R. Pitt and S. Clark
February 27, 2007
Module 4b: Characteristics and Monitoring of Stormwater Particulates
Introduction ................................................................................................................................................................... 2
Characteristics of Particulates in Stormwater ............................................................................................................... 2
Dissolved and Particulate Forms of Pollutants ......................................................................................................... 2
Litter and Floatables ................................................................................................................................................. 8
Litter, floatables, and settled solids control in CSOs and SSOs .......................................................................... 8
Sediment Sources and Effects ................................................................................................................................ 10
Erosion, Channel Stability, and Sediment ......................................................................................................... 11
Particle Size Distributions ...................................................................................................................................... 12
Particle Settling Velocities ..................................................................................................................................... 19
Pollutant Associations with Different Particle Sizes .............................................................................................. 22
Colloidal Analysis ............................................................................................................................................. 27
Strengths of Major Constituent and Heavy Metal Associations with Particulates ............................................ 27
Sediment Transport in Storm Drainage Systems ........................................................................................................ 29
Background ............................................................................................................................................................ 30
Modeling solids transport in sewers. ................................................................................................................. 30
Effects of Catchbasin and Street Cleaning ........................................................................................................ 31
Settling of Particulates in Flowing Water in Storm Drainage Systems .................................................................. 32
Resuspension of Settled Particulates in Storm Drainage........................................................................................ 36
Allowable Velocity and Shear Stress ................................................................................................................ 36
Criteria to Ensure Self-Cleaning in Sewerage ................................................................................................... 41
Accumulation of Sediment in Bellevue Inlet Structures and Storm Drainage ....................................................... 44
Summary of Sediment Transport in Sewers ........................................................................................................... 46
Required Sample Line Velocities to Minimize Particle Sampling Errors ......................................................... 46
Sediment Transport in Grass Swales .......................................................................................................................... 47
Methodology .......................................................................................................................................................... 47
Indoor Experiments ........................................................................................................................................... 48
Results and Discussions .................................................................................................................................... 50
Outdoor Swale Observations .................................................................................................................................. 52
Descriptions of the Site ..................................................................................................................................... 52
Sample Collection and Analytical Methods ...................................................................................................... 53
Results and Discussions .................................................................................................................................... 53
Sediment Trapping Model Development and Verification .................................................................................... 56
Concepts ............................................................................................................................................................ 56
Settling Frequency ............................................................................................................................................. 57
Predictive Model ............................................................................................................................................... 58
Comparisons of Indoor and Outdoor Swale Observations ..................................................................................... 61
Summary of Sediment Transport in Swales ........................................................................................................... 62
Sample Processing and Particulate Analyses in Water Samples ................................................................................. 62
Sample Preparation – Churn and Cone Splitters .................................................................................................... 62
Initial Laboratory Evaluations of the USGS/Dekaport Cone Splitter ................................................................ 63
Replicate Analyses of Solids Analyses using Churn and Cone Sample Splitters .............................................. 65
General Laboratory Processing for Particle Size Fractionation of Samples ........................................................... 66
Summary of Particulate Sampling and Analyses Issues ......................................................................................... 68
Summary of the Significance of Stormwater Particulates........................................................................................... 68
References ................................................................................................................................................................... 69
1
Introduction
The characteristics of particulates in stormwater determine the effects of stormwater pollutants in the receiving
waters and dramatically influence the treatability of stormwater. Many pollutants are mostly associated with
particulates, and the removal of the particulates in common sedimentation devices also removes many of the
pollutants of interest. The transport of particulates in urban areas is complex and the method of sampling and
analysis of the particulates can have large effects on the measured values. This discussion summarizes different
characteristics of stormwater particulates (pollutant associations, particle size distributions, and settling velocity),
the transport of particulates in storm drainage systems, including grass swales and grass filters, and describes
recommended sampling and analysis methods for stormwater particulates.
Characteristics of Particulates in Stormwater
Dissolved and Particulate Forms of Pollutants
Table 1 summarizes the filterable fraction of heavy metals found in stormwater runoff sheet flows from many urban
areas (Pitt, et al. 1995). Constituents that are mostly in filterable forms have a greater potential for affecting
groundwater and are more difficult to control using conventional stormwater control practices that mostly rely on
sedimentation and filtration principles. In most cases, most of the metals are associated with the non-filterable
(suspended solids) fraction of the stormwaters. Likely exceptions include zinc that may be mostly found in the
filtered sample portions. However, dry-weather flows in storm drainage tend to have much more of the toxicants
associated with filtered sample fractions.
Table 1. Reported Filterable Fractions of Stormwater Toxicants from Source Areas
Constituent
Cadmium
Chromium
Copper
Iron
Lead
Nickel
Zinc
Filterable Fraction (%)
20 to 50
<10
<20
small amount
<20
small amount
>50
source: Pitt, et al. 1995
Pitt, et al. (1998) analyzed 550 samples for a broad list of constituents, including the total and filtered observations
shown in Table 2. The samples were collected from telecommunication manhole vaults that were mostly affected by
stormwater. However, some other contaminating water, and groundwater, sources likely also influenced these
samples. These data are very similar to cold and warm season stormwater data collected during other projects. This
is the largest data base available that contains both total and filtered metal analyses. These samples were obtained
throughout the US and represent all seasons.
Table 2. Average Particulate Fraction of Selected Constituents from 550 Nationwide Samples
(mg/L, unless otherwise noted)
Constituents
Turbidity (NTU)
COD
Color (HACH)
Copper (g/L)
Lead (g/L)
Zinc (g/L)
Total
Concentration
13
25
34
29
14
230
Filtered Concentration (after
a 0.45 m membrane filter)
1.2
22
20
9.5
3
160
Percent Associated with
Filterable Fraction
8%
86%
59%
33%
21%
70%
Percent Associated
with Particulates
91%
14%
41%
67%
79%
30%
Pitt, et al. (1998)
Sansalone (1996) investigated the forms of heavy metals in stormwater and snowmelt. It was found that Zn, Cd, and
Cu were mainly dissolved in stormwater, while only Cd was mainly dissolved in snowmelt. Pb was associated with
the finer particulate fractions in both stormwater and snowmelt. The dissolved fraction of the metals should be
2
immobilized by sorption, while the particulate bound metals should be immobilized by filtration in a partial
exfiltration trench. Another study by Sansalone and Buchberger (1997) analyzed stormwater runoff at five sites on a
heavily traveled roadway in Cincinnati, Ohio. They found that the event-mean concentrations (EMC) of Zn, Cd, and
Cu exceeded surface-water-quality-discharge standards. Further, it was noted that Zn, Cd, and Cu are mainly in the
dissolved form while other metals, i.e., Pb, Fe, and Al are mainly bound to particles.
Sedlak, et al. (1997) used special analytical techniques to determine the speciation of Cu and Ni in point and nonpoint source (NPS) discharges. They found that the existence of strong metal-complexing ligands in wastewater
effluent, and to a lesser degree, surface runoff, must be considered when evaluating metal treatability.
Grout, et al. (1999) studied the colloidal phases in urban stormwater runoff entering Brays Bayou (Houston, Texas).
Colloids in the filtrate (after 0.45 μm filtering) and further separation by ultracentrifuging, accounted for 79% of the
Al, 85% of the Fe, 52% of the Cr, 43% of the Mn, and 29% of the Zn present in the filtrates. Changes in the
colloidal composition were caused by changes in colloidal morphologies, varying from organic aggregates to diffuse
gel-like structures rich in Si, Al, and Fe. Colloids were mostly composed of silica during periods of dry weather
flow and at the maximum of the stormwater flow, while carbon dominated the colloidal fraction at the beginning and
declining stages of the storm events. Garnaud, et al. (1999) examined the geochemical speciation of particulate
metals using sequential extraction procedures for different runoff sources in Paris, France. They found that most
metals were bound to acid soluble particulates in the runoff but that Cu was almost entirely bound to oxidizable and
residual fractions.
Barry, et al. (1999) identified salinity effects on the partitioning of heavy metals in the stormwater canals entering
Port Jackson (Sydney), Australia. Cu, Pb, and Zn were found increasingly in dissolved phases as the salinity
increased in the lower sections of the canals. During high flows, most of the metals seemed to be rapidly exported
from the estuary as a discrete surface layer, while low flows contributed most of the metals to the estuary.
Shafer, et al. (1999) investigated the partitioning of trace metal levels (Al, Cd, Cu, Pb, and Zn) in Wisconsin rivers
and found that the concentrations in the rivers were comparable to recent data collected in the Great Lakes and other
river systems where ‘modern’ clean methods were used for sampling and analysis. They also found that the variation
in the partitioning coefficients of each metal between sampling locations could be explained by the amount of
anthropogenic disturbance in the watershed and by the concentration of dissolved organic carbon (DOC) in the
water.
Parker, et al. (2000) analyzed the particulates found in urban stormwater runoff in the Phoenix, Arizona,
metropolitan area. They found that the inorganic content of the particulates was similar to that in soils that were not
impacted by urban runoff. The metals concentrations (Cd, Cu, Pb, and Zn) were higher, but below levels that would
require remediation. Arsenic concentrations were above control levels; however, this contribution likely was
geologic, and not anthropogenic. Sediment toxicity was seen, but could not be explained based on their chemical
results.
Krein and Schorer (2000) investigated heavy metals and PAHs in road runoff and found, as expected, an inverse
relationship existed between particle size and particle-bound heavy metals concentrations. Sutherland, et al. (2000)
investigated the potential for road-deposited sediments in Oahu, Hawaii, to bind contaminants, and thus transporting
these bound contaminants to the receiving water as part of the runoff. In the sediment fractions smaller than 2 mm in
diameter, the origins of the Al, Co, Fe, Mn and Ni were determined to be geologic. Three of the metals
concentrations (Cu, Pb and Zn) were found to be enhanced by anthropogenic activities. Sequential extraction of the
sediment determined the associations of the metals with the following fractions: acid extractable, reducible,
oxidizable, and residual.
The fate and transport of metallic pollutants through a watershed were related to the characteristics of the solid
particles to which they are bound (Magnuson, et al. 2001). Because the particles are most often associated with
metal pollution having nominal diameters of < 50 µm, split-flow thin-cell (SPLITT) fractionation was investigated
as a means to study the metal loading as a function of particle settling rate. Sansalone, et al. (2001) reported that
urban stormwater levels of Zn, Cu, Cd, Pb, Cr, and Ni can be significantly above ambient background levels, and for
many urban and transportation land uses, often exceed surface water discharge criteria for both dissolved and
particulate-bound fractions. They advocated a multiple-unit-operation approach to stormwater treatment.
3
Glenn, et al. (2001a and 2001b) described their research at highway test sites in Cincinnati, OH, investigating the
effects of traffic activities and winter maintenance on the behavior of particulates in the runoff. They found that
urban snow has a much greater capacity to accumulate traffic-related pollutants, as compared to stormwater, due to
longer residence times before melting, and the snow’s porous matrix. Parameters such as residence time, solids
loadings, alkalinity, hardness and pH influence the heavy metal partitioning in the snow. They found that Pb, Cu,
Cd, Zn, Al, Mg, and Fe were mostly particulate bound, while Na and Ca were mostly dissolved. Partition
coefficients for most heavy metals in snowmelt water ranged from 103 to 106 L/kg.
Significant amounts of non-point source runoff were shown to enter the Santa Monica Bay (CA) from the Ballona
Creek Watershed during wet weather flow during monitoring by Buffleben, et al. (2001). The watershed is
developed mostly with residential, commercial and light industrial land uses. They found that the suspended solids
phase primarily transported the mass for five of the six metals studied: Cd, Cr, Cu, Pb, and Ni. Arsenic was found
primarily in the aqueous phase.
Mosley and Peake (2001) characterized urban runoff from a catchment in Dunedin, New Zealand, during base flows
and storm flows from five rainfall events. Fe and Pb were found to be predominantly particle-associated (>0.4 µm)
with concentrations being significantly elevated at the beginning of storm event. In contrast, the majority of Cu and
Zn was found in the <0.4 µm fraction prior to rain but a significant proportion was present in the > 0.4 µm fraction
during the initial period of storm flows. The results indicate that Cu and Zn may be more bioavailable, and more
difficult to remove by stormwater treatment, than Pb. The pH level and the concentration of major ions (Ca +2, Na+,
Mg+2, K+), dissolved PO4-P, and NO3 generally decreased during storm flows due to rainwater dilution.
Concentrations of total N and P often increased during the initial period of storm runoff, likely because of washoff
of particulate plant material.
Fan, et al. (2001b) reviewed the transport of toxic pollutants through multiple media and drainage systems in urban
watershed during wet-weather periods and found that a major portion of priority pollutants (including benzene,
polynuclear aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs), pesticides, and heavy metals) are in
particulate form, or sorbed onto particles.
Table 3 summarizes inlet and outlet concentrations for total and dissolved metals (Cd, Cr, Cu, Pb, Ni, and Zn), TSS,
hardness and total oil and grease for a retrofitted pond in Sunnyvale, CA. The metals removal data indicated that the
removal of total Cr, Cu, Pb, Ni and Zn were well correlated with TSS removal. Tables 4 through 6 also summarize
the particulate and filterable fraction of stormwater heavy metals from a number of studies. In almost all cases, the
heavy metals are mostly associated with particulates, except for Zn that is mostly associated with the filterable
fraction for most areas. Interesting exceptions are noted, however. As an example, zinc stormwater concentrations
from Birmingham, AL, industrial storage areas were found to be almost completely associated with the particulate
fraction. These samples were apparently not affected by runoff from areas having galvanized metals, but were
affected by heavy truck traffic, where the particulate forms of Zn would be mostly from tire wear.
4
Table 3. Average Inlet and Outlet Observed Concentrations and Pollutant Removals (Sunnyvale,
CA)
nf: non-filtered (total)
f: filtered (“dissolved”)
removals are only given if most observations were >PQL
5
Table 4. Filterable Fraction of Heavy Metals Observed at the Inlet to the Monroe St. Wet Detention
Pond, Madison, WI (average and standard deviation)
Number of observations
Average total
concentration
Average filtered
concentration
Average percentage
filterable
Average percentage
associated with
particulates
Copper
60 to 64
Lead
59 to 64
Zinc
57 to 64
50 (14)
85 (52)
152 (136)
6.4 (3.3)
3.5 (1.7)
51 (34)
13%
4.1%
34%
87%
96%
66%
Data from: House, et al. 1993.
Table 5. Milwaukee and Long Island NURP Source Area Heavy Metal Associations (based on mean
concentrations observed)
Residential roof
runoff
% filt
% part
Arsenicb
Cadmiumb
Chromiumb
Leada
Leadb
a
b
6
8
92
Commercial parking
runoff
% filt
% part
25
75
18
82
24
76
3
97
16
84
Bannerman, et al. 1983 (Milwaukee) (NURP)
STORET Site #59629A6-2954843 (Huntington-Long Island, NY) (NURP)
Table 6. Birmingham, AL, Source Area Heavy Metal Particulate Associations (based on mean
concentrations observed)
a
na: not available, too few detectable observations for calculation
Pitt, et al. 1999
7
Johnson, et al. (2003) conducted tests to examine the treatability of stormwater heavy metals, and some major
constituents and nutrients, and their associations with different-sized particulates in stormwater. The binding
strengths of these pollutants to the particulates were also examined by using a sequential extraction procedure using
different acids and bases under a wide range of pH and Eh conditions. Experiments examining the characteristics of
the filterable (<0.45 m) portion of the heavy metals were also conducted. Some of these results are summarized in
later sections of this module.
Litter and Floatables
Floatables are another form of the solids in runoff that are receiving attention. They obviously interfere with
aesthetics and recreation, along with having potential adverse biological effects. A current ASCE tack force, headed
by Betty Rushton, is developing a protocol for monitoring floatables in stormwaters, including methods to evaluate
floatable trapping benefits of stormwater control measures. In all cases, a complete mass balance is necessary to
adequately evaluate controls (and to verify the accuracy of the monitoring and lab analyses). Influent, effluent, and
trapped material must all be quantified for the period of study.
Armitage and Rooseboom (2000a) demonstrated that large quantities of litter are being transported in South Africa
stormwater runoff, and that the amount of litter produced was related to land use, vegetation, level of street cleaning
and type of rainfall. The benefits of litter reduction were documented using work from Australia and New Zealand,
and design equations for sizing litter traps were proposed (Armitage and Rooseboom 2000b). Newman, et al.
