganic material such as sand was higher for 
these samples.  The runoff sample from the inlet to the extended detention basin on July 
27, 2005 had the highest average particle density because its SSC was relatively high 
compared to its small volume concentration; it seems possible that the SSC is erroneously 
40 
high.  To reiterate the most significant finding, almost all of the stormwater runoff 
samples collected had an average particle density below the density of sand.   
 
4.4 Relative Distribution by Mass 
The relative distribution by mass was calculated from the Coulter Counter data and SSCs.  
The procedure was to convert the volume distribution to a cumulative basis and then 
scale it according to the relative mass in the SSC
75µm
 versus SSC
total
 samples.  Figure 4.2 
shows the relative distribution by mass for the bridge approach highway and bridge deck 
samples collected on January 27, 2005.   
0
0.1
0.2
0.3
0.4
0 1020304050607080
Highway Approach (SSC=249mg/L)
Bridge (SSC=30mg/L)
R
e
l
a
t
i
ve
 D
i
st
r
i
but
i
on 
by M
a
ss 
(
%
 
F
i
ner
)
Particle Diameter (μm)
 
Figure 4.2 Comparison of the Relative Distribution by Mass 
 
The figure illustrates that only 35-40% of the total mass in the stormwater runoff samples 
were measured by the Coulter Counter.  Even though the SSCs were significantly 
different for the bridge deck and bridge approach highway samples, the fraction of 
suspended particles measured by the Coulter Counter was comparable.  However, the 
41 
bridge deck sample had a slightly larger percentage of its particles by mass measured by 
the Coulter Counter (d
p 
< 75 µm).  The relative distribution by mass must be bimodal, 
since the curves leveled out around 40% instead of elevating towards 100% in Figure 4.2.  
Another significant increase in mass must occur at a larger particle size so that the 
relative distribution by mass reaches 100%.   
 
4.5 Comparing Bridge Approach Highway Samples 
The SSCs and particle size distributions for the bridge approach highway samples 
collected on four different dates were compared.  The SSCs varied from 77 mg/L to 
249 mg/L, which was a notable difference between storms.  The particle size distribution 
functions of these samples are displayed in Figure 4.3.  The sample collected on January 
4, 2005, which also had the lowest SSC, contained more small particles (d
p 
< 1.8 µm or 
log d
p
 < 0.25) compared to the other samples.  Also, the sample collected on January 27, 
2005, which had the highest SSC, had more particles in the mid-size range (1.8 µm < d
p 
< 
25 µm or 0.25 < log d
p
 < 1.4), comparatively.  All of the samples had a similar particle 
size distributions for the large particles (25 µm < d
p 
< 75 µm or 1.4 < log d
p
 < 1.9).  Thus, 
it seems that the small particles did not affect the SSCs; however, the particles with a 
diameter in the mid-size range, as well as those larger than 75 µm, were primarily 
responsible for the differences in the SSCs.  
 
 
42 
0
1
2
3
4
5
6
7
-.500.511.52
1/4/05 (SSC=77mg/L)
1/27/05 (SSC=249mg/L)
3/26/05 (SSC=129mg/L)
5/8/05 (SSC=111mg/L)
Pa
r
t
ic
l
e
 Size Di
s
t
r
i
b
u
t
i
on 
Fu
nc
t
i
on
lo
g
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
c
m
-3
μ
m
-1
) 
log d
p
 (d
p
 in μm)
         
Figure 4.3 Particle Size Distribution Function of the                                                   
Bridge Approach Highway Samples 
 
4.6 Before and After Treatment 
4.6.1 Full Sedimentation Sand Filter  
 
Two samples were gathered at the bridge approach highway site on January 27, 2005: 
bridge approach highway runoff and effluent of the full sedimentation sand filter.  This 
type of full sedimentation sand filter is commonly referred to as an Austin sand filter.  It 
is an extended detention basin followed by a sand bed filter.  The particle size distribution 
functions and volume distributions for these coupled samples are displayed in Figure 4.4.  
The SSC for the effluent sample was drastically lower than the SSC for the bridge 
approach highway runoff.  The effluent sample’s suspended sediment concentration was 
too small to measure for the sieved sample, and the sample that was not sieved had a SSC 
of 2.9 mg/L.  The SSC for the bridge approach highway sample that was sieved with a 75 
µm sieve was 249 mg/L, and the total SSC was 702 mg/L.  Between the inflow and 
43 
outflow samples, the particle size function distributions shows an efficient removal of 
particles with a diameter larger than 1 µm (log d
p
 > 0), and the volume distributions 
illustrates a significant decrease in volume for the particles with a diameter larger than 
2.5 µm (log d
p
 > 0.4).  The information gathered supported the notion that the full 
sedimentation sand filter treatment process dramatically improved the bridge approach 
highway stormwater runoff. 
 
       
Figure 4.4 (A) Particle Size Distribution Function and (B) Volume Distribution 
Before and After Treatment with an Austin Sand Filter for Samples Collected on 
January 27, 2005 
0
1 10
8
2 10
8
3 10
8
-0.5 0 0.5 1 1.5 2
Vo
lum
e
 D
i
s
t
r
i
but
i
o
n
Δ
V/
Δ
 l
og 
d
p
 (V
 
i
n
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
B.
0
1
2
3
4
5
6
7
Highway Approach
BMP
Pa
r
t
ic
l
e
 S
i
ze
 Di
str
i
b
u
t
i
on 
Fu
nct
i
o
n
log
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in 
c
m
-3
μ
m
-1
) 
SSC=249 mg/L
SSC=2.9 mg/L
A.
44 
4.6.2 Extended Detention Basin  
 
The Anderson Mill Basin site has an extended detention basin installed to treat the runoff.  
Three runoff samples were gathered at the inlet and outlet of the extended detention basin 
as well as three grab samples of the water within the basin.  Figure 4.5 displays the 
inflow and outflow particle size distributions for the samples collected from the storm 
event on July 27 and 28, 2005.  The particle size distributions for the other samples are 
located in Appendix B and show similar trends. 
0
1 10
7
2 10
7
3 10
7
-0.5 0 0.5 1 1.5 2
Vo
l
u
me
 Di
s
t
r
i
butio
n
Δ
V/
Δ
 l
og 
d
p
 (
V
 in
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
0
1
2
3
4
5
6
7
7-27-05 Inflow
7-28-05 Outflow
P
a
r
t
ic
l
e
 Size Di
s
t
r
ib
ut
i
on 
Fu
nction
log
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
cm
-3
μ
m
-1
) 
SCC = 4 mg/L 
VSS = 2 mg/L 
SSC = 42 mg/L 
VSS = 7 mg/L
A.
B.
 
Figure 4.5 (A) Particle Size Distribution Function and (B) Volume Distribution Before and 
After the Extended Detention Basin of Samples Collected on July 27 & 28, 2005 
 
45 
Figure 4.5 shows a reduction in particles with a diameter larger than 3 µm (log d
p
 > 0.5) 
between the inlet and outlet of the BMP for this particular storm event; all the data shown 
are samples that had been sieved through the 75 µm sieve.  The particle volume 
concentration decreased between the inlet and outlet samples as well; from a value of 
15 ppm
v
 for the inlet to 3 ppm
v
 for the outlet, for particles with a d
p
 < 75 µm (log d
p
 < 
1.9).  The SSC decreased from 42 mg/L to 4 mg/L, and the VSS reduced from 7 mg/L to 
2 mg/L by treatment.  A significant portion of the suspended sediment in the outflow 
sample was volatile.  Unfortunately, the density of the particles in the outflow sample 
could not be calculated because the suspended sediment concentration was too low.  A 
slight error in the SSC could drastically affect the calculated density.  Refer to Appendix 
A for additional SSC and VSS data.   
 
This research supported the treatment process theory that a sand filter basin plus an 
extended detention basin has a better treatment capability compared to solely an extended 
detention basin.  The extended detention basin alone does not reduce the particle volume 
as efficiently as the full sedimentation sand filter that was discussed in the prior section.  
It also did not remove particles with a diameter between 1 µm and 3 µm while the Austin 
sand filter did reduce those particles; however, this detention basin did remove larger 
particles quite well.  The excellent removal of large particles could be due in part to the 
nature of the storm.  The system seems to work well when all of the runoff in the basin 
can sit undisturbed for several hours so that batch sedimentation is efficient.  The 
duration of the storm event on July 27, 2005 was 3 hours, and the total rainfall was 1.02 
in.  The short storm event allowed for effective particle removal due to batch 
sedimentation.  A storm with a longer duration might not work as well because the runoff 
will be going into the basin at the same time it is leaving.  The outflow valve could be 
shut to hold the runoff in the basin longer to maximize the treatment efficiency.   
 
 
46 
4.6.3 Vegetated Filter Strip  
 
Highway stormwater runoff samples were collected at the vegetated filter strip site 1 on 
January 27, 2005.  This storm event has 1.5 inches of rainfall over approximately 16 
hours (Wunderground, 2005).  The collected samples were analyzed with the Coulter 
Counter to examine the alteration in particle size distribution from 0 m to 8 m across the 
grassy shoulder, which can be seen in Figure 4.6.  
 
         
Figure 4.6 (A) Particle Size Distribution Function and (B) Volume Distribution of 
the Samples Collected Along the Buffer Strip at Site 1 on January 27, 2005 
0
1
2
3
4
5
6
7
0 m
2 m
4 m
8 m
P
a
r
t
ic
l
e
 Size 
Di
s
t
r
ib
ut
i
on 
Fu
nction
log
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
cm
-3
μ
m
-1
) 
0
1 10
8
2 10
8
3 10
8
-0.5 0 0.5 1 1.5 2
Vo
lume
 Di
str
ibutio
n
Δ
V/
Δ
 log
 d
p
 (V
 i
n
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
A.
B.
 
47 
Figure 4.6 demonstrates that most of the particles in the highway runoff were removed 
between 0 m and 2 m along the vegetated filter strip.  A significant volume concentration 
reduction occurred for the particles with a diameter larger than 8 µm (log d
p
 > 0.9) from 0 
m to 2 m.  All four samples had nearly identical size distributions for particles less than 2 
µm (log d
p
 < 0.3), although the 2 m samples seemed to have the fewest particles.  For 
large particles (log d
p
 > 0.5, or d
p
 > 3.2 µm), the three samples at 2 m, 4 m, and 8 m were 
quite similar, though the 4 m sample had the fewest particles in that range.  It is possible 
that particles in this size range were scoured by the flowing water between 4 m and 8 m 
to increase the concentration as shown.  Overall, the grassy shoulder showed effective 
removal of particles with a diameter larger than 2.5 µm, and the majority of the particle 
treatment happened between 0 m and 2 m along the buffer strip.  Refer to Table 4.3 to 
look at the correlated SSCs.    
 
Table 4.3 SSCs for Samples Collected from the Storm Event on January 27, 2005 
Location Sieve Size SSC  (mg/L)
0m 75 µm 176 
0m None 285 
0m None 292 
2m None 20 
2m None 17 
4m None 14 
4m None 16 
8m None 4 
8m None 6 
 
Table 4.3 shows that a significant reduction in SSC occurred between 0 m and 2 m along 
the grassy shoulder.  The total SSCs for the particles in the runoff samples continued to 
decrease down the filter strip; however the decrease in concentration was not as 
pronounced as it was from 0 to 2 m.  Thus, the SSCs supported the Coulter Counter work 
and the conclusion that notable particle removal occurs between 0 m and 2 m along the 
buffer strip.    
 
48 
Samples were also collected at the second vegetated filter strip site after a storm event on 
May 28 and 29 of 2005.  The particle size distributions of the samples collected at 0 m 
and 8 m along the vegetated filter strip are illustrated in Figure 4.7.  Refer to Appendix B 
to see the particle size distributions of the samples collected at 2 m and 4 m along the 
buffer strip.  The particle size distributions for 0 m, 2 m, 4 m, and 8 m were relatively 
similar.   
 
 
           
Figure 4.7 (A) Particle Size Distribution Function and (B) Volume Distribution of 
the Samples Collected Along the Buffer Strip at Site 2 on May 28 and 29, 2005 
 
0
1 10
8
2 10
8
3 10
8
-.500.511.52
Vo
lume Di
st
r
i
but
io
n
Δ
V/
Δ
 l
og 
d
p
 (
V
 in
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
0
1
2
3
4
5
6
7
0m (SSC= 76mg/L)
8m (SSC= 22mg/L)
P
a
r
t
i
c
l
e
 Size
 
Di
s
t
r
i
b
ut
i
on 
Fu
nc
tion
log
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
c
m
-3
μ
m
-1
) 
49 
Figure 4.7 suggests an increase in particles with a diameter below 5 µm (log d
p
 < 0.7) 
from 0 m to 8 m along the vegetated filter strip, and the volume concentration also 
increased for this size range.  However, the particles with a diameter between 5 µm and 
75 µm (0.7 < log d
p
 < 1.9) decrease both in number and volume concentration from 0 m 
to 8 m down the buffer strip.  It is possible that the Coulter Counter sample, which was 
only a few milliliters of the whole sample, was not representative of the overall 0 m 
sample.  The particle size distributions of the samples shows that the vegetated filter strip 
was not effective at treating all of the particles with a diameter below 75 µm; however, 
the SSCs reduced from 76 mg/L to 22 mg/L from 0 m to 8 m down the grassy shoulder.  
It seemed as if the heavier particles with a diameter below 75 µm settled-out of the 
suspension, but the runoff collected more particles that had low densities and small 
diameters. 
   