(2000a) characterized the floatables found in urban stormwater runoff.
Riverine litter and floatables occupy a spatial and temporal position in any systematic analysis of river systems and
represent a problem that was increasing in scale. Quantifiable source factors of litter in the river Taff, South Wales,
United Kingdom, system were found to be mainly sewage inputs through CSO and fly tipping (Williams and
Simmons 1997a). Sewage-derived material constituted approximately 23% of all items on the river, large quantities
of waste, especially plastic sheeting, originated from fly tipping sites (Williams and Simmons, 1999). River bank
clearances provided valuable information on litter accumulation. Tracking also showed a distinctive correlation
between flood events and litter movement. Some litter types have an increased input during flood events, e.g.,
sanitary-wastewater-derived material from combined sewer outfalls, while accumulation of other litter types could
be due to their distribution throughout the catchment (Williams and Simmons 1997b).
Litter, floatables, and settled solids control in CSOs and SSOs
Much information exists on the removal of litter and floatables from combined sewer overflows, while less is
available concerning litter control from separate stormwater. Some of the CSO data may be applicable to stormwater
conditions and processes.
McGregor (2003) made the case for cleaning sanitary sewers to prevent SSOs. The case study presented in the paper
showed that an overflow problem might not have occurred if the municipality had adhered to a more proactive
maintenance program. Pisano, et al. (2003) summarized the design of passive automatic flushing systems installed
in the City of Cambridge's storm and sanitary sewer system tributary to the Alewife Brook as part of a $75 million
sewer separation program. The flushing systems were sized to achieve wave velocity of 1 m/s at the end of the
flushing segment with flush vault volumes ranging from 11 to 40 m3 for the storm drain systems and 6 m3 for the
sanitary system. Pollert and Stransky (2003) evaluated the separation efficiency of solids from CSOs through the use
of a combined 1D and 3D models. Charts of concentration dynamics of suspended solids on the CSO inflow,
outflow and overflow were presented (for particles with different characteristics).
Fischer and Turner (2002) reviewed the North Bergen, NJ Solids and Floatables Control facility, which uses a
system of nine Netting TrashTrap® system and one mechanical screen, which cost 300% less than the next least
expensive technology evaluated. Operations and maintenance costs on the facility are approximately $57,000 per
year. Irvine (2002) discussed the Buffalo River (NY) floatables control program that uses a floatables trap and
continuous water quality monitoring. Mean accumulations rates in the trap were significantly greater for CSO
periods than for dry weather and the distributions in the traps had more wood and less plastic than floatables traps in
8
New Jersey. The average mass trapped per unit volume was also less for the Buffalo watershed than for the two New
Jersey watersheds monitored, possibly due to the inlet hoods installed in the Buffalo watershed.
Dettmar, et al. (2002) investigated the performance and operation of flushing devices in both German field and
laboratory studies. The results showed that the generated flushing wave shear stresses of the investigated gates
surpass 14 N/m2. The yields from the cleaning operations netted 0.5 kg sediment per m3 of flushing water.
The paper by Fan, et al. (2001a) reviewed the causes of sewer deterioration and the control methods that can be used
to prevent or postpone this deterioration. The remedy reviewed in the paper was inline and combined sewer
overflow tank flushing systems to remove sediments and thus minimize hydrogen sulfide production in the sewer
system. The two technologies reviewed were the tipping flusher and the flushing gate, both of whom have been
demonstrated to perform well in facilities in Germany, Canada and the United States. Fan and Field (2002) extended
their overview of the causes of combined-sewer deterioration and associated large pollutant discharges, along with a
discussion of their control methods. Their results indicated that both the tipping flusher and the flushing gate
technologies appear to be cost-effective for flushing solids and debris from CSO storage tanks, while the flushing
gate appears to be the most efficient method for controlling combined-sewer sediment. The use of slotted invert
traps for managing sewer sediments was reviewed by Buxton, et al. (2002). The paper presented the results from a
laboratory investigation that compared the trapping performance of three slot-size configurations of a laboratoryscale invert trap and from 2-D modeling of stochastic particle tracking. The comparison of the two procedures
showed that the modeling consistently overpredicted retention efficiencies compared to the laboratory observations.
The effects of a manhole drop on the hydraulics of combined sewer flow were investigated by De Martino, et al.
(2002). The main wave features were analyzed to give expressions that contain both the upstream filling ratio and
the Froude number of the approach flow. The discharge capacity was also investigated and photographs were
presented to show typical drop flow in combined sewer manholes.
Sewer solids which are deposited in combined and sanitary sewers during dry weather, and are then flushed during
overflows, make up a significant component of CSO and SSO pollutant discharges. Detailed research investigating
all aspects of solids in sewer systems has been underway in Europe for nearly two decades. Recent research has
characterized the nature of the solids getting into sewer systems, how they behave in terms of transport, and some of
the main aspects of their effects. Ashley, et al. (2000) has demonstrated that in a number of catchments, the majority
of pollutants found in suspension during storms, and likely to be discharged from overflows, originate from the
predominantly organic 'near bed solids' which accumulate in systems during dry weather. This knowledge has been
used to improve the modeling of sewer solids behavior. The USEPA has also investigated the causes of sewer solids
deposition and the development/evaluation of control methods to prevent sediment accumulation. Control of sewer
sediment not only protects urban receiving water quality but also protects sewer structural integrity (Fan, et al.
2000b). Chen and Leung (2000) demonstrated that the sediment phase played a key role in oxygen utilization in a
sanitary gravity sewer. Their study indicated that the sediment phase contained more active biomass than the sewage
phase. Del Giudice, et al. (2000) examined supercritical flow in a bend manhole used in sanitary sewers, and
introduced the bend cover, which allows air entrainment into the downstream sewer. The bend cover may
significantly increase the performance and capacity of bend flow and may easily be added to existing manholes. A
three-dimensional particle settling model was used to predict the benthic exposure zone to sewage solids for a large
marine outfall and diffuser serving Nanaimo, Canada. The predicted exposure zone was determined largely by
horizontal advection, and to a lesser extent by the relative amounts of floc and non-floc solids settling at two
different rates (Hodgins, et al. 2000).
Burgess, et al. (1999) reported that the City of Indianapolis developed five CSO control facilities as "pilot" facilities,
for the purpose of establishing and demonstrating the local suitability and effectiveness of the various technologies
that they employ. Preliminary results, with emphasis on the capture characteristics of the floatable debris netting trap
facilities, were presented. Bridgeport's (Connecticut) Water Pollution Control Authority has recently faced the
problem of floating debris and the sound’s high tide. These drainage problems have been successfully solved by
building new flow-control regulators with integrated automatic radio (wireless) and telephone-modem (wired)
control, water-flow, and instrumentation technologies (Morrill and Sound, 1999).
9
Arthur and Ashley (1998) examined near-bed solids transport in combined sewers, developed a method to estimate
the rate of sediment transport near the bed in sewers and concluded that near-bed transport of sediment may
contribute to the first foul flush phenomenon in sewers. A study to quantify the first flush phenomenon in
stormwater runoff failed to identify any consistent, strong indication that the first flush runoff from a storm contains
higher concentrations of pollutants than the event average. Analysis of SS in the River Seine (Paris, France), both
upstream and downstream of major CSOs, showed that SS from wet-weather events settled rapidly, after which they
are gradually mobilized from the river bed during dry weather (Estèbe, et al. 1998). Dimensionless mass versus
volume curves for 197 runoff events in 12 separate and combined sewer systems depended on the pollutant, the site,
the rainfall event and the workings of the sewer system. No general relationships were established because of the
complexity of the phenomena involved and the multiplicity of influencing factors (Bertrand-Krajewski, et al. 1998).
Only slight first flush effects for SS and conductivity of storm runoff were observed at single road inlets for two
catchments in Yugoslavia and Sweden and no first flush effect was recorded for pH or temperature. After a year of
study, conclusions were that the first flush, if strongly present at the end of a drainage system, is not generated by
the first flush of pollution input and that regression curves are not very reliable for prediction of the first flush load
of pollution input into drainage systems (Deletic 1998).
Sediment Sources and Effects
Rossi, et al. (2005) performed stochastic modeling of TSS in urban areas during rain events. The model predicted
the probability of TSS loads arising from CSOs as well as from stormwater in separate sewer systems. The different
washoff behaviors of coarse versus fine suspended solids was observed in highway runoff by Aryal, et al. (2005a).
The concentration of the fines was more consistent compared to the coarse fraction, as was the PAH content
measured in the fine fraction.
Fan, et al. (2003) reviewed the sewer sediment control projects conducted by the Wet-Weather Research Program of
the US EPA. The projects focused on the relationship between wastewater characteristics and flow-carrying
velocity, on the abatement of solids deposition and solids resuspension in sewers, and on the sewerline flushing
systems for removal of sewer sediment.
Goodwin, et al. (2003) investigated the temporal and spatial variability of sediment transport and yields in the
Bradford Beck catchment (UK). The results demonstrated that for individual storms the sediment yields from the
urban sub-catchment were generally higher than those from the rural system although the annual yields were
comparable. For large events, urban sediment transport was dominated by the impact of CSO discharges. Cristina, et
al. (2003) defined “first-flush” criteria for solids from small impervious watersheds, dividing the category into two
types. They were a concentration-based first flush and a mass-based first flush, and the operational definition should
define treatment design and operational efficiency.
Paul, et al. (2002) evaluated the quantitative relationships that had been previously developed between landscape
metrics and sediment contamination in 75 small estuarine systems across the mid-Atlantic and southern New
England regions of the U.S. The landscape metrics important for explaining the variation in sediment metals levels
(R2 = 0.72) were the percent area of nonforested wetlands (negative contribution), percent area of urban land, and
point source effluent volume and metals input (positive contributions). The metrics important for sediment organics
levels (R2 = 0.5) and total PAHs (R2 = 0.46) were percent area of urban land (positive contribution) and percent area
of nonforested wetlands (negative contribution).
Nelson (1999) described the sediment budget of Issaquah Creek, a 144 km2 mixed-use, urbanizing watershed near
Seattle, WA. The water quality of Lake Sammamish, located at the outlet of the basin, was degrading with time, and
fine sediment entering the lake from the watershed was a likely source of phosphorus during periods of lake anoxia.
Another potential in-channel concern was the effect of fine sediment on spawning gravel for the salmon species that
occupy Issaquah Creek. The sediment balance was being used to identify the major sources of sediment, and thus
guide the most effective remedial measures. Nelson and Booth (2002) reported on sediment sources in the Issaquah
Creek watershed. Human activity in the watershed, particularly urban development, has caused an increase of nearly
50% in the annual sediment yield, now estimated to be 44 tons km-2 yr-1. The main sources of sediment in the
watershed are landslides (50%), channel-bank erosion (20%), and road-surface erosion (15%).
10
Benoit, et al. (1999a) studied sources of sediment entering Jordan Cove, Connecticut. Recent sediment accumulation
rates were found to be decreasing (from 0.84 cm/yr to 0.40 cm/yr) but were slightly faster than relative sea-level rise
at this site (0.3 cm/yr). Long Island Sound was found to be an important source of sediment to the cove; a minor part
of total sediment was supplied from the local watershed. Benoit, et al. (1999b) also studied sources and the history
of heavy metal contamination and sediment deposition in Tivoli South Bay, Hudson River, New York. The
measured sedimentation rate ranged from 0.59 to 2.9 cm/yr suggesting that rapid accumulation occurred during the
time period represented by the length of the cores (approximately the past 50 yrs). The sources of this material were
expected to be upland streams, or the Hudson River, during storm events. Concentrations of Pb, Cu and Zn
correlated with each other within individual cores at five of the six sites tested, suggesting a common proximate
source.
Erosion, Channel Stability, and Sediment
Tetzlaff, et al. (2005) used hydrological criteria to assess changes in the flow dynamic in urbanized catchments.
Flow acceleration, peak discharge and discharge dosage were studied. Miller, et al. (2005) investigated the impact of
urbanization and the associated hydrology and channel morphology on fundamental ecosystem processes such as
leaf litter decomposition and transport, which is a crucial part of restoring urban watersheds and streams. Yeager, et
al. (2005) discussed urbanization-induced stream erosion in southern California. The variability in stream response
to urbanization was assessed.
Eyles, et al. (2003) investigated the impacts of urbanization on the geophysical and sedimentological characteristics
in a Lake Ontario watershed. The urban-impacted watershed empties into the shallow, semi-enclosed coastal lagoon
of Frenchman's Bay, which served as a trap for fine-grained contaminated sediment. Blazier, et al. (2003)
investigated particle agglomeration in runoff and treatment processes, including the effects of pH, ionic strength and
Camp and Stein’s velocity gradient. Agglomeration rates increased with lower pH and were optimized when mixing
rates were set to avoid particle sedimentation and to avoid fluid shears.
An emerging concept in channel design uses sediment transport as the basis for quantifying channel stability. Byars
and Kelly (2001) examined channel stability in the Austin, TX, area. They concluded that a channel that is not
undergoing excessive erosion or sedimentation has a function and form similar to a natural stream and should be the
goal of good channel design. This approach, however, requires a more comprehensive understanding of the local
climate, geology, hydrology, and stream mechanics than historically utilized.
Hunt and Grow (2001) describe a field study conducted to determine the qualitative and quantitative impact to a
stream from a poorly controlled construction site. They used fish electroshocking, Qualitative Habitat Evaluation
Index, and zigzag pebble count studies. The 33 acre construction site consisted of severely eroded silt and clay loam
subsoil and was located within the Turkey Creek drainage, Scioto County, OH. The number of fish species declined
(26 to 19) and the number of fish found (525 to 230) decreased significantly when comparing upstream unimpacted
reaches to areas below the heavily eroding site. The Index of Biotic Integrity and the Modified Index of Well-Being,
common fisheries indexes for stream quality, were therefore reduced from 46 to 32 and 8.3 to 6.3, respectively.
Upstream of the area of impact, Turkey Creek had the highest water quality designation available (Exceptional
Warm Water Habitat); but fell to the lowest water quality designation (Limited Resource Water) in the area of the
construction activity. Water quality chemical analyses conducted on samples from upstream and downstream sites
verified that these impacts were not from chemical affects alone.
Morrisey, et al. (2000) described the sampling program used to confirm a predictive model of metal contaminant
build-up (Cu, Pb, and Zn) in the sediments of sheltered urban estuaries in Aukland, New Zealand that have been
subjected to urban runoff inflows. The results of their testing showed good general agreement between the model
predictions and the observed concentrations of metals in the sediments. The paper by Butcher, et al. (2000)
described the problems encountered when developing mercury TMDLs for the Arivaca and Pena Blanca Lakes in
Arizona. These two lakes lacked point-source discharges of mercury; however, the concentrations of mercury in fish
bodies were sufficiently high to trigger TMDL development. The resultant TMDL addressed the problems inherent
with controlling pollutants entering the lake when the lake sediment was found to be a primary source of the
pollutant. Rate, et al. (2000) investigated the concentration of heavy metals in sediments of the Swan River estuary
in Perth, Australia. They found that the concentration of lead was elevated near stormwater drain outfalls when
11
compared to areas away from the outfalls, likely due to vehicular material; no similar effect was seen for copper or
cadmium. They also noted that since the vast majority of all heavy metals were bound to iron oxides or organic
sediments, most of the metals are not bioavailable. The results of the study performed by Rochfort, et al. (2000) on
the effects of stormwater and CSO discharges on the benthic community showed that the levels of metals and PAHs
in sediments below these discharges were high. However, biological effects were not seen - neither the toxicity
endpoints nor the benthic community descriptors could be related to the sediment contaminant levels.
Ghani, et al. (1999) found that the thickness of a sediment deposit on the bottom of a rigid rectangular channel
greatly affects its erodibility of the deposits. They developed channel erosion equations that included terms for the
deposit’s thickness. Keshavarzy and Ball (1999) studied the entrainment of sediment particles in water and found
that the number of entrained particles per unit time per unit area was found to be related to the instantaneous shear
stress at the bed. These results were used to modify the Shields diagram. Rhoads and Cahill (1999) studied the
elevated concentrations of chromium, copper, lead, nickel and zinc that were found in sediments near storm sewer
outfalls. They noted that copper and zinc concentrations were greater in the bedload compared to the bed material
and therefore were more likely to be mobilized during runoff events.