As mentioned in the density section of this chapter, compounded error most likely 
occurred with either the volume concentration or SSC measurements which led to a low 
calculated density for the sample at 8 m.  The SSC for the 8 m sieved sample was only 
29% of the SSC of the 0 m sieved sample.  Since the density of organic material is 
approximately 1.1 g/cm
3
, the SSC of the 8 m sieved sample would have had to been 30 
mg/L to get that density, using the same volume concentration (27 ppm
v
).  If the SSC was 
30 mg/L, the 8 m sieved sample would have been 40% of the SSC at 0 m.  The volume 
concentration data support the notion that the SSC could have been lower than the actual 
concentration because the volume concentration of the 8 m sieved sample was 42% of the 
volume concentration for the 0 m sieved sample. 
 
The total SSCs for the 0 m and 8 m samples were 193 mg/L and 255 mg/L, respectfully, 
so it appears that the total SSC of the samples increased along the buffer strip.  As the 
runoff progressed along the vegetated filter strip, the runoff collected particulate matter.  
This occurrence could have been due to high velocity flow or flooding.  The National 
Weather Service recorded that Austin had approximately 1.5 inches of rainfall for this 
50 
storm event, and the storm event lasted for about a day and a half with several peaks in 
rainfall (NWSF, 2005; Wunderground, 2005), as shown in Figure 4.8.  A significant 
amount of rain for one event or long storm duration could reduce this BMP’s efficiency; 
also, a sudden downpour would produce high velocity runoff and decrease the 
effectiveness of the BMP compared to a long slow rain event.   
0
0.2
0.4
0.6
0.8
1
1-27-05
12243648
5-28-05
Ra
in
f
a
ll
  
R
a
t
e
 (
in
/h
r
)
Time (hr)
 
Figure 4.8 Rainfall Rate of the Storm Events on January 27
th
 and May 28
th
 of 2005 
 
 
Figure 4.8 illustrates that the storm event on May 28, 2005 had periods with high 
intensity rainfall.  The storm event on January 27, 2005 had a more constant rainfall 
amount that did not reach the same intensity as the storm on May.  The duration of the 
storm event in January was also much shorter than the storm event in May.  Thus, the 
intensity and duration of the storm seemed to impair the vegetated filter strips capability 
to remove the particles in the runoff. 
 
Seasonal change, in addition to possible flooding, affected the BMP’s treatment 
efficiency because the samples collected at 2 m, 4 m, and 8 m down the shoulder 
possessed grassy seeds and other vegetation.  To examine the seasonal influence on the 
51 
samples, the volatile suspended solid (VSS) concentrations were measured on these 
samples.  Table 4.4 displays the SSC and VSS concentrations of the collected samples.  
 
Table 4.4 SSCs and VSS concentrations for the samples collected May 2005 
Sieve 
Size 
SSC 
(mg/L) 
VSS 
(mg/L) Location 
0 m None 193
*
48 
8 m None 255
*
50 
0 m 75 µm 76
*
15 
8 m 75 µm 22
*
8 
*
 represents the average SSC  
 
The VSS concentrations reduced from 15 mg/L to 8 mg/L along the grassy shoulder for 
the 75 µm sieved samples.  However, a tiny change in measured weight of the residue 
and/or filter paper could significantly affect the VSS concentration, so the actual VSS 
concentration for the 8 m sieved sample could have been closer to 15 mg/L.  The 8 m 
sieved sample had a larger ratio of the volatile suspended solids to the total suspended 
solids compared to the 0 m sieved sample, but the compounded error in the value should 
still be taken into consideration.  Organic vegetated matter swept into the runoff impacted 
the suspended solid concentration and the particle size distribution of the samples 
collected along the grassy shoulder. 
 
A possibility also exists that the difference in SSCs at 0 m and 8 m is an accurate 
representation of the values at those locations.  Other outside factors, such as fire ant 
mounds near the collection pipes or build up of debris in the pipes at the beginning of the 
storm event due to irregular maintenance, should not be completely eliminated.  The 
phenomenon was only observed once and only at one location. 
 
In conclusion, the SSCs suggested that the vegetated filter strips effectively removed the 
heavy particles with a diameter below 75 µm in the highway runoff through settlement; 
however, the particle size distributions showed that this treatment method was not 
consistently effective at removing the less dense particles.  The total SSCs increased, so 
52 
the BMP was not efficient at removing the total amount of suspended solids in the runoff.  
Runoff treatment efficiencies may vary between storm events, and this occurrence can be 
caused by storm intensity or seasonal change. 
 
4.6.4 Comparison of BMP Outflow samples 
 
All of the BMPs examined did treat the highway runoff; however, their optimal capability 
to treat the stormwater runoff varied.  Table 4.5 illustrates the percent removal of SSC 
and volume concentration by the different BMPs. 
     
Table 4.5 Comparison of BMP Treatment Efficiencies 
 VFS1 VFS2 ASF EDB EDB EDB 
Storm Date 1/27/05 5/28/05 1/27/05 7/27/05 8/05/05 8/08/05 
SSC 
total in 
(mg/L) 288 193 702 114 605 60 
SSC 
total out 
(mg/L) 5 255 3 3 2 5 
% Removal 
SSC total
98% -32% >99% 97% >99% 92% 
SSC 
(dp<75µm) in 
(mg/L) 176 76 249 42 16 15 
SSC 
(dp<75µm) out 
(mg/L) - 22 - 2 1 4 
% Removal 
SSC (dp<75µm)
- 71% - 95% 94% 73% 
Volume 
(dp<75µm) in 
(ppm
v
) 125 64 146 15 12 11 
Volume 
(dp<75µm) out 
(ppm
v
) 7 27 4 3 2  8 
% Removal 
Volume (dp<75µm)
94% 58% 97% 80% 83% 27% 
VFS1 = Vegetated Filter Strip at Site 1; VFS2 = Vegetated Filter Strip at Site 2; ASF = Austin Sand Filter; 
EDB = Extended Detention Basin; Volume = Volume Concentration 
 
The BMP with the worst treatment efficiency was the vegetated filter strip at site 2.  
Although the vegetated filter strip at site 1 showed effectively particle treatment, site 
location and conditions should be examined closely when determining whether to 
implement a buffer strip because the seasonal change and storm intensity can notably 
impact this BMP’s treatment efficiency.  The extended detention basin reduced both the 
53 
SSC and volume concentration of the particles in the runoff.  The low percentage of 
volume concentration reduction is not due to this BMP’s inefficiency.  This value is low 
because the influent runoff had a significantly lower volume concentration of particles 
compared to the influent samples at the other sites.  The extended detention basin and 
Austin sand filter had similar total SSC out and volume concentration out (d
p
 < 75 µm) 
values, which were all lower than the values for the vegetated filter strip sites.  The 
Austin sand filter had high removal efficiencies, and in a previous section, the particle 
size distributions showed that this BMP treated particles with a diameter above 1 µm.  
The other BMPs could not treat particles with a diameter below 3 µm, which was also 
discussed in the previous sections.   
 
Overall, all of the BMPs work for runoff treatment, and more data should be collected to 
be able to differentiate more between the different types of BMPs.  After analyzing the 
limited data from this research, it was concluded that the Austin sand filter seemed to be 
the most effective BMP at removing the particles in runoff, and the vegetated filter strip 
BMP was the least efficient treatment process.  It is important that future work is 
conducted to support or disprove this conclusion since the results were data limited.  
Additionally, it is necessary to compare the price of construction and maintenance of 
each type BMP for a particular site when deciding which BMP to implement.    
 
4.7 Summary 
The most notable finding from this research was that the average particle density for 
almost all of the collected runoff samples was less than the density of sand.  BMPs are 
typically designed according to the surface overflow rate theory, and the associated 
54 
calculations are based on the particle density.  If a laboratory experiment is performed 
using particles with density of 2.65 g/cm
3
, the controlled treatment experiments could 
yield high efficiencies for particle removal.  This same treatment system placed in the 
field could yield lower treatment efficiencies if particle density for highway runoff was 
overestimated.   
 
This research also supports the idea that the Austin sand filter was the most efficient 
treatment process at removing particles from the highway runoff compared to the 
vegetated filter strips and extended detention basins, but more research should be 
conducted to support this conclusion.  This conclusion makes logical sense because an 
Austin sand filter is an extended detention basin followed by a filter.  It is necessarily 
better than a comparable extended detention basin alone.  Vegetated filter strips depend 
considerably on the design, i.e., slope and degree of vegetation, and could easily be 
overwhelmed in a large storm event.  However, research should be continued on this 
subject to analyze how the characteristics of the storm event influence these BMPs’ 
treatment capabilities.    
  
 
 
55 
CHAPTER 5 SUMMARY AND CONCLUSIONS   
 
The project objective of this research was an analysis of suspended sediment 
concentrations (SSCs) and particle size distributions of stormwater runoff samples in 
order to calculate and to document the average particle densities.  Runoff samples were 
also examined before and after three treatment systems.  These treatment processes 
included a full sedimentation sand filter, extended detention basin, and vegetated filter 
strip.  The particle size distribution is correlated to the surface area of the particulate 
matter in runoff.  Since pollutants sorb onto the particles, understanding the particle size 
distribution of runoff is important so that advancements can be made in designing 
treatment systems.  A limited amount of published articles exist that address the issue of 
particle size distribution and density of particles in stormwater runoff, so the key findings 
from this research added essential information to the field to improve the design of 
BMPs. 
 
5.1 Conclusions 
The evaluation of the data collected in this research led to the following conclusions: 
 
1. Suspended sediment concentration (SSC) was a better technique than total 
suspended solids (TSS) when measuring concentrations of particulate material in 
stormwater runoff because TSS measurements often fail to include large particles. 
  
2. SSCs in stormwater runoff varied significantly between storm events as well as 
location.  Mid-size particles (1.8 µm < d
p
 < 25 µm) affected the SSC, but small 
particles (d
p 
< 1.8 µm) did not have much impact.  Large concentrations of small 
particles contribute little to the particle volume (mass) in the overall sample.  
56 
3. Almost all of the densities of the particles in the collected runoff samples were 
less than the density of sand (ρ = 2.65 g/cm
3
). 
 
4. The full sedimentation sand filter effectively treated the stormwater runoff and 
removed particles with a diameter larger than 1 µm.  The extended detention 
basin was less effective at removing the smallest particles, but did provide 
substantial treatment of the runoff. 
 
5. The vegetated filter strip decreased the SSC and VSS of runoff as the stormwater 
progressed along the grassy shoulder.  This system, however, did not consistently 
reduce the volume concentration of the particles in the runoff, because the runoff 
could pick up organic particles.  Seasonal change and storm intensity also 
influenced treatment efficiency of this BMP.  
 
6. The full sedimentation sand filter was the most effective of the three BMPs 
studied at decreasing particles in runoff based on the limited data of this research. 
This result is consistent with treatment process theory.  However, additional data 
should be obtained to verify this conclusion. 
 
Laboratory experiments are often performed with particles having a density of sand, and 
designs of real systems often incorporate the assumption that the density of the particles 
in stormwater is equal to the density of sand.  The data collected in this research showed 
otherwise.  Thus, the theoretical particle removal efficiencies for the BMPs, which 
commonly include the sedimentation process, will be overestimated compared to the 
actual particle removal efficiencies observed in the field, if the particle density of runoff 
is assumed to be 2.65 g/cm
3
.  The designer must take several factors into consideration 
when implementing a BMP: particles in stormwater runoff are less dense than sand, 
organic material may be swept into the runoff, and biological growth may occur in the 
system.   
57 
5.2 Recommendations for Future Research 
Future work should focus on developing additional information about particle size 
distributions of stormwater runoff and the effectiveness of on-site treatment processes.  
Samples should be collected before and after additional on-site treatment processes to 
establish a data bank of efficiencies for removal of small particles by various types of 
BMPs.  Future studies also could examine the influences of storm size on the SSCs and 
the resulting average particle density of stormwater runoff.  In addition, the pollutants 
adsorbed to particles in different size ranges could be explored. 
 