Stormwater impacts to streams are not limited to the relatively short duration of runoff events. As an example,
sediments can dominate the aquatic physiochemical and biological processing of nutrients; sediment contaminated
by stormwater pollutants has a detrimental effect on the receiving-water-biological community. The EPA and other
regulatory agencies are attempting to develop sediment quality criteria to determine where excessive concentrations
of chemical constituents are present in sediments at sufficient concentrations and in chemical forms to be
significantly adverse to the designated beneficial uses of the associated water body. Lee and Jones-Lee (1996a)
presented the issues that need to be considered in evaluating the results of a sediment quality assessment procedure
to determine whether the toxicity or excessive concentration found is a potentially significant cause of real water
quality deterioration in the water body of concern.
Particle Size Distributions
Knowing the settling velocity characteristics associated with stormwater particulates is necessary when designing
wet detention ponds and other stormwater control devices that rely on sedimentation processes. Particle size is
directly related to settling velocity (using Stokes law, for example, and using appropriate shape factors, specific
gravity and viscosity values) and is usually used in the design of detention facilities. Particle size can also be much
more rapidly measured in the laboratory than settling velocities. Settling tests for stormwater particulates need to be
conducted for about three days in order to quantify the smallest particles that are of interest in the design of wet
detention ponds. If designing rapid treatment systems (such as grit chambers or vortex separators for CSO
treatment), then much more rapid settling tests can be conducted. Probably the earliest description of conventional
particle settling tests for stormwater samples was made by Whipple and Hunter (1981).
Whipple and Hunter (1981) contradicted the assumption sometimes used in modeling detention pond performance
that pollutants generally settle out in proportion to their concentrations (first-order rate equations). However,
Grizzard and Randall (1986) have shown a relationship between particulate concentrations and particle size
distributions. High particulate concentrations were found to be associated with particle size distributions that had
relatively high quantities of larger particulates, in contrast to waters having low particulate concentrations. The
water having high particulate concentrations would therefore have increased particulate removals in detention
ponds. This relationship is expected to be applicable for pollutants found mostly in particulate forms (such as
suspended solids and most heavy metals), but the relationship between concentration and settling would be much
poorer for pollutants that are mostly in soluble forms (such as filterable residue, chlorides and most nutrients).
Therefore, the partitioning of specific pollutants between the “particulate” and “dissolved” forms, and eventually for
different particulate size fractions, is needed.
Smith (1982) also states that settleability characteristics of the pollutants, especially their particle size distribution,
are needed before detention pond analyses can be made. Kamedulski and McCuen (1979) report that as the fraction
of larger particles increase, the fraction of the pollutant load that settles also increases. Randall, et al. (1982), in
settleability tests of urban runoff, found that non-filterable residue (suspended solids) behaves liked a mixture of
discrete and flocculent particles. The larger discrete particles settled out rapidly, while the flocculent particles
12
(having small specific gravities) were very slow to settle out. Therefore, simple particle size information may not be
sufficient when flocculent particles are also present. Particle size analyses should include identification of the
particle by microscopic examination to predict the extent of potential flocculation.
Detailed pond studies have provided additional in-sight to the particle sizes in stormwater and their behavior in wet
ponds. Greb and Bannerman (1997) reported on a long-term pond study for Madison, WI. They found that the
influent particle-size distribution affected the efficiency of the wet pond’s performance on sediment and associated
pollutant removal. Krishnappan, et al. (1999) monitored in-pond particle sizes using a submersible laser particle size
analyzer, reducing potential changes in particle characteristics that may occur during sampling. The suspended
solids were mostly composed of flocs, and were about 30 μm in size during winter surveys and about 210 μm during
the summer surveys. Three seasonal surveys of suspended solids were carried out in an on-stream stormwater
management pond, by means of a submersible laser particle size analyzer. Size distributions were measured at up to
17 points in the pond, and water samples collected at the same locations were analyzed for primary particles
aggregated in flocs. Using a relationship defining the floc density as a function of floc size and Stokes’ equation for
settling, an empirical relationship expressing the free fall velocity as a function of floc size was produced
(Krishnappan, et al. 1999). It is possible that the flocs formed were associated with in-pond chemical and biological
processes and were not representative of influent particle size conditions. Pettersson (2002) investigated the
characteristics of suspended particles in a small stormwater pond. The results showed the particle volume in the
stormwater (particle volume per stormwater volume) predominately consisted of very fine particles and that the
smallest particles comprised most of the surface area. The stormwater pollutants exhibited strong correlations with
particle characteristics. Petterson (1999) examined the partitioning of heavy metals in particulate-bound and
dissolved phases in a stormwater pond in Goteborg, Sweden. The results showed a clear variation in lead
partitioning affected by specific conductivity.
Figure 1 shows approximate stormwater particle size distributions derived from several upper Midwest and Ontario
analyses, from all of the Nationwide Urban Runoff Program (NURP) data (EPA 1983; Driscoll 1986), and for
several eastern sites that reflect various residue concentrations (Grizzard and Randall 1986). Pitt and McLean (1986)
microscopically measured the particles in selected stormwater samples collected during the Humber River Pilot
Watershed Study in Toronto. The upper Midwest data sources were two NURP projects: Terstriep, et al. (1982), in
Champaign/Urbana Ill. and Akeley (1980) in Washtenaw County, Michigan.
13
Figure 1. Particle size distributions for various stormwater sample groups.
Over the years, relatively few samples have been analyzed for stormwater particle sizes and no significant trends
have been identified relating the particle size distribution to land use or storm condition. However, the work by
Grizzard and Randall (1986) does indicate significantly different particle size distributions for stormwaters from the
same site having different suspended solids concentrations. The highest suspended solids concentrations were
associated with waters having relatively few small particles, while the low suspended solids concentration waters
had few large particles. The particle size distribution for the upper Midwest urban runoff samples falls between the
medium and high particulate concentration particle size distributions.
For many urban runoff conditions, the median stormwater particle size is estimated to be about 30 m, (which can
be much smaller than the median particle size of some source area particulates). Very few particles larger than 1000
m are found in stormwater, but particles smaller than 10 m are expected to make up more than 20 percent of the
stormwater total residue weight.
Specific conditions (such as source area type, rain conditions and upstream controls) have been shown to have
dramatic effects on particle size distributions. Randall, et al. (1982) monitored particle size distributions in runoff
from a shopping mall that was cleaned daily by street cleaning. Their data (only collected during the rising limb of
the hydrographs) showed that about 80 percent of the particles were smaller than 25 m, in contrast to about 40
percent that were smaller than 25 m during the outfall studies. They also only found about two percent of the
runoff particles in sizes greater than 65 m, while the outfall studies found about 35 percent of the particles in sizes
greater than 65 m.
Limited data is also available concerning the particle size distribution of erosion runoff from construction sites.
Hittman (1976) reported erosion runoff having about 70 percent of the particles (by weight) in the clay fraction (less
than four m), while the exposed soil being eroded only had about 15 to 25 percent of the particles (by weight) in
the clay fraction. When the available data is examined, it is apparent that many factors affect runoff particle sizes.
Rain characteristics, soil type, and on-site erosion controls are all important.
14
Tests have also been conducted to examine the routing of particles through the Monroe St. detention pond in
Madison, Wisconsin (Roger Bannerman, Wisconsin Department of Natural Resources, personal communication).
This detention pond serves an area that is mostly comprised of medium residential, with some strip commercial
areas. This joint project of the Wisconsin Department of Natural Resources and the U.S. Geological Survey has
obtained a number of inlet and outlet particle size distributions for a wide variety of storms. The observed median
particle sizes ranged from about 2 to 26 m, with an average of 9 m. The following list shows the average particle
sizes corresponding to various distribution percentages for the Monroe St. outfall:
Percent larger
than size
Particle Size
(m)
10 %
25
50
75
90
450
97
9.1
2.3
0.8
These distributions included bedload material that was also sampled and analyzed during these tests. This
distribution is generally comparable to the “all NURP” particle size distribution presented previously. The critical
particle sizes corresponding to the 50 and 90 percent control values are as follows for the different data groups:
Monroe St.
All NURP
Midwest
Low solids conc.
Medium solids conc.
High solids conc.
90 %
50%
0.8
1
3.2
1.4
3.1
8
9.1 m
8
34
4.4
21
66
The particle size distributions of stormwater at different locations in an urban area greatly affect the ability of
different source area and inlet controls in reducing the discharge of stormwater pollutants. A series of U.S.
Environmental Protection Agency (USEPA) funded research projects has examined the sources and treatability of
urban stormwater pollutants (Pitt, et al. 1995). This research has included particle size analyses of 121 stormwater
inlet samples from three states (southern New Jersey; Birmingham, Alabama; and at several cities in Wisconsin) in
the U.S. that were not affected by stormwater controls. Particle sizes were measured using a Coulter Counter MultiSizer IIe and verified with microscopic, sieve, and settling column tests. Figures 2 through 4 are grouped box and
whisker plots showing the particle sizes (in m) corresponding to the 10th, 50th (median) and 90th percentiles of the
cumulative distributions. If 90 percent control of suspended solids (by mass) was desired, then the particles larger
than the 90th percentile would have to be removed, for example. In all cases, the New Jersey samples had the
smallest particle sizes (even though they were collected using manual “dipper” samplers and not automatic samplers
that may miss the largest particles), followed by Wisconsin, and then Birmingham, Alabama, which had the largest
particles (which were collected using automatic samplers and had the largest rain intensities). The New Jersey
samples were obtained from gutter flows in a residential neighborhood that was xeroscaped, the Wisconsin samples
were obtained from several source areas, including parking areas and gutter flows mostly from residential, but from
some commercial areas, and the Birmingham samples were collected from a long-term parking area on the UAB
campus. Generally, source area samples have larger particles, while outfall samples do not have the largest particles
that washoff from the source areas. The largest particles accumulate in the drainage system and the runoff water
does not have the power to transport these materials to the outfall.
15
Figure 2. Tenth percentile particle sizes for stormwater inlet flows (Pitt, et al. 1997).
Figure 3. Fiftieth percentile particle sizes for stormwater inlet flows (Pitt, et al. 1997).
16
Figure 4. Ninetieth percentile particle sizes for stormwater inlet flows (Pitt, et al. 1997).
The median particle sizes ranged from 0.6 to 38m and averaged 14m. The 90th percentile sizes ranged from 0.5 to
11m and averaged 3m. These particle sizes are all substantially smaller than have been typically assumed for
stormwater.
Stormwater particle size distributions typically do not include bed load components because automatic sampler
intakes are usually located above the bottom of the pipe where the bed load occurs. During the Monroe St.
(Madison, WI) detention pond monitoring, the USGS and WI DNR installed special bed load samplers that trapped
the bed load material for analysis. The bedload samplers were liter-sized wide mouth containers which were placed
in bored holes in the bottom of the enclosed flat bottomed concrete small box channels right before the pond. Three
units were placed in the channel bottom, each having different width slots cut in their lids. The mass of material
trapped was directly related to the ratio of the width of the slot to the width of the channel. The material was
removed, dried and weighed, and sieved. This particle size distribution was combined with the flow-weighted
particle distributions obtained for the runoff events monitored during the same exposure period. Practically all
events were monitored, with little flow not represented in these analyses. Figure 5 shows the measured particle size
distributions for 16 seasonal samples (each having several runoff events), also including the bedload particle size
distributions. The bedload samplers were in place for several weeks at a time in order to accumulate sufficient
sample for analyses. The bedload material was comprised of the largest material represented on this figure
(generally about 300 or 400 m and larger particulates) and comprised about 10 percent of the annual total solids
loading, but ranged from about 2 to 25 percent for individual periods. The bed load component in Madison was most
significant during the early spring rains when much of the traction control sand that could be removed by rains was
being washed from the streets. This is not a large fraction of the solids, but it represents the largest particle sizes
flowing in the stormwater and it can be easily trapped in most detention ponds or catchbasins.
17
Figure 5. Inlet particle size distributions observed at the Monroe St. wet detention pond.
Figure 6 shows a typical “delta” of large material immediately near the influent to a wet detention pond, along with
an accumulation of material along the invert of the corrugated steel drainage pipe. This photo was taken at a pond in
Snowmass, CO, in an area of heavy sand applications for traction control in the winter. The bedload sediment
material in this photo is quite large, several mm in diameter, and near the upper range of the particle size distribution
shown previously. This drainage pipe is relatively short, connected to an adjacent parking area. It is rare for this
large material to be transported great distances in drainage systems. If it does enter a pond, or any type of sediment
device, it is easily trapped near the inlet. However, moderate sized particulates can easily be transported quite some
distance in the drainage system and be deposited well away from the inlet when discharged into wet detention
ponds. The use of forebays, or small pre-settlement ponds, will trap much of the moderate-sized particulates (down
to about 25m) within a relatively small area for easier removal during maintenance operations. The finer
particulates (down to just a few m), containing most of the pollutants, would be trapped in the main pond area,
although the sediment mass is relatively small.
18
Figure 6. Bedload sediment accumulation in sewerage and near outlet in pond (Snowmass, CO).
Particle Settling Velocities
The settling velocities of discrete particles are shown in Figure 7, based on Stoke’s and Newton’s settling
relationships. Probably more than 90% of all stormwater particulates are in the 1 to 100 m range, corresponding to
laminar flow conditions, and appropriate for using Stoke’s law. This figure also illustrates the effects of different
specific gravities on the settling rates. In most cases, stormwater particulates have specific gravities in the range of
1.5 to 2.5. This corresponds to a relatively narrow range of settling rates for a specific particle size. Particle size is
much easier to measure than settling rates and it is generally recommended to measure particle sizes using
automated particle sizing equipment (such as a Coulter Counter Multi-Sizer III) and to conduct periodic settling
column tests to determine the corresponding specific gravities. If the particle counting equipment is not available,
then small scale settling column tests (using 50 cm diameter Teflon columns about 0.7 m long) can be easily used.
19
Figure 7. Type 1 (discrete) settling of spheres in water at 10 C (Reynolds 1982).
Particle settling observations in actual detention ponds have generally confirmed the ability of well designed and
operated detention ponds to capture the “design” particles. Gietz (1983) found that particles smaller than 20 m
were predominate (comprised between 50 to 70 percent of the sediment) at the outlet end of a “long” monitored
pond, while they only made up about ten to 15 percent of the sediment at the inlet end. Particles between 20 and 40
m were generally uniformly distributed throughout the pond length, and particles greater than 40 m were only
found in the upper (inlet) areas of the pond. The smaller particles were also found to be resuspended during certain
events.
Pisano and Brombach (1996) summarized numerous solids settling curves for stormwater and CSO samples. They
are concerned that many of the samples analyzed for particle size are not representative of the true particle size
distribution in the sample. As an example, it is well known that automatic samplers do not sample the largest
particles that are found in the bedload portion of the flows. Particles having settling velocities in the 1 to 15 cm/sec
range are found in grit chambers and catchbasins, but are not seen in stormwater samples obtained by automatic
samplers, for example. It is recommended that bedload samplers be used to supplement automatic water samplers in
order to obtain more accurate particle size distributions (Burton and Pitt 2002). Selected US and Canadian settling
velocity data are shown in Table 7. The CSO particulates have much greater settling velocities than the other
samples, while the stormwater has the smallest settling velocities. The corresponding “Stoke’s” particle sizes for the
20
geometric means are about 100 m for the CSOs, about 50 m for the sanitary sewage, and about 15 m for the
stormwater.
Table 7. Settling Velocities for Wastewater, Stormwater, and CSO (Pisano and Bromback 1996)
Samples
CSO
dry weather wastewater (sanitary sewage)
stormwater
Geometric Means of
Settling Velocities
Observed (cm/sec)
0.22
0.045
0.011
Range of Medians of
Settling Velocities
Observed (cm/sec)
0.01 to 5.5
0.030 to 0.066
0.0015 to 0.15
More than 13,000 CSO control tanks have been built in Germany using the ATV 128 rule (Pisano and Bromback
1996). This rule states that clarifier tanks (about 1/3 of these CSO tanks) are to retain all particles having settling
velocities greater than 10 m/hr (0.7 cm/sec), with a goal of capturing 80% of the settleable solids. Their recent
measurements of overflows from some of these tanks indicate that the 80% capture was average for these tanks and
that the ATV 128 rule appears to be reasonable.
Bzdusek, et al. (2005) investigated the potential for fine-grained sediment transport and deposition in Wisconsin
using sediment-core tracer profiles. Cesium-137 activity versus depth showed distinct peaks related to peak river
discharges in specific years. Sediment deposition has continued near one urban bridge.
Furumai, et al. (2002b) studied the dynamic behavior of suspended pollutants and particle size distributions in
highway runoff. Except for Pb, the concentrations of TSS and heavy metals in runoff were within the range of the
EMC reported in recent highway runoff research. Particle-bound heavy metals (Zn, Pb, and Cu) accounted for more
significant pollutant loads than soluble fractions. Their content decreased with increasing total SS concentration in
runoff samples. The results of particle size distribution (PSD) analysis of runoff samples indicate that high TSS
concentration samples contained coarser particles. Based on the PSD results, a stepwise wash-off phenomenon of
TSS under varying runoff rate conditions was explained by the different washoff behavior of fine (less than or equal
20 μm) and coarser particles.
Hijioka, et al. (2000) investigated the behavior of suspended solids in runoff from a 67-ha urban watershed by
looking at the behavior of two separate fractions (fine – less than 45 um; coarse – greater than 45 um). They found
that the two fractions behaved similarly; however, the pollutant loadings were different with the coarse fraction
needing a more intense storm to generate runoff containing that fraction.
The composition and morphology of colloidal materials entering an urban waterway (Brays Bayou, Houston, Texas)
during a storm event was investigated. Analyses of organic carbon, Si, Al, Fe, Cr, Cu, Mn, Zn, Ca, Mg, and Ca were
performed on the fraction of materials passing through a 0.45 μm filter. This fraction, traditionally defined as
"dissolved", was further fractionated by ultra-centrifugation into colloidal and dissolved fractions (Grout, et al.
1999).
Lloyd and Wond (1999) presented examples of particulate size fraction distributions for road and highway runoff
collected in Australia and compared the information with United States and European samples. They found that the
particle size distribution of suspended solids in stormwater runoff from roads and highways in Australia were
relatively finely graded. Sansalone and Hird (1999), in contrast, found that the particle sizes of stormwater
particulates investigated at a freeway site in Cincinnati, Ohio were much larger than typically found elsewhere. In
their samples, particles several hundred μm in size were common. They stressed the need to carefully collect
stormwater samples for particle size analyses considering the difficulty of representing large particles in samples
collected with automatic samplers. Pitt, et al. (1999), during pilot-scale testing of a critical-source-area-treatment
device, monitored particle size characteristics both in the influent and effluent. The parking lot stormwater had
median particle sizes ranging from 3 to 15 μm. They also monitored particle sizes from 75 source areas in the
Birmingham, Alabama, area as part of treatability tests during an earlier phase of this research and found similar
21
small-sized particles. Corsi, et al. (1999) measured stormwater particle sizes as part of a treatment system evaluation
at a public works yard in Milwaukee, Wisconsin, and found median particle sizes in the influent of about 18 μm.
Andral, et al. (1999) analyzed particle sizes and particle settling velocities in stormwater samples collected from
eight storm events from the A9 motorway in the Kerault Region of France. They concluded that to effectively treat
runoff, particles smaller than 50 μm in diameter (which represented approximately three-quarters of the particulates
analyzed, by weight) must be captured. The median particle size for their samples averaged about 15 μm. Settling
velocities of these particulates were also studied. The median settling velocities of the particulates smaller than 50
μm ranged from 2.5 to 3.3 m/h, while the larger particles between 50 to 100 μm in diameter had median settling
velocities ranging from 5.7 to 13 m/h. Krishnappan, et al. (1999) examined particle size distributions of suspended
solids in a wet detention pond. They used a submersible laser particle size analyzer that enabled them to examine the
particulate characteristics without disturbance by sampling. They found that the suspended solids were mostly
composed of flocs, with maximum sizes ranging from 30 (winter) to 212 μm (summer). They concluded that flocs in
the size range from 5 to 15 μm would settle faster than both smaller primary particles of higher density, and
somewhat larger flocs of lower density. The larger flocs were also found to be susceptible to break up by turbulence.
Organic matter in sediments from pipes and silt traps in combined sewers was divided into fractions with different
settling velocities. The largest fraction of organic material was found in the faster settling material, however the
faster settling material had a slower biodegradability than the slower settling fraction (Vollersten, et al., 1998). The
efficiencies estimated using particle size distribution and settling velocities as suggested in the Ministry of the
Environment and Energy Stormwater Management Practices Planning and Design Manual of Ont., Can., which is
used to design best management practices (BMP), were less than those estimated using the observed particle
distribution and settling velocities estimated from these sites. Suspended solid removal estimation based on the
manual would result in removal efficiencies more conservative than observed (Liang, et al. 1998).
Jones and Washburn (1998) mapped the three-dimensional distribution of dissolved and particulate components
associated with stormwater runoff in Santa Monica Bay, Calif. The three major particles types in the bay were
particles associated with the stormwater runoff, phytoplankton in the water column, and resuspended sediments.
Water quality and particle-size distribution were characterized from urban-stormwater runoff from two storms that
indicated potential relationships between zinc (Zn)/organic carbon and iron (Fe)/macrocolloid (0.45 μm — 20 μm)
pairs. Results also indicated that concentrations of particle ion number, organic carbon, suspended solids (SS), Fe,
and Zn increased during storms but showed no evidence of the “first flush” (Characklis and Wiesner, 1997).
Solids-settling characteristics are very important for designing many CSO and stormwater-sedimentation-control
facilities. Pisano (1996) summarized more than 15 years of settling data obtained in the United States, separated by
wastewater types.
Pollutant Associations with Different Particle Sizes
During analyses conducted by Johnson, et al. (2003), all samples were divided into three portions for particle size
analyses: >106 m, <0.45 m, and the intermediate range from 0.45 to 106 m. The Coulter Counter (Multi-Sizer 3)
results for the intermediate size range were compiled with results from the other two size ranges to obtain the
complete particle size distribution. The last section of this module describes a recommended analytical protocol for
particulate analyses. In a follow-up study to Johnson, et al. (2003), Morquecho (2005) examined about 10 separate
particle size categories for chemical analyses.
Most of the analytes had large decreases with filtration, especially for the more contaminated samples (turbidity,
phosphorus, phosphate, magnesium, chromium, copper, iron, lead, and zinc). Total solids and COD had much
smaller changes with filtration, with substantial fractions associated with the filterable (<0.45 m) fraction.
Constituents that did not change significantly with filtration included nitrate and sodium, as expected. Other analytes
(COD and cadmium) also had little change, except for a single sample in each case.
Table 8 summarizes the calculated “potency” factors for several size ranges for the heavy metals during these tests.
These are the masses of these constituents associated with the particulates, expressed as mg constituent per kg of
22
suspended solids. As expected, these values increase with decreasing particle size, except for iron that is seen to
generally remain constant. The iron measurements show that these stormwater particulates were about 1.5 to 3%
iron, by mass. There were also large zinc components for the smallest sizes. Typical soil values for iron are about
50,000 mg Fe per kg solids, somewhat higher than these observations, while observed copper, lead, and zinc are all
higher than typical soil values of about 100 mg/kg (Lindsay 1979).
Typical street dirt (mostly for the smallest <43 m size particles) contents of these heavy metals are about 100 to
500 mg/kg for copper, about 500 to 7,500 mg/kg for lead, and about 250 to 1,200 mg/kg for zinc (Bannerman, et al.
1983, Pitt 1979, Pitt 1985, Pitt and McLean 1986, Pitt and Sutherland 1982, and Terstrip, et al. 1982). The lead
values found were much less than the earlier street dirt values, due to the decreased use of lead in gasoline.
However, the copper and zinc values observed are much greater than the street dirt values.
In Toronto, Pitt and McLean (1986) found copper values of about 30 to 200 mg/kg for most residential and
commercial source area particulates (<125 m), but from 100 to over 1,000 mg/kg at industrial areas. Similar values
for residential and commercial area lead soil concentrations were generally from 200 to 1,000 mg/kg and slightly
higher in industrial areas. Zinc ranged from 50 to 2,000 mg/kg in all land use areas. Again, the observed copper and
zinc values were much greater than these previous measurements, while the lead values were much less.
Table 8. Summary Table Pollutant Associations for Different Particle Sizes (Morquecho 2005)
particle size (m)
>250
106 to 250
45 to 106
10 to 45
2 to 10
0.45 to 2
Copper
mg Cu/kg
COV
SS
50
na
2,137
1.45
1,312
1.16
735
0.97
4,668
1.60
2,894
1.21
Iron
mg Fe/kg
COV
SS
28,604
1.50
21,730
0.85
14,615
0.72
26,221
0.54
18,508
1.16
29,267
1.31
particle size (m)
>250
106 to 250
45 to 106
10 to 45
2 to 10
0.45 to 2
COD
mg COD/kg
COV
SS
672,000
1.4
373,000
1.3
586,000
1.0
629,000
1.7
1,940,000
0.7
2,330,000
1.6
Total Phosphorus
Mg TP/kg
COV
SS
6,440
1.2
17,200
3.4
2,160
0.9
6,640
0.9
78,800
1.3
57,500
1.8
Lead
mg Pb/kg
SS
117
375
226
229
868
199
COV
0.58
1.03
0.85
0.50
0.78
1.40
Zinc
mg Zn/kg
SS
266
3,486
2,076
1,559
13,641
13,540
COV
0.88
0.79
0.88
0.74
1.88
1.56
Randall, et al. (1982), recognized the strong correlation between pollutant removal effectiveness in wet detention
ponds and pollutant associations with suspended solids. High lead removals were related to lead's affinity for
suspended solids, while much smaller removals of BOD5 and phosphorus were usually obtained because of their
significant soluble fractions.
Wet detention ponds also are biological and chemical reactors. Changes in many pollutants can take place in the
water column or in the sediments of ponds. Dally, et al. (1983) monitored heavy metal forms in runoff entering and
leaving a wet detention pond serving a bus maintenance area. They found that metals entering the monitored pond
were generally in particulate (nonfilterable) forms and underwent transformations into filterable (smaller than 0.2
m in size) forms. The observed total metal removals by the pond were generally favorable, but the filterable metal
outflows were much greater than the filterable metal inflows. This effect was most pronounced for cadmium and
lead. Very little changes in zinc were found, probably because most of the zinc entering the pond was already in
filterable forms. These metal transformations may be more pronounced in wet detention ponds that in natural waters
because of potentially more favorable (for metal dissolution) pH and ORP conditions in wet pond sediments. Other
studies have found similar transformations in the forms and availability of nutrients in wet detention ponds, usually
depending on the extent of algal growth and algal removal operations.
23
Table 9 can be used to estimate the approximate controls for pollutants in wet detention ponds, related to particulate
removals. These ratios of pollutant removals to suspended solids removals are based on many field observations
(mostly from the NURP studies, EPA 1983) of detention pond performance and can vary significantly. Three
general groupings were identified: total lead and total copper were most efficiently removed, while organic nitrogen
was the least efficiently removed. Many of the nutrients showed “negative” removals during monitoring, possibly
because of biological cycling of the nutrients in the ponds. Wet detention ponds should not be expected to provide
significant removals of any pollutants in “soluble” forms (associated with very small particles, colloids, or truly
dissolved).
Table 9. Approximate Control of Stormwater Pollutants in Wet Detention Ponds
Constituent Group
Lead and copper
COD, BOD5, soluble and total phosphorus, nitrates, and zinc
Organic nitrogen
Control as a Fraction of
Suspended Solids Control
0.75 to 1.00+
0.6
0.4
Example: If 85% control of suspended solids, then:
Lead and copper: 0.75 to 1.0+ of 85% = 64 to 85+%
COD, etc.: 0.6 of 85% = 51%
Organic nitrogen: 0.4 of 85% = 34%
The relationship between solids retention and pollution retention is important for wet detention ponds. Becker, et al.
(1995) used settling column tests to measure the settling characteristics of different pollutants in sanitary sewage.
They found that the majority of the particulate fractions of COD, copper, TKN, and total phosphorus was associated
with particles having settling velocities of 0.04 to 0.9 cm/sec. Figure 8 is an example plot showing the relationship
of particulate COD and different settling velocity fractions.
Figure 8. COD and particulate settling velocity (Butler, et al. 1995).
Particulate pollutant strength (or potency factor) is the ratio of a particulate pollutant concentration to the suspended
solid concentration, expressed in mg/kg. The strengths of stormwater particulates were calculated for each pollutant
with a particulate form and plotted on a probability versus strength chart for the Madison, WI, data from House, et
al. (1993) (Figure 9 for Zn). All pollutants had higher outlet than inlet strength values due to preferential removals
of large particles in the detention pond, leaving relatively more small particles in the discharge water. The small
particles in stormwater have higher pollutant strengths than the large particles.
24
Figure 9. Particulate pollutant strengths for zinc (data from House, et al. 1993).
Vignoles and Herremans (1995) also examined the heavy metal associations with different particles sizes in
stormwater samples from Toulouse, France. They found that the vast majority of the heavy metal loadings in
stormwater were associated with particles less than 10 m in size, as shown on Table 10. They concluded that
stormwater control practices must be able to capture the very small particles.
Table 10. Percentages of Suspended Solids and Distribution of Heavy Metal Loadings Associated
with Various Stormwater Particulate Sizes (Toulouse, France) (Percentage associated with size
class, concentration in mg/kg).
Suspended
solids
Cadmium
Cobalt
Chromium
Copper
Manganese
Nickel
Lead
Zinc
>100 m
15%
50 to 100 m
11%
40 to 50 m
6%
32 to 40 m
9%
20 to 32 m
10%
10 to 20 m
14%
<10 m
35%
18 (13)
9 (18)
5 (21)
7 (42)
8 (86)
8 (31)
4 (104)
5 (272)
11 (11)
5 (16)
4 (25)
8 (62)
4 (59)
5 (27)
4 (129)
6 (419)
6 (11)
4 (25)
2 (26)
3 (57)
3 (70)
4 (31)
2 (181)
3 (469)
5 (6)
6 (20)
6 (50)
4 (46)
3 (53)
5 (31)
4 (163)
5 (398)
5 (5)
6 (18)
3 (23)
4 (42)
4 (54)
5 (27)
5 (158)
5 (331)
9 (6)
10 (22)
9 (39)
11 (81)
7 (85)
10 (39)
8 (247)
16 (801)
46 (14)
60 (53)
71 (134)
63 (171)
71 (320)
63 (99)
73 (822)
60 (1,232)
Source: Vignoles and Herremans (1995)
Table 11 lists the remaining concentrations for several monitored pollutants after controlling for particle sizes
ranging from 20 to 0.45 m (Morquecho 2005). The corresponding percentage reductions are shown in Table 12.
Some pollutants can be reduced by a reduction in particulates, such as suspended solids, total phosphorus and most
heavy metals. Other pollutants, such as nitrates, are reduced much less, even after filtration down to 0.45 m.