 
 
58 
APPENDIX A: SUSPENDED SEDIMENT CONCENTRATIONS AND 
RAINFALL DATA 
 
LIST OF TABLES  
 
Table A-1 Comparison of SSCs for Bridge Approach Highway Samples................. 60 
Table A-2 Comparison of SSCs for Bridge Deck Samples.......................................... 60 
Table A-3 SSCs for Bridge Deck and Bridge Approach Highway Samples          
Collected from Storm Event on January 27, 2005.............................................. 61 
Table A-4 SSCs of the Grab Samples from Anderson Mill Basin.............................. 61 
Table A-5 SSCs and VSS concentrations for the Inflow and Outflow Samples 
Collected at Anderson Mill Basin......................................................................... 62 
Table A-6 Comparison of the Storm Duration and Rainfall Amount at the  
Anderson Mill Basin .............................................................................................. 62 
Table A-7 SSCs for the Samples Collected from the Vegetated Filter Strip at        
Site 2 from Storm Event on May 28 to 29 of 2005 .............................................. 63 
 
59 
Table A-1 Comparison of SSCs for Bridge Approach Highway Samples 
SSC  
(mg/L) Date Sieve Size
1/4/2005 None 116 
1/4/2005 None 93 
1/4/2005 105 µm 87 
1/4/2005 75 µm 77 
1/4/2005 75 µm 80 
3/26/2005 None 2803 
3/26/2005 125 µm 141 
3/26/2005 105 µm 141 
3/26/2005 75 µm 129 
5/8/2005 None 1007 
5/8/2005 125 µm 129 
5/8/2005 105 µm 126 
5/8/2005 75 µm 111 
 
 
 
Table A-2 Comparison of SSCs for Bridge Deck Samples 
SSC  
(mg/L) Date Sieve Size
11/15/2004 None 25 
11/15/2004 None 29 
11/15/2004 105 µm 18 
11/15/2004 105 µm 16 
11/15/2004 75 µm 13 
11/1/2004 None 26 
11/1/2004 None 25 
11/1/2004 105 µm 13 
11/1/2004 105 µm 10 
11/1/2004 75 µm 12 
11/1/2004 75 µm 10 
60 
Table A-3 SSCs for Bridge Deck and Bridge Approach Highway Samples Collected 
from Storm Event on January 27, 2005 
Location Sieve Size 
SSC  
(mg/L) 
Bridge Approach Highway: 
Coulter Counter Sample None 764 
Bridge Approach Highway None 702 
Bridge Approach Highway 105 µm 268 
Bridge Approach Highway 75µm 249 
Sand Filter Discharge None 3 
Bridge Deck:              
Coulter Counter Sample None 76 
Bridge Deck None 76 
Bridge Deck 105 µm 32 
Bridge Deck 75 µm 30 
Bridge Deck 75 µm 29 
 
Table A-4 SSCs of the Grab Samples from Anderson Mill Basin 
Date Sieve Size SSC  (mg/L) VSS  (mg/L) 
5/9/2005 None 7 1 
5/9/2005 125 µm 6 0 
5/9/2005 75 µm 4 1 
5/9/2005 75 µm 5 2 
5/28/2005 None 19 8 
5/28/2005 125 µm 8 3 
5/28/2005 75 µm 6 3 
5/28/2005 75 µm 5 2 
5/29/2005 None 3 0 
5/29/2005 125 µm 5 3 
5/29/2005 75 µm 2 1 
5/29/2005 75 µm 4 2 
  
61 
Table A-5 SSCs and VSS concentrations for the Inflow and Outflow Samples                                    
Collected at Anderson Mill Basin 
Coupled Storm  Date Location Sieve Size 
SSC  
(mg/L) 
VSS  
(mg/L) 
7/27/2005  Inflow None 114 29 
7/27/2005  Inflow 125 µm 58 10 
7/27/2005  Inflow 75 µm 42 7 
7/28/2005  Outflow None 6 3 
7/28/2005  Outflow 125 µm 5 2 
# 1 7/28/2005  Outflow 75 µm 4 2 
8/5/2005 Inflow None 605 32 
8/5/2005 Inflow 125 µm 28 8 
8/5/2005 Inflow 75 µm 16 1 
8/5/2005 Outflow None 2 2 
8/5/2005 Outflow 125 µm 2 2 
# 2 8/5/2005 Outflow 75 µm 1 0 
8/8/2005 Inflow None 60 18 
8/8/2005 Inflow 125 µm 26 5 
8/8/2005 Inflow 75 µm 15 5 
8/8/2005 Outflow None 5 2 
8/8/2005 Outflow 125 µm 5 2 
# 3 8/8/2005 Outflow 75 µm 4 2 
 
 
 
Table A-6 Comparison of the Storm Duration and Rainfall Amount at the Anderson 
Mill Basin 
Duration 
(hr) 
Rainfall Amount 
(in) Storm Date 
7/27/2005 3 1.02 
8/5/2005 16.5 0.41 
8/8/2005 15 1.41 
 
62 
Table A-7 SSCs for the Samples Collected from the Vegetated Filter Strip at                                     
Site 2 from Storm Event on May 28 to 29 of 2005 
Location 
SSC  
(mg/L) 
VSS 
(mg/L) Sieve Size 
0m None 182 48 
0m None 203 - 
2m None 193 - 
4m None 268 - 
8m None 273 50 
8m None 236 - 
0m 125 µm 113 - 
2m 125 µm 85 - 
4m 125 µm 55 - 
8m 125 µm 59 - 
0m 75 µm 78 - 
0m 75 µm 74 15 
2m 75 µm 56 - 
4m 75 µm 30 - 
8m 75 µm 22 - 
8m 75 µm 24 8 
8m 75 µm 20 8 
 
63 
APPENDIX B: PARTICLE SIZE DISTRIBUTIONS                                      
OF THE RUNOFF SAMPLES 
  
LIST OF FIGURES 
 
Figure B-1 (A) Particle Size Distribution Function and (B) Volume Distribution          
of Bridge Approach Highway Sample Collected on January 4, 2005............... 65 
Figure B-2 (A) Particle Size Distribution Function and (B) Volume Distribution        
of Bridge Approach Highway Sample Collected on March 26, 2005................ 66 
Figure B-3 (A) Particle Size Distribution Function and (B) Volume Distribution       
of Bridge Approach Highway Sample Collected on May 8, 2005. .................... 67 
Figure B-4 (A) Particle Size Distribution Function and (B) Volume Distribution          
of Bridge Deck Sample Collected on January 27, 2005...................................... 68 
Figure B-5 (A) Particle Size Distribution and (B) Volume Distribution of Bridge 
Approach      Highway Sample versus Bridge Deck Sample that was               
Collected on January 27, 2005.............................................................................. 69 
Figure B-6 (A) Particle Size Distribution Function and (B) Volume Distribution of 
Grab Samples from Anderson Mill Basin Collected on May 8, 28, & 29 of 2005
................................................................................................................................. 70 
Figure B-7 (A) Particle Size Distribution Function and (B) Volume Distribution            
of Anderson Mill Basin Inflow and Outflow Samples Collected on            
August 5, 2005 ........................................................................................................ 71 
Figure B-8 (A) Particle Size Distribution Function and (B) Volume Distribution of 
Anderson Mill Basin Inflow and Outflow Samples Collected on                     
August 8, 2005 ........................................................................................................ 72 
Figure B-9 (A) Particle Size Distribution Function and (B) Volume Distribution of 
Inflow       Samples at Anderson Mill Basin Collected on 7-27-05, 8-5-05,           
& 8-8-05 .................................................................................................................. 73 
Figure B-10 (A) Particle Size Distribution Function and (B) Volume Distribution of 
Outflow  Samples at Anderson Mill Basin Collected on 7-27-05, 8-5-05,               
& 8-8-05 .................................................................................................................. 74 
64 
Figure B-11 (A) Particle Size Distribution Function and (B) Volume Distribution         
of the Samples Collected on May 28 and 29 at the Second Vegetated Filter 
Strip Site ................................................................................................................. 75 
Figure B-12 (A) Particle Size Distribution Function and (B) Volume Distribution         
of Vegetated Filter Strip Samples at 2 m from Sites: 1 (Collected on 1-27-05) 
and 2 (Collected on 5-28-05) ................................................................................. 76 
 
 
 
  
Figure B-1 (A) Particle Size Distribution Function and (B) Volume Distribution of 
Bridge Approach Highway Sample Collected on January 4, 2005.                           
 