25
Table 11. Average Remaining Concentrations after Controlling for Different Sized Particles
(Morquecho 2005)
Total Solids
Suspended Solids
Turbidity (NTU)
Total-P
Total-N
Nitrate
Phosphate
COD
Ammonia
pH (pH units)
Cadmium (g/L)
Chromium (g/L)
Copper (g/L)
Iron (g/L)
Lead (g/L)
Zinc (g/L)
Residual Concentration after Removing all Particulates Greater than Size
Shown (mg/L, unless otherwise noted)
Unfiltered
20 m
5 m
1 m
0.45 m
178
106
102
86
84
94
22
18
2
0
53
30
24
4
2
0.38
0.12
0.07
0.04
0.03
2.13
1.50
1.25
1.38
1.63
0.6
0.60
0.60
0.53
0.50
0.8
0.23
0.18
0.15
0.10
99
51
48
47
52
0.26
0.17
0.14
0.12
0.11
7.35
7.45
7.66
7.39
7.53
4.9
3.9
3.8
3.8
3.8
15.4
4.7
3.8
2.8
2.5
24.6
18.3
16.2
17.8
15.6
2830
1350
1050
140
80
15.7
9.2
6.0
3.7
2.8
179
64
53
53
51
Table 12. Average Percentage Reduction in Pollutants after Controlling for Different Particle Sizes
(Morquecho 2005)
Total Solids
Suspended Solids
Turbidity
Total-P
Total-N
Nitrate
Phosphate
COD
Ammonia
Cadmium
Chromium
Copper
Iron
Lead
Zinc
Percent Pollutant Reduction after Removing all
Particulates Greater than Size Shown
20 m
5 m
1 m
0.45 m
40%
43%
52%
53%
76
81
98
100
43
55
92
96
68
82
89
92
30
41
35
23
0
0
12
17
71
78
81
88
48
52
52
47
35
46
54
58
20
22
22
22
69
81
82
84
26
34
34
37
52
63
95
97
41
62
76
82
64
70
70
72
Most well designed wet detention ponds remove most particulates down to about 1 to 5 m for long-term
conditions, depending on the rain conditions and drainage area. Smaller ponds may only be able to remove the
particulates down to about 20 m, while no pond can remove the filterable fraction by physical removal processes
alone. The following list shows which constituents can be controlled to different levels and the associated particulate
removal targets, based on many field observations of actual wet pond performance:
>90%
suspended solids (0.45 to 1 m)
turbidity (0.45 to 1 m)
total phosphorus (0.45 m)
iron (0.45 to 1 m)
26
80 to 89%
suspended solids (5 m)
turbidity (2 m?)
total phosphorus (1 to 5 m)
phosphate (0.45 to 1 m)
chromium (0.45 to 5 m)
lead (0.45 m)
70 to 79%
suspended solids (20 m)
turbidity (3 m?)
phosphate (5 to 20 m)
lead (1 m)
zinc (0.45 to 5 m)
all <50%
total nitrogen
nitrate
COD
cadmium
copper
Of course, samples for other locations and situations would likely result in different levels of control. However, the
predicted level of control associated with particulate control at least to 1 to 5 m are close to what is expected for a
well-designed wet detention pond designed for the control of these particulates, with the exception of copper which
would normally be better controlled than indicated with this data. These data enable us to specify the level of
treatment of particulates to obtain specific pollutant treatment goals.
Colloidal Analysis
Chelex-100 resin was added to portions of the filtered (0.45 m filter) stormwater samples collected in Birmingham
and Tuscaloosa, AL (Johnson, et al, 2003 and Morquecho 2005). After shaking for one hour, the samples were again
filtered (again using a 0.45 m filter). Ionic forms of the metals will bond to the Chelex, while colloidal bound
metals and metals strongly bound to ligands will remain in solution and be filtered. Therefore, the “after” Chelex
exposed concentrations reflects these bound metals, while the differences in the “before” and “after” exposure
concentrations reflects the portions associated with free ionic forms.
The Chelex-100 tests indicated that almost all of the filterable (through 0.45 m filters) heavy metals (and major
ions) were not associated with colloids and organo-metallic complexes, but occurred as free ions. The main
exception was cadmium, with only about 30% of the filterable cadmium in ionic forms, and copper with about half
of the filterable copper in ionic forms. The UV exposure tests used to indicate likely organo-metallic complexes
were quite uncertain due to the low concentrations of all complexed metals in the filterable sample fraction and the
imprecision of the toxicity tests at such low values.
Strengths of Major Constituent and Heavy Metal Associations with Particulates
Most of the heavy metals in stormwater are associated with the particulates larger than 0.45 m in size, although
some exceptions exist (usually zinc, and sometimes copper, can predominantly be associated with the filterable
fraction of stormwater). The fate of these particulate-bound heavy metals is of interest when designing stormwater
controls. Specifically, do these metals stay bound to the particulates under typical environmental conditions of pH
and Eh? This is of special interest when considering sediment in urban streams, and captured sediment in
stormwater controls. Additional tests were therefore conducted by Johnson, et al. (2003) and Morquecho (2005) to
examine the strengths of association between the pollutants and the particulates. It was expected that the filtered
27
concentrations after exposure to varying pH conditions would be intermediate between the original filterable
(filtered through 0.45 m filters and digested) and the initial total (unfiltered and digested) concentrations. If the
concentrations under all pH conditions were similar to the concentrations of the original filtered samples, then the
particulate bound metals are strongly associated with the particulates and are unlikely to enter the overlying water
column in a pond. However, if the concentrations approached the total metal concentrations, and were much higher
than the filtered metal concentrations, then the metals are weakly associated with the particulates and my become
disassociated during normal environmental conditions, with potentially adverse water quality consequences.
Experiments were conducted to examine the likelihood of the metals disassociating from the particulates under pH
conditions ranging from about 4 to 11. In addition, tests were conducted with both weak and strong acids. Related
tests were conducted to measure the disassociation potential of heavy metals (and nutrients) under aerobic and
anaerobic conditions having extreme Eh values. Figure 10 is an example plot for the copper tests.
Table 13 summarizes the results of these tests, indicating if the major constituents and heavy metals can be
disassociated from the particulates under different pH conditions. The major constituents (calcium, potassium and
sodium) had very little change for any pH exposure condition; all concentrations were very similar. Magnesium also
had several samples that had better retention at high pH, compared to low pH conditions. There was also no apparent
difference between exposure to the strong or the weak acids at the two low pH conditions for any constituent.
Figure 10. Leachability of copper under different pH conditions.
28
Table 13. Summary of Strengths of Associations of Major Constituents and Heavy Metals
with Particulates under Varying pH Exposure Conditions (Morquecho 2005)
Calcium
Magnesium
Potassium
Sodium
Cadmium
Chromium
Copper
Iron
Lead
Zinc
pH exposure conditions (after 3 days on shaking table)
pH = 4
pH = 5.5
neutral
pH = 9
pH = 11
dissolved1
dissolved
dissolved
dissolved
dissolved
weak2
moderate3
moderate
moderate
strong4
strong
strong
strong
strong
strong
dissolved
dissolved
dissolved
dissolved
dissolved
precipitated
precipitate
precipitated precipitated precipitated
5
d
strong
strong
strong
strong
strong
strong
strong
strong
strong
strong
strong
strong
strong
strong
strong
strong
strong
strong
strong
strong
weak
strong
strong
precipitated precipitated
1
dissolved constituents occur then the exposed concentrations were close to the original filterable concentration, and most of the
constituent was in the dissolved (filterable) fraction.
2
weak associations occur when the exposed concentrations are close to the original total concentration which was substantially
larger than the dissolved concentration.
3
moderate associations occur when the exposed concentrations are intermediate between the original total and filterable
concentrations.
4
strong associations occur when the exposed concentrations remain close to the original filterable concentration which is
substantially smaller than the total concentration.
5
precipitation may have occurred when the exposed concentrations are less than the original filterable concentration.
These tests indicate that the heavy metals of concern remain strongly bound to the particulates during long
exposures at the extreme pH conditions likely to occur in receiving water sediments. They will also likely remain
strongly bound to the particulates in stormwater control device sumps or detention pond sediments where
particulate-bound metals are captured. The associated tests examining metal binding to filtration media under
aerobic and anaerobic conditions also found that the heavy metals likely remain strongly associated with a variety of
organic and inorganic media under varying Eh conditions. However, those other tests did indicate that captured
nutrients are likely to come off the media under anaerobic conditions.
Sediment Transport in Storm Drainage Systems
Much research has been conducted on the transport, settling, and scour of gross solids in sanitary sewers and in
combined sewers, especially in Europe, during recent years (Ashley, et al. 1999; 2000; 2002; Butler and
Karunaratne 1995; Butler, et al. 1995; Cigana, et al. 1998a, 1998b, 1998c, 1999, 2000, 2001; plus reviews by
Pisano, et al. 1998, amongst others). However, relatively little research has been conducted concerning the fate of
larger particulates that enter separate stormwater drainage systems. Historical design approaches are intended to
minimize particulate deposition in sewerage, and usually present a minimum pipe slope or a minimum velocity
objective. If followed, these guidelines usually result in minimal maintenance problems associated with particulate
accumulations in the storm drainage system. However, it is still important to minimize erosion sources in the
watershed. In addition, catchbasin sumps have also been used to trap the larger particulates that enter storm drainage
inlets. Therefore, particulate transport in separate storm drainage has not been considered to be a significant problem
for public works managers. However, much more is needed to be known about particulate transport in stormwater
drainage systems when conducting stormwater quality investigations, especially when examining the effects of
source area controls on outfall quality. Prior studies have conducted mass balances in urban drainage systems and
have found significant accumulations of solids in the drainage systems. These accumulations are mostly of the
largest particulates that enter the drainage system, effectively preventing these from being discharged to the
receiving waters. Many source area stormwater quality controls, such as street cleaning and the use of catchbasins,
also preferentially remove the largest particulates. Simple modeling of the benefits of these controls therefore
typically leads to inaccurate conclusions concerning reduced discharges of these particulates, and associated
pollutants. The pollutants being removed would not likely be effectively transported through the drainage system,
but would instead accumulate. In addition, source area and inlet samples frequently indicate much larger amounts of
29
large particulates than would be discharged to the receiving water. This also misleads management strategies
pertaining to stormwater quality. This module section presents observations from an urban area mass balance
investigation along with some traditional particulate transport approaches in an attempt to explain some of these
observations.
McIlhatton, et al. (2002) reviewed the influence of solids eroded from the bed of the combined sewer on receiving
waters and treatment plants using data obtained from field studies carried out in the main Dundee interceptor sewer
in Scotland. In addition, the paper described some of the methods used to investigate the pollutant characteristics
associated with the solids erosion in combined sewers. Sakrabani, et al. (2002) described the efforts using an
endoscope that allowed visualization of combined sewer deposits (Near-Bed Solids) in a non-destructive way. Based
on this work, several models have been proposed to estimate near-bed solids and predict the pollutant loads of
consequence from CSOs. Pollert and Stansky (2002) reported on the use of computational techniques (a 1-D
MOUSE model and the 3-D FLUENT model) to evaluate the separation efficiency of suspended solids in a CSO
system. A combination of the two models was used to describe the CSO behavior.
Modeling the transport of sediment and other debris in sewers was investigated by Babaeyan-Koopaei, et al. (1999).
Using the correct velocity distributions was found to be critical in order to obtain accurate predictions of sediment
transport. A sensitivity analysis investigated the influence of some important parameters involved in the model,
especially the drag coefficient, the lift coefficient, solid density, and pipe roughness. Arthur, et al. (1999) proposed a
new design approach to minimize sedimentation in sewers. They also compared the results of laboratory
investigations with real sewer conditions. Johnstone, et al. (1999) described on-going research concerning the
disposal of large sanitary solids in combined sewers, assessing the relative sustainability of conventional disposal
methods using an integrated, holistic approach, incorporating economics, sociology, life cycle data, and a risk
assessment. They also described a project designed to study the behavior of the sanitary solids in the sewerage
system through laboratory, field, and modeling studies. Skipworth, et al. (1999) sampled and analyzed combined
sewer sediment deposits and found coarse, loose, granular, predominantly mineral material that was overlain by a
mobile, fine-grained cohesive-like sediment deposit in the invert of pipes. The erosion of this more mobile fraction
was identified as the major source of the first flush of pollutants associated with CSO events. They presented a new
approach to model the erosion and subsequent transport of these mobile sediments.
Rushforth, et al. (1999) examined the relationships between the erosion of in-sewer organic deposits and the
sediment composition. They found that sewer sediments consist of mixtures of organic and inorganic material and
exhibit a much wider range of particle sizes and densities than typically assumed. Specifically, the characteristics of
the fine organic sediment found in combined sewer deposits fall outside the applicable range of grain size and
density for typically used sediment transport models. Their laboratory experiments showed that the addition of
granular material in sewer deposits significantly increases the amount of organic material eroded, compared to a
deposit composed entirely of organic material.
Background
Modeling solids transport in sewers.
Many computational models are available to simulate the flow and transport of suspended solids and other water
quality pollutants through urban drainage systems to wastewater treatment facilities and CSOs. However, models
capable of predicting the quantity and temporal distribution of gross solids entering treatment works and CSOs are
less available. Digman, et al. (2002) developed a model to predict the movement of gross solids in a combined
sewerage system. The Gross Solids Simulator (GSS) links directly with Hydroworks to obtain network data and
flow depth and velocity. The model also uses a series of diurnal solids profiles based on work performed at the
Imperial College in London. The GSS has been calibrated using an extensive set of data sampled from combined
sewerage systems. In a separate study, results from a thorough analysis of the water quality modeling components of
MOUSE and HydroWorks/InfoWorks CS suggests that results obtained for pollutant transport in the sewers is not as
accurate as for hydrodynamics (Bouteligier, et al. 2002). Consequently, caution should be exercised when using the
water quality modules of the two software packages.
Zug, et al. (2002) are creating an integrated model (sewer networks, storage tank, and wastewater treatment plant
(WWTP)) for the Grand Couronne sewer system and parts of the Berlin sewer system. The integrated model has
30
been built, calibrated, and validated for hydraulic and pollution load prediction during dry and wet weather. Mark, et
al. (2002) used integrated modeling of the drainage system, WWTP, and receiving water to evaluate real time
control and alternative WWTP operations to protect beachfront water quality. The integrated modeling used
MOUSE TRAP to simulate the drainage system, historical data to represent the WWTP operation, and MIKE21 to
represent the beach water quality in two dimensions. An assessment of the benefits of an integrated urban drainage
modeling approach linking the operation of the sewer system, WWTP, and receiving water was evaluated by Erbe,
et al. (2002). Mailhot, et al. (2002) evaluated the impacts of watershed management and wastewater treatment
activities on water quality of the Chaudiere River using an integrated modeling system. Wolff, et al. (2002) assessed
the use of hydraulic models of the collection system coupled with steady state models of the WWTP hydraulics.
Willems and Berlamont (2002b) used deterministic and probabilistic modeling to simulate the impact of combined
urban drainage and WWTP system discharges on receiving water quality.
Effects of Catchbasin and Street Cleaning
A study was conducted in Bellevue, Washington (Pitt 1985), in two mixed residential and commercial study areas,
as part of the Nationwide Urban Runoff Program (EPA 1983). One task of this research included the monitoring of
catchbasins, simple inlets, man-holes, and sewerage sediment accumulations at more than 200 locations for a period
of three years. The sediment in the catchbasins and the sewerage was found to be the largest particles that were
washed from the streets. The sewerage and catchbasin sediments had a much smaller median particle size than the
street dirt and were therefore more potentially polluting than the particulates that can be removed by street cleaning.
Cleaning catchbasins twice a year was found to allow the catchbasins to capture particulates most effectively. This
cleaning schedule was found to reduce the total residue and lead urban runoff yields at the outfalls by between 10
and 25 percent, and COD, total Kjeldahl nitrogen, total phosphorus, and zinc by between 5 and 10 percent (Pitt and
Shawley 1982).
This research examined two study areas, Lake Hills and Surrey Downs, both similar medium density residential
areas. Each study area was examined with four separate experimental conditions: no controls, street cleaning alone,
catchbasin cleaning alone, and both street cleaning and catchbasin cleaning together. This research was therefore
conducted in a replicated complete block design, allowing runoff quality comparisons between periods having these
different public works practices. The eight experimental categories were as follows:
1. Lake Hills, Active CB, No SC (catchbasins were accumulating material, but no street cleaning operations
were being conducted during this project period).
2. Lake Hills, Active CB, SC (catchbasins were accumulating material, and street cleaning operations were
being conducted during this project period).
3. Lake Hills, Full CB, No SC (catchbasins were full and not accumulating material, and no street cleaning
operations were being conducted during this project period).
4. Lake Hills, Full CB, SC (catchbasins were full and not accumulating material, street cleaning operations
were being conducted during this project period).
5. Surrey Downs, Active CB, No SC (catchbasins were accumulating material, but no street cleaning
operations were being conducted during this project period).
6. Surrey Downs, Active CB, SC (catchbasins were accumulating material, and street cleaning operations were
being conducted during this project period).
7. Surrey Downs, Full CB, No SC (catchbasins were full and not accumulating material, and no street cleaning
operations were being conducted during this project period).
8. Surrey Downs, Full CB, SC (catchbasins were full and not accumulating material, street cleaning operations
were being conducted during this project period).
31
Catchbasins were cleaned and surveyed at the beginning of the project. The accumulation of material was then
monitored through periodic measurements. The project periods were therefore categorized as “active” or “full.” The
active periods were when accumulation was taking place in the catchbasins, while the full periods were when the
catchbasins were at equilibrium, with no additional accumulation of material.
The following are Student t test results to measure the significance of the difference between selected data groups
for outfall total solids concentrations. There would have to be a 50 to 75% difference between the sample means of
the two categories to identify a significant difference, with 10 to 15 storms representing each of the two categories
for each test site, using a power of 80%, and assuming a typical COV of about 0.75 (Burton and Pitt 2002). P values
smaller than 0.05 are usually considered as being significantly different (at the 95% confidence level), while larger P
values indicate that not enough data are available to distinguish the data groups at the measured differences.
Student’s t-test results:
2 vs. 6: both street and catchbasin cleaning in both areas, LH vs. SD
P value: 0.71 (not enough data to detect a difference)
3 vs. 7: nothing in both areas, LH vs. SD
P value: 0.031 (significantly different)
2 vs. 3 LH both street and catchbasin cleaning vs. nothing
P value: 0.037 (significantly different)
6 vs. 7 SD both street and catchbasin cleaning vs. nothing
P value: 0.99 (not enough data to detect a difference)
When both street and catchbasin cleaning was being conducted in both areas, the outfall total solids concentrations
appeared to be the same (as expected). However, when no controls were in use in either area, the outfall total solids
concentrations were significantly different (Lake Hills had lower total solids concentrations compared to Surrey
Downs), which was not expected. When both street and catchbasin cleaning was conducted in Lake Hills, the outfall
total solids concentrations were significantly larger than when no cleaning was being conducted, which also was not
expected. In Surrey Downs, no differences were detected when cleaning was conducted compared to no cleaning.
These results are counter-intuitive. The hypothesis was that the two watersheds would behave in a similar manner
when similar activities were being conducted in each, and that the cleaning would reduce the outfall total solids
discharges. Over the years, a number of reasons have been given for the observed odd behavior. Older street
cleaning equipment was not very efficient in removing the particles that are washed off, and in fact, have been found
to actually remove the larger particles that actually armour the finer materials, potentially increasing the solids
discharges. However, the catchbasins are removing particles that have washed off the watershed area and have been
transported to the drainage system, but this material likely would not have been transported all the way to the outfall.
Ashley, et al. (1999, 2000, 2002) has extensively researched the transport of solids in combined sewerage.
Unfortunately, similar information is currently lacking for separate storm drainage. The initial objective for the use
of catchbasin sumps was to reduce the accumulation of coarse debris in the sewerage. These Bellevue tests seem to
indicate the substantial benefit of the removal of this material that may otherwise cause potential flow obstruction
problems in the drainage system. However, it is quite likely that this large material would rarely flow completely to
the outfalls, at least under the relatively mild Bellevue rain conditions and during the time frame of this study.
Settling of Particulates in Flowing Water in Storm Drainage Systems
After the stormwater particulates enter the storm drainage, they will tend to settle as they flow towards the outfall to
the receiving water. If they settle slowly, such as occurs for small particles, they will remain suspended and not
become part of the bed load or sediment in the sewerage. However, if they settle to the bottom of the pipe before
reaching the outfall, they may become part of the bed load which will bounce along the pipe bottom, or become
trapped with other settled debris. When the flow stops, the sediments will tend to dry and become more
consolidated. The next runoff event may cause some of this settled material to become resuspended and may move
32
towards the outfall. Therefore, there are three phases to particulate transport in storm drainage systems: 1) settling of
the particulates in the flowing water, 2) movement as bed load during the event, 3) accumulating as sediment and
potentially subsequent scour. The following discussion describes these sediment transport phases.
Table 14 presents settling conditions for particulates moving in pipes ranging from 1 to 5 ft in diameter, and for flow
depths ranging from 10% to 100% of the pipe diameters. Pipe slopes ranging from 0.1 to 2% and particles from 1 to
10,000 µm, all with specific gravities of 2.5, are used in these calculations. This table shows the distances the
particles would travel before they would settle to the bottom of the pipe, if starting from the surface of the flow,
using Manning’s equation to calculate the stormwater velocity, and the combination of Stoke’s and Newton’s laws
for settling rates. A particle settling to the pipe bottom doesn’t imply that the particles would be permanently trapped
as sediment, but the particles may move (relatively slowly) as part of a mobile bedload. Obviously, the flow
distances required for settling for the smallest particles are very long and would remain suspended. Some of the 100
µm particle flow conditions are shown to result in relatively short settling distances, while most of the 1,000 and
10,000 µm particles would most certainly settle to the pipe bottom before reaching the outfall. If the flow and
associated bed load movement stops before the particles reach the outfall, they will form a more compact sediment
requiring more energy during subsequent rains to scour.
33
Table 14. Settling Distance (ft) for Particles Flowing in Pipes having Various Diameters and Slopes (n=0.013)
1 µm particles
(2.5 specific gravity)
10 µm particles
(2.5 specific gravity)
Pipe
Diameter
(ft)
flow
depth/pipe
diameter
ratio
1
0.1
>10,000
>10,000
>10,000
>10,000
180
390
550
780
1
0.3
>100,000
>100,000
>100,000
>100,000
1000
2300
3200
4600
1
0.5
>100,000
>100,000
>100,000
>100,000
2200
4895
6900
>10,000
1
0.8
>100,000
>100,000
>100,000
>100,000
4000
8929
>10,000
>10,000
1
1
>100,000
>100,000
>100,000
>100,000
4400
9790
>10,000
>10,000
1.5
0.1
>10,000
>10,000
>100,000
>100,000
350
770
1100
1500
1.5
0.3
>100,000
>100,000
>100,000
>100,000
2000
4500
6400
9000
1.5
0.5
>100,000
>100,000
>100,000
>100,000
4300
9629
>10,000
>10,000
1.5
0.8
>100,000
>100,000
>100,000
>100,000
7900
>10,000
>10,000
>10,000
1.5
1
>100,000
>100,000
>100,000
>100,000
8600
>10,000
>10,000
>10,000
2
0.1
>10,000
>100,000
>100,000
>100,000
560
1300
1800
2500
2
0.3
>100,000
>100,000
>100,000
>100,000
3300
7300
>10,000
>10,000
2
0.5
>100,000
>100,000
>100,000
>100,000
7000
>10,000
>10,000
>10,000
2
0.8
>100,000
>100,000
>100,000
>100,000
>10,000
>10,000
>10,000
>10,000
2
1
>100,000
>100,000
>100,000
>100,000
>10,000
>10,000
>10,000
>10,000
3
0.1
>100,000
>100,000
>100,000
>100,000
1100
2500
3500
4900
3
0.3
>100,000
>100,000
>100,000
>100,000
6400
>10,000
>10,000
>10,000
3
0.5
>100,000
>100,000
>100,000
>100,000
>10,000
>10,000
>10,000
>10,000
3
0.8
>100,000
>100,000
>100,000
>100,000
>10,000
>10,000
>10,000
>100,000
3
1
>100,000
>100,000
>100,000
>100,000
>10,000
>10,000
>10,000
>100,000
5
0.1
>100,000
>100,000
>100,000
>100,000
2600
5800
8100
>10,000
5
0.3
>100,000
>100,000
>100,000
>100,000
>10,000
>10,000
>10,000
>10,000
5
0.5
>100,000
>100,000
>100,000
>100,000
>10,000
>10,000
>100,000
>100,000
5
0.8
>100,000
>100,000
>100,000
>100,000
>10,000
>100,000
>100,000
>100,000
5
1
>100,000
>100,000
>100,000
>100,000
>10,000
>100,000
>100,000
>100,000
0.001
slope
0.005
slope
0.01
slope
0.02
slope
0.001
slope
0.005
slope
0.01
slope
0.02
slope
34
Table 14. Settling Distance (ft) for Particles Flowing in Pipes having Various Diameters and Slopes (n=0.013) (cont.)
100 µm particles
(2.5 specific gravity)
1,000 µm particles
(2.5 specific gravity)
Pipe
Diameter
(ft)
flow
depth/pipe
diameter
ratio
1
0.1
1.9
4.4
6.2
8.7
1
0.3
11
25
36
51
1
0.5
24
54
77
1
0.8
44
99
140
1
1
49
110
150
1.5
0.1
3.8
8.6
1.5
0.3
22
50
1.5
0.5
48
1.5
0.8
0.001
slope
0.005
slope
0.01
slope
0.02
slope
0.001
slope
10,000 µm particles
(2.5 specific gravity)
0.005
slope
0.01
slope
0.02
slope
0.001
slope
0.005
slope
0.01
slope
0.02
slope
0.1
0.2
0.3
0.4
0.0
0.0
0.1
0.1
0.5
1.1
1.6
2.3
0.1
0.2
0.3
0.5
110
1.1
2.4
3.5
4.9
0.2
0.5
0.7
1.0
200
2.0
4.5
6.3
8.9
0.4
0.9
1.3
1.8
220
2.2
4.9
6.9
10
0.4
1.0
1.4
2.0
12
17
0.2
0.4
0.5
0.8
0.0
0.1
0.1
0.2
71
100
1.0
2.3
3.2
4.5
0.2
0.5
0.6
0.9
110
150
210
2.2
4.8
6.8
10
0.4
1.0
1.4
1.9
87
200
280
390
3.9
8.8
12
18
0.8
1.8
2.5
3.5
1.5
1
96
210
300
430
4.3
10
14
19
0.9
1.9
2.7
3.9
2
0.1
6.2
14
20
28
0.3
0.6
0.9
1.2
0.1
0.1
0.2
0.2
2
0.3
36
81
120
160
1.6
3.6
5.2
7.3
0.3
0.7
1.0
1.5
2
0.5
77
170
250
350
3.5
7.8
11
16
0.7
1.6
2.2
3.1
2
0.8
140
320
450
630
6.4
14
20
28
1.3
2.8
4.0
5.7
2
1
160
350
490
690
7.0
16
22
31
1.4
3.1
4.4
6.2
3
0.1
12
27
39
55
0.5
1.2
1.7
2.5
0.1
0.2
0.3
0.5
3
0.3
71
160
230
320
3.2
7.2
10
14
0.6
1.4
2.0
2.9
3
0.5
150
340
480
680
6.9
15
22
31
1.4
3.1
4.3
6.1
3
0.8
280
620
880
1200
13
28
40
56
2.5
5.6
7.9
11
3
1
310
680
960
1400
14
31
43
61
2.7
6.1
8.7
12
5
0.1
29
64
90
130
1.3
2.9
4.1
5.8
0.3
0.6
0.8
1.2
5
0.3
170
370
530
750
7.5
17
24
34
1.5
3.4
4.8
6.7
5
0.5
360
800
1130
1600
16
36
51
72
3.2
7.2
10
14
5
0.8
650
1500
2100
2900
29
66
93
130
5.9
13
19
26
5
1
710
1600
2300
3200
32
72
100
140
6.4
14
20
29
35
Resuspension of Settled Particulates in Storm Drainage
This discussion presents some particulate transport information that can be used to predict if settled particles
forming a sediment in a pipe may be resuspended or scoured during subsequent events. This information does not
allow predictions to be made concerning the accumulation of particulates in the sediment, only the likelihood that
previously settled material may scour.
Allowable Velocity and Shear Stress
Allowable Velocity Data. The concept of allowable velocities for various soils and materials dates from the early
days of hydraulics. Table 15 is an example of allowable velocities from U.S. Bureau of Reclamation research
(Fortier and Scobey 1926, reprinted by McCuen 1998), that also shows the corresponding allowable shear stresses
and Manning’s roughness values. If these velocities are exceeded for an extended period, it is assumed that the
channel lining material can become unstable. These values are not directly applicable to pipe flows, but the typically
used maximum velocity of about 3 ft/sec for storm drainage design is similar to the values for stiff clays and silts. It
is interesting that clays can withstand higher velocities than sands.
Table 15. Maximum Permissible Velocities and Corresponding Unit Tractive Force (Shear Stress) (U.S.
Bureau of Reclamation research, Fortier and Scobey 1926)
Clear Water
Material
n
V
(ft/sec)
1.50
1.75
2.00
2.00
2.50
2.50
3.75
3.75
6.00
2.50
3.75
4.00
4.00
5.00
o
(lb/ft2)
0.027
0.037
0.048
0.048
0.075
0.075
0.26
0.26
0.67
0.075
0.38
0.43
0.30
0.91
Water Transporting
Silts
V
(ft/sec)
2.50
2.50
3.00
3.50
3.50
3.50
5.00
5.00
6.00
5.00
5.00
5.50
6.00
5.50
Fine sand, colloidal
0.020
Sandy loam, noncolloidal
0.020
Silt loam, noncolloidal
0.020
Alluvial silts, noncolloidal
0.020
Ordinary firm loam
0.020
Volcanic ash
0.020
Stiff clay, very colloidal
0.025
Alluvial silts, colloidal
0.025
Shales and hardpans
0.025
Fine gravel
0.020
Graded loam to cobbles when noncolloidal
0.030
Graded silts to cobbles when noncolloidal
0.030
Coarse gravel, noncolloidal
0.025
Cobbles and shingles
0.035
Note:
 an increase in velocity of 0.5 ft/sec can be added to these values when the depth of water is greater than 3 ft.
 a decrease in velocity of 0.5 ft/sec should subtracted then the water contains very coarse suspended sediments.
 for high and infrequent discharges of short duration, up to 30% increase in velocity can be added
Colloidal
o
(lb/ft2)
0.075
0.075
0.11
0.15
0.15
0.15
0.46
0.46
0.67
0.32
0.66
0.80
0.67
1.10
Allowable Shear Stress Data. By the 1930’s, boundary shear stress (sometimes called tractive force) was generally
accepted as a more appropriate criterion than allowable velocity for channel stability. The average boundary shear
stress in uniform flow is calculated by
 o  RS
(lb/ft2)
where:
γ = specific weight of water (62.4 lbs/ft3)
R = hydraulic radius (ft)
S = hydraulic slope (ft/ft)
Flow characteristics predicting the initiation of motion of sediment in noncohesive materials are usually presented in
nondimensional form in the Shield’s diagram (Figure 11). This diagram indicates the initial movement, or scour, of
36
noncohesive uniformly graded sediments on a flat bed. The diagram plots the Shield’s number (or mobility number),
which combines shear stress with grain size and relative density, against a form of the Reynolds number that uses
grain size as the length variable. The ASCE Sedimentation Manual (1975) uses a dimensionless parameter, shown
on Figure 11, to select the dimensionless stress value. This value is calculated as:
 
d  s
0.1  1 gd 
  
 
0.5
where:
d = particle diameter (meters)
g = gravitational constant (9.81 m/sec2)
 = kinematic viscosity (1.306 x 10 –6 m2/sec for 10oC)
s = specific gravity of the solid
 = specific gravity of water
A series of parallel lines on Figure 11 represent these calculated values. The dimensionless shear stress value (*) is
selected where the appropriate line intersects the Shield’s curve. The critical shear stress can then be calculated by:
 c   *  s   d
Figure 11. Shield’s diagram for dimensionless critical shear stress (COE 1994).
37
An example evaluation is given by the COE (1994) in their assessment manual. In their example, the use of the
Shield’s diagram is shown to likely greatly over-predict the erodibility of the channel bottom material. The expected
reason they give is that the Shield’s diagram assumes a flat bottom channel and the total roughness is determined by
the size of the granular bottom material. The actual Manning’s roughness value is likely much larger because it is
largely determined by bed forms, channel irregularities, and vegetation, and not grain size. They recommend, as a
more realistic assessment, that empirical data based on field observations be used. In the absence of local data, they
present Figure 12 (from Chow 1959) for applications for channels in granular materials. This figure shows the
permissible unit tractive force (shear stress) as a function of the average particle diameter, and the fine sediment
content of the flowing water.
Figure 12. Allowable shear stresses (tractive forces) for canals in granular materials (U.S. Bureau of
Reclamation, reprinted in Chow 1959).
Table 16 shows calculated shear stresses, velocities, and discharge quantities for various pipe conditions. Also
shown are the estimated maximum particles sizes that would not be scoured during these flow conditions. For the
smallest slopes, almost all settled particles would likely remain and not be scoured, while almost all unconsolidated
38
sediments would be scoured for pipe greater than 1% in slope. This table shows that pipe conditions resulting in at
least 3 ft/sec stormwater velocities would also have shear stresses of at least 0.08 lb/ft 2. This shear stress would
likely cause scour of particles up to 2,000 µm in size for water having a low content of fine sediment. If the water
had a high content of fine sediment, the maximum particles likely to be scoured from the sediment may be only
about 100 µm in size. If the sediment was somewhat consolidated (as expected to occur during dry periods between
runoff events), than the necessary shear stress to cause sediment scour would be substantially greater, as noted in the
following discussion.
39
Table 16. Calculated Shear Stress and Particle Scour
Pipe
Dia.