Average SSC was 79 mg/L. 
0
1 10
8
2 10
8
3 10
8
-0.5 0 0.5 1 1.5 2
Vo
lum
e
 D
i
st
r
i
but
i
o
n
Δ
V/
Δ
 l
og 
d
p
 (V
 
i
n
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
0
1
2
3
4
5
6
7
30 um aperture
100 um aperture
400 um aperture
Pa
r
t
i
c
l
e
 S
i
ze
 Di
str
i
b
uti
on 
Fu
nct
i
o
n
lo
g
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in 
c
m
-3
μ
m
-1
) 
A.
30 μm aperture
100 μm aperture
400 μm aperture
B.
65 
 
  
Figure B-2 (A) Particle Size Distribution Function and (B) Volume Distribution of 
               
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Bridge Approach Highway Sample Collected on March 26, 2005.                 
SSC was 129 mg/L. 
0
1 10
8
2 10
8
-.500.511.52
Vo
l
u
me
 D
i
s
t
ri
bu
ti
o
Δ
V/
Δ
 l
og 
d
p
 (V 
i
n
 
μ
m
3
log d
p
 (d
p
 in μm)
3 10
8
n
cm
-
3
) 
0
1
2
3
4
5
6
7
30 um aperture
100 um aperture
400 um aperture
Pa
r
t
i
c
le Si
z
e
 D
i
str
i
b
u
ti
on 
Fu
nct
i
on
lo
g
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
c
m
-3
μ
m
-1
) 
A.
B.
66 
0
1
2
3
4
5
6
7
30 um aperture
100 um aperture
400 um aperture
Pa
rt
i
c
le
 Siz
e
 
D
i
st
r
i
b
u
ti
o
n
 
Fu
nc
t
i
o
n
lo
g
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 c
m
-3
μ
m
-1
) 
0
1 10
8
2 10
8
3 10
8
-0.5 0 0.5 1 1.5 2
Vo
lum
e
 Dis
t
r
i
bu
t
i
o
n
Δ
V/
Δ
 lo
g
 d
p
 (
V
 
in
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
30 μm aperture
100 μm aperture
400 μm aperture
30 μm aperture
100 μm aperture
400 μm aperture
A.
B.
 
Figure B-3 (A) Particle Size Distribution Function and (B) Volume Distribution of 
Bridge Approach Highway Sample Collected on May 8, 2005.  SSC was 111 mg/L. 
 
67 
0
1
2
3
4
5
6
7
30 um aperture
100 um aperture
400 um aperture
Pa
r
t
ic
le Size
 Di
st
r
i
b
ut
i
on 
Fu
nc
tion
log
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
cm
-3
μ
m
-1
) 
0
1 10
8
2 10
8
3 10
8
-0.5 0 0.5 1 1.5 2
V
o
lu
m
e
 D
is
t
r
ib
u
t
io
n
Δ
V/
Δ
 lo
g
 
d
p
 (V i
n
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
30 μm aperture
100 μm aperture
400 μm aperture
A.
B.
  
Figure B-4 (A) Particle Size Distribution Function and (B) Volume Distribution of 
Bridge Deck Sample Collected on January 27, 2005.  SSC was 30 mg/L. 
 
 
68 
0
1
2
3
4
5
6
7
Bridge (SSC=30mg/L)
Highway Approach (SSC=249mg/L)
P
a
rt
ic
le
 S
iz
e
 D
i
s
t
r
ib
u
ti
o
n
 
F
u
n
c
t
io
n
lo
g
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
c
m
-3
μ
m
-1
) 
0
1 10
8
2 10
8
3 10
8
-.500.511.52
Vo
lu
m
e
 Dis
t
ri
bu
t
i
o
n
Δ
V/
Δ
 log
 
d
p
 (V i
n
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
A.
B.
  
Figure B-5 (A) Particle Size Distribution and (B) Volume Distribution of Bridge 
Approach Highway Sample versus Bridge Deck sample that was Collected on 
January 27, 2005 
 
 
69 
0
1
2
3
4
5
6
7
5/9/05
5/28/05
5/29/05
Pa
r
t
ic
le Size
 Di
st
r
i
b
u
t
i
on 
Fu
nct
i
on
log
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
cm
-3
μ
m
-1
) 
0
1 10
7
2 10
7
3 10
7
-0.5 0 0.5 1 1.5 2
V
o
lu
m
e
 Dis
t
r
i
b
u
t
io
n
Δ
V/
Δ
 l
og 
d
p
 (V i
n
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
A.
B.
  
Figure B-6 (A) Particle Size Distribution Function and (B) Volume Distribution of 
Grab Samples from Anderson Mill Basin Collected on May 8, 28, & 29 of 2005 
70 
0
1
2
3
4
5
6
7
8-5-05 Inflow
8-5-05 Outflow
P
a
r
t
i
c
l
e
 S
i
ze 
Di
st
rib
u
t
i
o
n
 
Fu
nction
lo
g
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
c
m
-3
μ
m
-1
) 
0
1 10
7
2 10
7
3 10
7
-.500.511.52
Vo
lu
m
e
 Di
st
r
i
b
u
t
io
n
Δ
V/
Δ
 l
og 
d
p
 (V
 
i
n
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
A.
B.
 
Figure B-7 (A) Particle Size Distribution Function and (B) Volume Distribution of 
Anderson Mill Basin Inflow and Outflow Samples Collected on August 5, 2005 
 
71 
         
Figure B-8 (A) Particle Size Distribution Function and (B) Volume Distribution of 
Anderson Mill Basin Inflow and Outflow Samples Collected on August 8, 2005 
0
1 10
7
2 10
7
3 10
7
-0.5 0 0.5 1 1.5 2
Vo
l
u
me Dist
r
i
but
io
n
Δ
V/
Δ
 l
og 
d
p
 (V i
n
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
0
1
2
3
4
5
6
7
8-8-05 Inflow
8-8-05 Outflow
Pa
r
t
ic
l
e
 Size Di
s
t
r
i
b
ut
i
on 
Fu
nc
t
i
on
log
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
cm
-3
μ
m
-1
) 
B.
A.
 