(ft)
flow
depth/pipe
dia. ratio
shear stress lb/ft2), velocity (ft/sec), and discharge (ft3/sec) for different slopes:
Max.
Max.
particle
particle
shear at
size not
shear at size not
0.001
scoured
velocity
discharge
0.005
scoured
velocity
slope
(µm)
at 0.001 at 0.001
slope
(µm)
at 0.005
1
0.1
0.0041
<100
0.57
0.028
0.020
1
0.3
0.011
<100
1.1
0.21
1
0.5
0.016
<100
1.4
0.55
1
0.8
0.019
<100
1.6
1
1
0.016
<100
1.5
0.1
0.0061
1.5
0.3
0.016
1.5
0.5
1.5
0.8
discharge
at 0.005
shear
at
0.01
slope
velocity
at 0.01
discharge
at 0.01
shear
at 0.02
slope
velocity
at 0.02
discharge
at 0.02
<100
1.3
0.063
0.041
1.8
0.089
0.081
2.6
0.13
0.053
250
2.5
0.48
0.11
3.5
0.68
0.21
5.0
0.96
0.078
1,500
3.2
1.29
0.12
4.5
1.7
0.31
6.4
2.5
1.2
0.094
2,800
3. 7
2.6
0.20
5.2
3.6
0.38
7.3
5.5
1.4
1.1
0.078
1,500
3.2
2.5
0.12
4.5
3.6
0.31
6.4
5.1
<100
0.75
0.083
0.030
<100
1.7
0.19
0.061
2.4
0.26
0.12
3.4
0.37
<100
1.5
0.63
0.080
1,500
3. 3
1.4
0.16
4.6
2.0
0.32
6.6
2.8
0.023
<100
1. 9
1.63
0.12
3,500
4.2
3.6
0.23
6.0
5.2
0.47
8.4
7.3
0.028
<100
2.1
3. 4
0.14
5,000
4.8
7.6
0.28
6. 8
11
0.57
9.6
15
1.5
1
0.023
<100
1. 9
3.3
0.12
3,500
4.2
7.4
0.23
6.0
11
0.47
8.4
15
2
0.1
0.0081
<100
0.91
0.18
0.041
<100
2.03
0.40
0.081
2.9
0.57
0.16
4.1
0.80
2
0.3
0.021
<100
1. 8
1.4
0.11
3,300
4.0
3.1
0.21
5.6
4.3
0.42
8.0
6.1
2
0.5
0.031
<100
2.3
3.5
0.16
4,500
5.1
7.98
0.31
7.2
11
0.62
10
16
2
0.8
0.038
<100
2.6
7.3
0.19
>5,000
5.8
16
0.38
8.2
23
0.76
12
33
2
1
0.031
<100
2.3
7.2
0.16
4,500
5.1
16
0.31
7.2
23
0.62
10
32
3
0.1
0.012
<100
1.2
0.53
0.061
700
2. 7
1.2
0.12
3.8
1. 7
0.24
5. 4
2.4
3
0.3
0.032
<100
2.3
4.0
0.16
5,000
5.2
9.0
0.32
7.4
13
0.65
10
18
3
0.5
0.047
<100
3.0
10
0.23
>5,000
6.7
23
0.47
9.5
33
0.94
13
47
3
0.8
0.057
450
3.4
22
0.28
>5,000
7.7
48
0.57
11
68
1.1
15
97
3
1
0.048
<100
3.0
21
0.23
>5,000
6.7
47
0.47
9. 5
67
0.94
13
95
5
0.1
0.020
<100
1.7
2.1
0.10
3,000
3.8
4.6
0.20
5.4
6.5
0.41
7.5
9.3
5
0.3
0.053
400
3.3
16
0.27
>5,000
7.4
35
0.53
10
50
1.1
16
70
5
0.5
0.078
1,500
4.2
41
0.39
>5,000
9.4
92
0.78
13
130
1.6
19
180
5
0.8
0.094
2,800
4.8
84
0.47
>5,000
114
190
0.94
15
270
1.9
22
380
5
1
0.078
1,500
4.2
83
0.39
>5,000
9.4
190
0.78
13
260
1. 6
19
370
assuming n=0.013
water with low content of fine sediment in water (if high content, then require higher shear stress)
40
Erodability of Previously Settled Material after Consolidation. Figure 13 is an example of allowable shear stresses
for a range of cohesive materials having a range of void ratios corresponding to varying amounts of consolidation.
Again, the COE recommends that local field observation or laboratory testing results be given preference. If the void
ratio was about 0.4, corresponding to a compact sediment, the shear stress would have to be greater than about 0.3
lb/ft2 to affect particles larger than clays. This shear stress may only occur for larger and full-flowing pipes having
2% slopes (and about 7 ft/sec velocities).
Figure 13. Example of allowable shear stresses (tractive forces) for cohesive materials (COE 1994) (a Leon
clayey soil is hardpan. Hardpan is a condition of the soil or subsoil in which the soil grains become
cemented together by such bonding agents as iron oxide and calcium carbonate, forming a hard, impervious
mass).
Criteria to Ensure Self-Cleaning in Sewerage
Pisano, et al. (1998) described the mechanisms associated with deposition and scour of sediment in sewerage. The
following paragraphs are summarized from that report.
Not all of the particles of a given size at the flow/sediment boundary dislodge and move at the same time because
the flow is turbulent and contains short term fluctuations in velocity. The limiting condition, below which sediment
movement is negligible, known as the threshold of movement, is usually defined in terms of either the critical bed
shear stress or the critical erosion velocity.
41
Once sediment is entrained in flowing water, it may travel in the sewer in one of two basic ways: 1) Finer, lighter
material tends to travel in suspension, while 2) heavier material travels in a rolling, sliding mode as bedload. In the
transport of suspended sediment, there is a continuous exchange between particles settling out and those being
eroded and re-entrained back into the flow. Under certain conditions, fine grained and organic particles can form a
highly concentrated mobile layer of “fluid mud” near the pipe invert (Ashley and Crabtree 1992).
If the flow velocity or turbulence level decreases, there will be a net reduction in the amount of sediment held in
suspension. Below a certain limit, the sediment will form a bed, with transport occurring only in the surface layer. If
the flow velocity is further decreased, sediment transport will cease completely. The flow conditions necessary to
prevent deposition depend on the pipe size and on properties of the sediment, mostly particle size and specific
gravity. Flow velocities needed to entrain sediment tend to be higher than those at which deposition occurs.
Sewer Self Cleansing Criteria. Average shear stress is an important parameter in the criteria for sewer selfcleansing. The average shear stress is the amount of force the fluid exerts on the wetted perimeter of the pipe.
Another important parameter is the bed shear stress which is the amount of force the fluid exerts on the bed of
sediment in the pipe. Bed shear stress is related to bed load scour and movement.
Historically, the design of drainage systems has included the prevention of deposition of sediments in pipes. These
conditions are based on a minimum velocity of flow, or a minimum shear stress that the flow should exert on the
walls of the pipe to maintain self-cleansing conditions. The minimum velocity of flow, or the minimum shear stress,
corresponds to a particular depth of flow, or with a particular frequency of occurrence. Available design criteria
were reported by the Construction Industry Research and Information Association (CIRIA), and a summary of these,
outlined by Nalluri and Ab Ghani (1996), are shown in Tables 17 and 18.
Table 17. Minimum Velocity Criteria
Source
Country
American Society of Civil Engineers (1970) USA
British Standard (1987)
UK
Bielecki (1982)
Germany
Sewer Type
Sanitary
Storm
Sanitary
Combined
Not noted
Minimum
Velocity
(m/s) (ft/sec)
0.6
2.0
0.9
3.0
1.0
3.3
1.0
3.3
1.5
5.0
Pipe
Conditions
Full/Half-full
Full/Half-full
Full
Full
Full
Table 18. Minimum Shear Stress Criteria
Source
Country
Lysne (1969)
ASCE (1970)
Yao (1974)
USA
USA
USA
Maguire
Bischof (1976)
Sewer Type
Not noted
Not noted
Storm
Sanitary
UK
Not noted
Germany Not noted
Minimum Shear Stress
(N/m2 )
2-4
1.3-12.6
3.0-4.0
1.0-2.0
6.2
2.5
(lb/ft2)
0.04 – 0.08
0.027 – 0.26
0.061 – 0.083
0.021 – 0.042
0.13
0.052
Pipe
Conditions
Not noted
Not noted
Not noted
Not noted
Full/Half-full
Not noted
These criteria do not take into account any of the characteristics of the sediment, of the suspended sediment
concentration, the bed load, or of any cohesion between the sediment particles. Nonetheless, minimum velocity and
minimum shear stress levels of 1 m/s (3.3 ft/sec) and 2 N/m2 (0.04 lb/ft2) are commonly accepted criteria. However,
recent work by many researchers has shown that a single value of minimum velocity or shear stress cannot
adequately describe the self cleansing conditions in all pipes of different size, roughness and gradient for a range of
sediment characteristics and flow conditions.
In practice, sewer pipes will not be maintained self-cleansing at all times. The diurnal pattern of the dry weather
flow and the temporal distribution and nature of sediments found in sewer flows may result in the deposition of
some sediments at times of low flow and the subsequent erosion and transport of these sediments, either as
42
suspended load or bed-load, at times of higher flow. The deposited sediments will exhibit additional strength due to
cohesion and, provided that the peak dry weather flow velocity or bed shear stress is of sufficient magnitude to
erode these sediments, the sewer will maintain self cleansing operation at times of dry weather. May, et al. (1996)
presented a definition to describe a self cleansing sewer as “an efficient self-cleansing sewer is one having a
sediment-transporting capacity that is sufficient to maintain a balance between the amounts of deposition and
erosion, with a time-averaged depth of sediment deposit that minimizes the combined costs of construction,
operation and maintenance.” To achieve such self-cleansing performance, the following criteria apply:
1. Flows equaling or exceeding a limit appropriate to the sewer should have the capacity to transport a
minimum concentration of fine-grain particles in suspension (applicable for all types of sewerage systems).
2. The capacity of flows to transport coarser granular material as bed-load should be sufficient to limit the
depth of deposition to a specified proportion of the pipe diameter. This criteria generally relates to combined
and stormwater systems. Limit of deposition considerations, i.e., “no deposition” generally applies to sanitary
sewer designs. In this context, there must be sufficient shear in sanitary systems to avoid deposition of large
particles.
3. Flows with a specified frequency of occurrence should have the ability to erode bed particles from a
deposited granular bed that may have developed a certain degree of cohesive strength (applicable to all
systems).
To meet these criteria, new guidelines have been developed by CIRIA (Ackers, et al. 1996), and are being adopted
throughout Europe for the design of sewers to control sediment problems. Design criteria for the transport of fine
grained material in suspension, the transport of coarser sediments as bed load, and the erosion of cohesive sediment
deposits and guidelines on the minimum flow velocity and pipe gradient for different types and sizes of sewer are
outlined. To account for the effects of cohesion (Criterion 3), the design flow condition should produce a minimum
value of bed shear stress of 2.0 N/m2 (0.04 lb/ft2) on a flat bed for a rough concrete pipe (Ackers, et al. 1996).
The third criterion is of specific interest to the problem of cleansing accumulated mature sediment beds. Various
researchers have studied the flow conditions required to release particles from a deposited bed, which has developed
a degree of cohesion. Summaries of investigations forming much of the basis for Criterion 3 are as follows:
 Nalluri and Alvarez (1992), whose laboratory studies used synthetic cohesive sediments, concluded that there
were two ranges of bed shear stress at which erosion occurred: 2.5 N/m2 (0.05 lb/ft2) applying for the weakest
material, comprising a surface layer of fluid sediment, and 6 to 7 N/m2 (0.12 to 0.14 lb/ft2) for the more
granular and consolidated material below. It was found that, after erosion, the synthetic cohesive sediments
behaved very much like non-cohesive material.
 Ristenpart and Uhl (1993) found in field tests that during dry weather an average bed shear stress of 0.7 N/m 2
(0.015 lb/ft2) was required to initiate erosion, increasing to an average of about 2.3 N/m2 (0.05 lb/ft2) during
wet weather, or to 3.3 N/m2 (0.068 lb/ft2) after a prolonged period of dry weather and presumably,
consolidation of the deposited bed.
 Ashley and Crabtree (1992) suggested that the bonds between particles at the surface of a deposited bed are
weakened by the presence of the water, so that surface layers can be successively stripped away by the flow.
Measurements in the Dundee, Scotland, sewers indicated that it began to move at a fluid shear stress of about 1
N/m2 (0.02 lb/ft2), with significant erosion of a deposited bed occurring at bed shear of 2 to 3 N/m 2 (0.04 to
0.06 lb/ft2). Taking account of a review of work by other researchers, Ashley and Crabtree concluded that most
deposits whould be eroded at a shear stress exceeding 6 to 7 N/m2 (0.12 to 0.14 lb/ft2).
Pisano, et al. (1998) conducted a “desktop” analysis of the sedimentation and scour problems associated with the
3.05 m (10 ft) South Ottawa Tunnel, in Ottawa, Canada. This procedure analyses deposition on a daily basis,
determining accumulations of inorganic and organic deposits as well as scour and erosion of prior deposites over a
year’s period. The particles in this combined sewerage had settling velocities ranging from 0.1 to 30 cm/sec,
corresponding to 75 µm fine sand to 3/8 inch gravel. If the average shear stress is smaller than the critical shear
stress for that sieve size and density, deposition will occur. Erosion of the bed load is based on maximum hourly
43
shear stress. If the maximum shear stress is great than 2 N/m2, then erosion of the bed load occurs. Table 19 shows
the full range of depositon and scour that may occur for this large tunnel for a variety of discharge, velocity, and
shear stress conditons.
Table 19. Assessment of 3.05 Meter Tunnel (Pisano, et al. 1998)
Discharge
Velocity
(m3/sec)
(ft3/sec)
(m/sec)
(ft/sec)
2.6
93
0.36
1.2
3.5
130
0.47
1.5
4.4
160
0.60
2.0
5.3
190
0.72
2.4
6.1
220
0.84
2.8
7.0
250
1.06
3.5
8.8
320
1.20
4.0
10.5
380
1.44
4.7
13.2
470
1.80
5.9
15.8
570
2.17
7.2
17.5
630
2.41
7.9
Note: Kb = 1.2 millimeters = rough concrete
Fluid shear
(N/m2)
(lb/ft2)
0.27
0.0056
0.48
0.0099
0.74
0.015
1.04
0.021
1.43
0.029
1.86
0.038
2.85
0.059
4.15
0.085
6.47
0.13
9.3
0.19
11.5
0.24
Deposition or erosion potential
Severe deposition
Moderate to severe deposition
Mild to moderate deposition
Slight to mild deposition
Skims juvenile sediments
None to slight erosion top layer
Slight to mild erosion of consolidated beds (2-5%)
Mild erosion of consolidated beds (5-15%)
Moderate erosion of consolidated beds (15-25%)
Substantial erosion (25-35%)
Substantial erosion (35-50%)
Accumulation of Sediment in Bellevue Inlet Structures and Storm Drainage
An important part of the Bellevue NURP project was the measurement of the sediment accumulating in the inlet
structures (Pitt 1985). The storm drainage system inlets were cleaned and surveyed at the beginning of the project.
The 207 inlet structures were then surveyed nine times over two years to determine the depth of accumulating
material (from December 1979 through January 1981). The first year rate of accumulation was relatively steady
(based on 3 observation periods), while the sediment loading remained almost constant during the second year.
During the second year, there was about twice as much contaminated sediment in the separate storm drainage
system at any one time as there was on the streets. The scouring of the sewerage sediments out of the drainage
systems was not found to be significant during the project period. There was a period of heavy rains in October of
1981 (about 100 mm of rain during a week, very large for Bellevue) during the second year when the accumulated
material did not decrease, based on observations made before and after the rain (August 1981 and January 1982).
The lack of sediment movement from catchbasin sumps was also observed during earlier tests conducted in San Jose
by Pitt (1979). During that study, an idealized catchbasin and sump were constructed based on Lager, et al. (1974)
and was filled with clean material having the same particle sizes as typical sump material, along with fluorescent
tracer beads. During a year, freezing core samples were obtained and the sediment layers were studied to determine
any flushing and new accumulations of material. The sediment material was found to be very stable, except for a
very thin surface layer.