72 
0
1 10
7
2 10
7
3 10
7
-.500.511.52
Vo
l
u
me
 Di
st
ri
butio
n
Δ
V/
Δ
 lo
g d
p
 (V
 i
n
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
0
1
2
3
4
5
6
7
7-27-05 Inflow
8-8-05 Inflow
8-5-05 Inflow
Pa
r
t
ic
le Size 
Di
s
t
r
i
b
u
t
i
on 
Fu
nction
log
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
cm
-3
μ
m
-1
) 
B. 
A. 
 
Figure B-9 (A) Particle Size Distribution Function and (B) Volume Distribution of 
Inflow Samples at Anderson Mill Basin Collected on 7-27-05, 8-5-05, & 8-8-05 
73 
0
1
2
3
4
5
6
7
7-28-05 
8-5-05 
8-8-05 
Pa
r
t
i
c
l
e
 Si
ze Di
s
t
r
i
b
u
t
i
on 
Fu
nct
i
on
log
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
c
m
-3
μ
m
-1
) 
0
1 10
7
2 10
7
3 10
7
-0.5 0 0.5 1 1.5 2
Vo
lum
e
 Dist
r
i
but
i
o
n
Δ
V/
Δ
 lo
g
 
d
p
 (
V
 
in
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
A. 
B. 
 
Figure B-10 (A) Particle Size Distribution Function and (B) Volume Distribution of 
Outflow Samples at Anderson Mill Basin Collected on 7-27-05, 8-5-05, & 8-8-05 
 
74 
  
0
1
2
3
4
5
6
7
0m (SSC= 76mg/L)
2m (SSC= 56mg/L)
4m (SSC= 30mg/L)
8m (SSC= 22mg/L)
Pa
rt
ic
le
 Si
z
e
 D
i
s
t
r
i
b
u
t
i
o
n
 
F
u
nc
t
i
on
log
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in 
cm
-3
μ
m
-1
) 
0
1 10
8
2 10
8
3 10
8
-.500.511.52
Vo
lume
 Di
st
r
ibut
io
n
Δ
V/
Δ
 log 
d
p
 (
V
 
in
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
A.
B.
Figure B-11 (A) Particle Size Distribution Function and (B) Volume Distribution of 
the Samples Collected on May 28 and 29 at the Second Vegetated Filter Strip Site 
75 
0
1 10
8
2 10
8
3 10
8
-0.5 0 0.5 1 1.5 2
Vo
lum
e Dist
r
i
but
i
o
n
Δ
V/
Δ
 log
 
d
p
 (V
 
i
n
 
μ
m
3
cm
-
3
) 
log d
p
 (d
p
 in μm)
0
1
2
3
4
5
6
7
1-27-05
5-28-05
Pa
rt
ic
le
 Siz
e
 
D
i
s
t
r
i
b
u
t
i
on
 
Fu
nc
t
i
on
lo
g
(
Δ
N/
Δ
d
p
) (
Δ
N/
Δ
d
p
 in
 
cm
-3
μ
m
-1
) 
A.
B.
 
Figure B-12 (A) Particle Size Distribution Function and (B) Volume Distribution of 
Vegetated Filter Strip Samples at 2 m from Sites: 1 (Collected on 1-27-05) and 2 
(Collected on 5-28-05) 
 
 
76 
REFERENCES 
 
Allen, T. (1990). Particle Size Measurement. Chapman and Hall, London. 
 
American Society for Testing and Materials (ASTM). (2000 & Reapproved 2002). 
“Standard Test Methods for Determining Sediment Concentration in Water Samples: 
D3977-97.” vol. 11.02, Water (II), pp. 395-400. 
 
Andral, M. C., Roger, S., Montrejaud-Vignoles, M., and Herremans, L. (1999). “Particle 
Size Distribution and Hydrodynamic Characteristics of Solid Matter Carried by Runoff 
from Motorways.” Water Environment Research, 71 (4), 398-407. 
 
Arthur, S. and Ashley, R. (1998). “The Influence of Near Bed Solids Transport on First 
Foul Flush in Combined Sewers.” Water Science and Technology, 25(8), 1-12. 
 
Austin City Connection. (2005). “Austin Map.” Retrieved June 5, 2005, from 
http://www.ci.austin.tx.us/help/austinmap.htm
 
Barrett, M. E., Irish, L. B., Malina, J. F., and Charbeneau, R. J. (1998). “Characterization 
of Highway Runoff in Austin, Texas, Area.” Journal of Environmental Engineering, 
124(2), 131-137. 
 
Barrett, M., Lantin, A., and Austrheim-Smith, S. (2004) “Stormwater Pollutant Removal 
in Roadside Vegetated Buffer Strips.” Transportation Research Record No. 1890, 129-
140.  
 
Bachoc, A. (1992). La transfert des solides dans les reseaux d’assainissement unitaires. 
PhD thesis, INPT, Toulouse, 281 p. 
77 
Beckman Coulter’s Multisizer 3. (2002). “Multisizer 3 Operator’s Manual: PN 8321681 
Rev. B.” version 3.51, pp. 1-139. 
 
Bell, J. H., and Wanielista, M. P. (1979). “Use of Overland Flow in Storm-Water 
Management on Interstate Highways.” Transportation Research Record No. 736, 
13-21. 
 
Butler, D., Thedchanamoorthy, S., and Payne, J. A. (1992). “Aspects of surface sediment 
characteristics on an urban catchment in London.” Water Science and Technology, 
25(8), 13-19. 
 
Characklis, G. W. and Wiesner, M. R. (1997). “Particles, Metals, and Water Quality in 
Runoff from Large Urban Watershed.” Journal of Environmental Engineering, ASCE, 
123(8), 753-759. 
 
Clesceri, L. S., Greenberg, A. E., and Eaton, A. D. (1998). “2540 D. Total Suspended 
Solids Dried at 103 - 105
◦
C.”  Standard Methods for the Examination of Water and 
Wastewater (20
th
 Edition). Washington, DC: American Public Health Association, pp. 
2-57. 
 
Connecticut Department of Environmental Protection. (2004). “Dry Detention Ponds.” 
Connecticut Stormwater Quality Manual, pp. 11-S1-1 to 11-S1-4. 
 
Cristina, C., Tramonte, J., and Sansalone, J. (2002). “A Granulometry-Based Selection 
Methodology for Separation of Traffic-Generated Particles in Urban Highway 
Snowmelt Runoff.” Water, Air, and Soil Pollution, 136, 33-53. 
 
Dempsey, B. A., Tai, Y. L., and Harrison, S. G. (1993). ‘‘Mobilization and 
78 
Removal of Contaminants Associated with Urban Dust and Dirt.’’ Water Science and 
Technology, 28(3–5), 225–230. 
 
Dorman, M. E., Hartigan, J., Steg, R. F., and Quasebarth, T. F. (1996). “Retention, 
Detention and Overland Flow for Pollutant Removal from Highway Stormwater 
Runoff.” Volume I: Research Report, Federal Highway Administration Report FHWA-
RD-96-095. 
 
Fan, C., Field, R., and Lai, F. (2003). “Sewer-Sediment Control: Overview of an 
Environmental Protection Agency Wet-Weather Flow Research Program.” Journal of 
Hydraulic Engineering, ASCE, 129(4), 253-259. 
 
Furumai, H., Balmer, H., and Boller, M. (2002). “Dynamic Behavior of Suspended 
Pollutants and Particle Size Distribution in Highway Runoff.” Water Science and 
Technology, 46 (11-12), 413-418. 
 
Gray, J. R., Glysson, G. D., Turcious, L. M., and Schwarz, G. E. (2000). “Comparability 
of Suspended-Sediment Concentration and Total Suspended Solids Data.” Water-
Resources Investigations Report 00-4191, U.S. Department of the Interior and U.S. 
Geological Survey. 
 
Jacopin, C., Bertrand-Krajewski, J. L., and Desbordes, M. (1999). 
“Characterization and Settling of Solids in an Open, Grassed, Stormwater Sewer 
Network Detention Basin.” Water Science and Technology, 39(2), 135-144. 
 
Kang, J-H., Li, X-H., Lau, S-L., Kayhanian, M., and Stenstrom, M. K. (2005). 
“Application of Particle Size Distribution in Highway Runoff: Optimization of 
Settling Tank Design to Remove Particles.” The 4
th
 South Pacific Conference on 
Stormwater and Aquatic Resource Protection, Auckland, New Zealand, May 5, 2005. 
79 
 
Kobriger, N. P. and Geinopolos, A. (1984). “Sources and Migration of Highway Runoff 
Pollutants- Research Report.” Volume III, Report, FHWA/RD-84/05, (PB86-227915) 
FHWA, U.S. Department of Transportation.  
 
Lawler, D. F. and Benjamin, M. M. (2006). Chapter 11 in Water Quality Engineering: 
Physical, Chemical Treatment Processes, to be published by McGraw-Hill; New 
York. (Draft copy, 2004). 
 
Li, Y., Lau, S., Kayhanian, M., and Stenstrom, M. K. (in press, a). “Application of 
Particle Size Distribution in Highway Runoff: Optimization of Settling Tank Design to 
Remove Particles and Metals.” Journal of Environmental Engineering, ASCE. 
 
Li, Y., Lau, S., Kayhanian, M., and Stenstrom, M. K. (in press, b). “Dynamic 
Characteristics of Particles Size Distribution in Highway Runoff: Implications for 
Settling Tank Design.” Journal of Environmental Engineering, ASCE. 
 
NWSFO (National Weather Service Forecast Office). (2005). “Preliminary Local 
Climatological Data (WS Form: F-6).” Retrieved November 15, 2005 from 
http://www.srh.noaa.gov/ewx/html/climate.htm
 
Palmer, C. L. (1950). “The Pollution Effects of Storm-water Overflows from Combined 
Sewers.” Sewage and Industrial Wastes, 22(2), 154-165. 
 
Palmer, C. L. (1963). “Feasibility of Combined Sewer System.” Journal Water Pollution 
Control Federation, 35(2), 162-167. 
 
80 
Pitt, R. E., and Amy, G. (1973). ‘‘Toxic Materials Analysis of Street Contaminants.’’ 
Rep. No. EPA-R2-73-283, United States Environmental Protection Agency, 
Washington, D.C. 
 
Rickly Hydrological Company. (2004). “Dekaport Cone Sample Splitter.” Retrieved May 
20, 2005 from http://www.rickly.com/sai/dekaport.htm
 
Sansalone, J., and Buchberger, S. (1997). “Characterization of Solid and Metal Element 
Distributions in Urban Highway Stormwater.” Water Science and Technology, 36(8-
9), 155-160. 
 
Sansalone, J. and Tribouillard, T. (1999). “Variation in Characteristics of Abraded 
Roadway Particles as a Function of Particle Size.” Transportation Research Record 
No. 1690, pp. 153-163. 
 
Schueler, T.R. (1992). A Current Assessment of Urban Best Management 
Practices. Metropolitan Washington Council of Governments. 
 
USEPA (United States Environmental Protection Agency). (1999). “Storm Water 
Technology Fact Sheet: Sand Filters” Report 832-F-99-007. Retrieved on October 12, 
2005 from http://www.epa.gov/owm/mtb/sandfltr.pdf
 
U.S. Geological Survey. “The Fragile Fringe: A Guide for Teaching about Coastal 
Wetlands.” Retrieved October 31, 2005 from www.nwrc.usgs.gov/fringe/glossary.html
 
U.S. Silica. (1998). “Sil-Co-Sil 49: Fine Ground Silica.” Retrieved October 22, 2005 
from  
http://www.u-s-silica.com/PDS/Mill%20Creek/MiCSCS492000.PDF
 
81 
Vaze, J. and Chiew, F. H. S. (2004). “Nutrient Loads Associated with Different 
Sediment Sizes in Urban Stormwater and Surface Pollutants.” Journal of 
Environmental Engineering, ASCE, 130(4), 391-396. 
 
Wunderground. (2005). “Weather Station History: Weather Underground for South 
Austin.” Retrieved November 19, 2005 from 
http://www.wunderground.com/weatherstation/WXDailyHistory.asp?ID=KTXAUST
I8&month=5&day=28&year=2005
 
Wigington, P. J., Jr., Randall, C. W., and Grizzard, T. J. (1986). “Accumulation of 
Selected Trace Metals in Soils of Urban Runoff Swale Drains.” Water Resources 
Bulletin - American Water Works Association, 22(1), 73-79. 
 
Woodward-Clyde. (1994) ‘‘San Jose street sweeping equipment evaluation.’’ 
Rep., Prepared for City of San Jose, California. 
 
Young, G. K., Stein, S., Cole, P., Kammer, T., Graziano, F., and Bank, F. (1996). 
“Evaluation and Management of Highway Runoff Water Quality.” Publication No. 
FHWA-PD-96-032. Washington, DC: Federal Highway Administration, U.S. 
Department of Transportation. 
 
Zanders. J. M. (2005). “Road Sediment: Characterization and Implications for the 
Performance of Vegetated Strips for Treating Road Runoff.” Science of Total 
Environment, 339, 41-47. 
  
 
82