The first year accumulation rates (L/month per inlet) ranged from 1.4 in Lake Hills to 4.8 in Surrey Downs, as
shown on Table 20. The catchbasins and inlets had sumps (the catchbasin sumps were somewhat larger), while the
manholes were much larger, with more volume available for accumulation sediment. The stable volume that
occurred during the second year was about 60% of the total storage volumes of the catchbasins and inlets (sump
volume below the outlet pipe). If the sumps were very shallow, the maximum sediment depth was only about 12
mm, while the deeper sumps had about 150 mm of accumulated sediment. Individual inlet structures had widely
varying depths, but the depth below the outlet appeared to the most significant factor affecting the maximum sump
volume available. This “scour” depth generally was about 300 mm (1 ft.). If the sumps were deeper, they generally
were able to hold more sediment before their equilibrium depth was reached and would therefore require less
frequent maintenance. About 100 L/ha/yr accumulated in Surrey Downs, while only about 2/3 of this value
accumulated in Lake Hills. Nine of the most heavily loaded catchbasins in the first summer inventory in Surrey
Downs were located very near two streets that did not have curbs and had extensive nearby sediment sources
(eroding hillsides). These few catchbasins (about 10% of the total catchbasins) accounted for more than half of the
total Surrey Downs sediment observed during that survey. They also represented about 70% of the observed
increased loadings between the first winter and summer inventories.
44
Table 20. Accumulation Rate of Sediment in Inlet Structures in Bellevue, WA (Pitt 1985)
Number of structures
Surey Downs
(38.0 ha)
Catchbasins
Inlets
Manholes
Average
Lake Hills (40.7
ha)
Catchbasins
Inlets
Manholes
Average
total
per ha
Sediment accumulation
(L/month)
per ha
per unit
43
27
6
76 total
1.1
0.7
0.2
2.0 total
5.3
2.0
0.8
8.1
4.8
2.8
4.0
4.2
71
45
15
131 total
1.7
1.1
0.4
3.2 total
2.4
1.5
1.6
5.5
1.4
1.4
4.0
1.7
Approx. months
to stable volume
Stable volume (L)
per ha
per unit
13
20
19
15
68
40
15
123 total
62
57
76
62
18
14
23
18
43
22
36
101 total
25
20
90
31
In addition to the catchbasin surveys, pipe surveys were also conducted during the study. Very few storm drain pipes
in either test area had slopes less than one percent, the assumed critical slope for sediment accumulation. In Lake
Hills, the average slope of the 118 pipes surveyed was about 4 percent. Only 7 percent of the Lake Hills pipes had
slopes less than 1 percent. The 75 pipes surveyed in Surrey Downs had an average slope of 5 percent, and 12 percent
had slopes less than 1 percent. A pipe sediment survey was conducted in October of 1980. Very little sediment was
found in the storm drains in either study area. The pipes that had significant sediment were either sloped less than 11/2 percent, or located close to a source of sediment. The characteristics of the pipe sediments were similar to the
characteristics of the sediment from close inlets and catchbasins, indicating a common source, and the eventual
movement of the inlet sediments. The volume of sediment found in the Lake Hills pipes was about 1-1/2 m3, or
about 0.04 m3 per ha, or about 40% of the total sediment in the inlet structures (about 0.1 m 3 per ha stable volume).
This was equivalent to about 70 kg of sediment/ha. In Surrey Downs, much more sediment was found in the storm
drainage: more than 20 m3 of sediment was found in the pipes, or about 0.5 m3/ha or 1,000 kg/ha. Most of this
sediment was located in silted-up pipes along 108th St. and Westwood Homes Rd. which were not swept and were
close to major sediment sources.
The chemical quality of the captured sediment was also monitored. Tables 21 and 22 show the sediment quality for
Surrey Downs inlet structures sampled between January 13 and June 17, 1981. The sediment quality shown on this
table is very similar to the street dirt chemical quality that was simultaneously sampled and analyzed. It is interesting
to note that the COD values increase with increasing particle sizes, likely corresponding to increasing amounts of
organic material in the larger material. The nutrients are generally constant with size, while the metal concentrations
are much higher for the smaller particles, as expected for street dirt. As shown on this table, the lead values are
much higher when these samples were taken compared to current conditions. Current outfall lead concentrations are
now about 1/10 of the values they were in the early 1980s.
Table 21. Chemical Quality of Bellevue, WA, Inlet Structure Sediment (mg constituent/kg total solids) (Pitt
1985)
Particle Size (µm)
COD
TKN
TP
Pb*
Zn
<63
160,000
2,900
880
1,200
400
61-125
130,000
2,100
690
870
320
125-250
92,000
1,500
630
620
200
250-500
100,000
1,600
610
560
200
500-1,000
140,000
1,600
550
540
200
1,000-2,000
250,000
2,600
930
540
230
2,000-6,350
270,000
2,500
1,100
480
190
>6,350
240,000
2,100
760
290
150
* these lead values are much higher than would be found for current samples due to the decreased use of leaded gasoline since
1981.
45
Table 22. Annual Calculated Accumulation of Pollutants in Bellevue, WA, Inlet Structures (Pitt 1985)
L/ha/yr
Surrey
Downs
Lake Hills
Solids
kg/ha/yr
96
147
COD
kg/ha/yr
37
66
100
7.5
TKN
kg/ha/yr
0.17
TP
kg/ha/yr
0.25
Pb
kg/ha/yr
0.49
Zn
kg/ha/yr
0.10
0.07
0.07
0.07
0.02
These accumulations are substantial. As an example, the total suspended solids discharged at the outfalls was about
100 to 200 kg/ha/yr, with the amount accumulating in the catchbasin inlets having about the same quantity.
Summary of Sediment Transport in Sewers
To achieve self-cleaning in a drainage system, Pisano, et al. (1998) listed the following criteria:
1. Flows equaling or exceeding a limit appropriate to the sewer should have the capacity to transport a
minimum concentration of fine-grain particles in suspension (applicable for all types of sewerage systems).
2. The capacity of flows to transport coarser granular material as bed-load should be sufficient to limit the
depth of deposition to a specified proportion of the pipe diameter. This criteria generally relates to combined
and stormwater systems. Limit of deposition considerations, i.e., “no deposition” generally applies to sanitary
sewer designs. In this context, there must be sufficient shear in sanitary systems to avoid deposition of large
particles.
3. Flows with a specified frequency of occurrence should have the ability to erode bed particles from a
deposited granular bed that may have developed a certain degree of cohesive strength (applicable to all
systems).
Ashley and Crabtree (1992) made sediment movement measurements in the Dundee, Scotland, sewers and found
that sediment began to move at a fluid shear stress of about 1 N/m2 (0.02 lb/ft2), with significant erosion of a
deposited bed occurring at bed shear of 2 to 3 N/m2 (0.04 to 0.06 lb/ft2). Taking account of a review of work by
other researchers, they concluded that most deposits should be eroded at a shear stress exceeding 6 to 7 N/m2 (0.12
to 0.14 lb/ft2). This is likely to occur for most flow conditions in pipes at least 2 ft in diameter at a slope greater than
0.5%.
For the smallest slopes (<0.5% slopes), almost all settled particles would likely remain and not be scoured, while
almost all unconsolidated sediments would be scoured for pipe greater than 1% in slope. Pipe conditions resulting in
at least 3 ft/sec stormwater velocities would also have shear stresses of at least 4 N/m2 (0.08 lb/ft2). This shear stress
would likely cause scour of particles up to 2,000 µm in size for water having a low content of fine sediment. If the
water had a high content of fine sediment, the maximum particles likely to be scoured from the sediment may be
only about 100 µm in size.
Required Sample Line Velocities to Minimize Particle Sampling Errors
The collection of stormwater samples representing the complete range of likely particle sizes can be challenging.
Manual samples can be obtained as the flowing water drops into an inlet, or as the water cascades out of an outfall.
These samples, if frequent during an event, should well represent the complete particle size distribution. However, if
an automatic sampler using a pump system is used, there may be some losses of particles during the sampling
operation. Typical sample lines are Teflon lined polyethylene and are 10 mm in diameter. The water velocity in
the sample line is about 100 cm per second, enabling almost complete sampling of a wide range of particle sizes.
However, bed load sampling equipment is usually needed to adequately collect samples of bed load at outfalls.
Table 23 shows the particle sizes that would be lost in vertical sampling lines at a pumping rate of 30 and 100
cm/second (Burton and Pitt 2001).
46
Table 23. Losses of Particles in Sampling Lines (Burton and Pitt 2001)
30 cm/sec flow rate
Critical settling rate
Size range (m, for
(cm/sec)
 = 1.5 to 2.65
g/cm3)
100% loss
50% loss
25% loss
10% loss
1% loss
30
15
7.5
3.7
0.37
2,000 - 5,000
800 - 1,500
300 - 800
200 - 300
50 - 150
100 cm/sec flow rate
Critical settling rate
Size range (m, for
(cm/sec)
 = 1.5 to 2.65
g/cm3)
100
50
25
10
1
8,000 - 25,000
3,000 - 10,000
1,500 - 3,000
350 - 900
100 - 200
A water velocity of 100 cm/sec (about 3 ft/sec) would result in very little loss of stormwater particles. Particles of 8
to 25 mm would not be lifted in the sample line at all at this velocity, but these sized particles would not fit through
the openings of the intake or even fit in most sample lines. They are also not likely to be present in stormwater, but
may be a component of bedload in a stream, or gravel in the bottom of a stormdrain pipe, requiring special sampling.
Very few particles larger than several hundred micrometers occur in stormwater and these should only have a loss
rate of 10% at the most. Most particles in stormwater are between 1 and 100 m in diameter and have a density of
between 1.5 and 2.65 g/cm3. Even at 30 cm/sec, these particles should experience insignificant losses. A pumping
rate of about 100 cm/sec would add extra confidence in minimizing particle losses. ASTM (1995) in method D 4411
recommends that the sample velocity in the sampler line be at least 17 times the fall rate of the largest particle of
interest. As an example, for the 100 cm/sec example above, the ASTM recommended critical fall rate would be
about 6 cm/sec, enabling a particle of several hundred micrometers in diameter to be sampled with a loss rate of less
than 10%. This is certainly adequate for most stormwater sampling needs. Bed loads comprised of particles of up to
several mm can account for about 5 to 10% of the sediment load under some conditions, and special sampling would
be needed to collect these larger particulates. In many cases, this larger material would not move through the pipe,
but accumulate as a semi-permanent sediment deposit.
Sediment Transport in Grass Swales
Grass swales are vegetated open channels that collect and transport stormwater runoff. They are often used as
an alternative to concrete gutters to transport runoff along streets due to their low cost. However, they also
offer several advantages in stormwater quality management, especially in their ability to infiltrate runoff. This
section describes another benefit of grass swales: their ability to trap particulates during low flows. A series of
detailed laboratory tests were conducted to describe sediment transport processes for stormwater grass swales.
Field verifications of these processes are also described in this paper. As expected, runoff hydraulics,
especially depth of flow, along with swale length, affect the transport of particulates of different sizes. Shallow
flows (less than the grass height) provided consistently high removal rates, while deeper flows (and especially
along with relatively low sediment concentrations) had poorer sediment trapping abilities. Obviously, long
swales and large particle sizes are an effective combination, but the smallest particles are likely to be
effectively transported along most swales. There appeared to be equilibrium concentrations for different
particle sizes that were not further reduced, irrespective of swale length, likely associated with combinations of
scour of material from the underlying soil or of previously trapped sediment, and the carrying capacity of the
water.
Methodology
The Department of Civil and Environmental Engineering at the University of Alabama has been conducting research
investigating the effectiveness of grass swales for stormwater sediment transport for several years. This research was
initially supported by the Water Environment Research Foundation (WERF) (Johnson, et al. 2003) and more
recently by the University Transportation Center of Alabama (UTCA) (Nara and Pitt 2005). The aims of this
research are to understand the effects of different variables affecting sediment transport in grass swales, especially
considering different particle sizes, swale features, and flow conditions.
Controlled tests were conducted using specially constructed indoor grass swales and test solutions having known
concentrations of sediment with different particle sizes. The test solutions were made using particles of sieved sands
(locally acquired) and commercially-sized silica (from U.S. Silica Co.). The indoor swales were adjusted for
47
different slopes, and had three different grasses. Two series of indoor experiments were conducted. The first series
were exploratory in nature to identify the most important variables affecting sediment transport, while the second
series examined these variables in more detail and were used to develop a sediment transport model.
The first series of tests examined grass type, slope, swale length, time since the beginning of the flow, flow rate,
particle size, and sediment concentration. The second series of tests focused on fewer grass types, had less
variability in sediment concentrations, and composited samples over the complete test period. The sediment
transport processes described in this paper were derived during the second series of experiments.
The second series of indoor swale experiments included analyzing 108 samples for turbidity, total solids, total
suspended solids, total dissolved solids, and particle size distribution. The results of the indoor swale experiments
were verified during monitoring at an outdoor grass swale located adjacent to the Tuscaloosa (Alabama, U.S.A.)
City Hall, during actual storm events. The outdoor swale tests included collecting and analyzing 69 samples during
13 storm events from August to December 2004. These samples were analyzed for turbidity, total solids, total
suspended solids, total dissolved solids, and particle size distribution.
Indoor Experiments
Descriptions of the Experimental Set-up. Experimental swales were constructed in an indoor greenhouse facility
located in the Bevil building on the campus of the University of Alabama, as part of a prior research project
(Johnson, et al. 2003). Artificial sunlight (ambient UV and visible) is provided, and room temperature is maintained
at approximately 78oF (25oC) at this facility. The experimental set up consists of three identical rectangular channels
on a base which is adjustable for varying channel slopes. A mixture of 70% top soil and 30% sand (by weight) was
placed in the channel sections which are completely sealed by non-reactive marine-epoxy paint to prevent leakage.
Each channel is 2 ft wide (0.6 m), 6 ft long (1.82 m), and 6 in (15 cm) deep and has a specific type of lawn grass.
Jason Kirby constructed these swales and tested the grasses for hydraulic resistance during his MSCE thesis (2003).
The swale system hydraulic components produced a well-mixed sheetflow two feet wide. The flow of the test
mixture was controlled using a sediment slurry tank, a pump in a 150-gallon (0.57m3) tank, a mixing chamber, and a
preliminary wooden chute, as shown in Figure 14. The sediment slurry tank constantly fed a mixture of test
sediment into the mixing chamber where it was mixed with water pumped from the 150-gallon (0.57 m3) tank before
it flowed into the swale segment.
Sediment Characteristics of the Sediment-water Mixture. Fine-ground silica (SIL-CO-SIL® from US Silica Co.),
along with sieved sands, were used in the test mixture. Fine-ground silicas obtained from U.S. Silica are naturally
bright white, low in moisture, and chemically inert. Two different types of silca, SIL-CO-SIL®106 and SIL-COSIL®250 were used. Sieved sands were used to provide larger particles, ranging from 90 to 250 µm and 300 to 425
µm. The sediment concentration of the test flow was set at 500 mg/L, a concentration higher than for most
stormwaters in order to obtain sufficient particles for accurate sizing during the tests. Table 24 shows the percentage
contribution of the test sediments in different particle ranges, while Figure 15 shows the particle size distribution of
the test sediment mixture. This mixture therefore had a wide range of particle sizes represented, enabling sediment
transport processes to be described for several different size ranges that are represented in typical stormwaters.
48
Figure 14. Overview of indoor experimental set-up
Table 24. Percentage Contribution and Specific Gravity of the Test Sediments
Sediment
Ground silica “flour” (SIL-COSIL®106)
Ground silica “flour” (SIL-COSIL®250)
Fine silica sand (90 to 250 µm)
Coarse silica sand (300 to 425 µm)
Total
Percentage
contribution
Specific
gravity
15 %
2.65
50 %
2.65
25 %
10 %
2.65
2.65
100 %
49
100
Cumulative mass (%)
90
80
70
60
50
40
30
20
10
0
1
10
100
1000
Particle diameter (µm)
Figure 15. Particle size distribution of the composite test sediments mixture
Experimental Factors Tested During the Indoor Experiments. Four main factors were tested during the second
series of indoor experiments in order to quantify the effects of these variables on the transport of specifically-sized
sediment in grass swales. These factors were grass type, slope, flow rate, and swale length:




Grass types: The three different types of grass used were synthetic turf, Zoysia, and Kentucky bluegrass.
Zoysia has a dense root structure that has high wear-tolerance, has a smooth and tough blade, and is
commonly found in the Southern part of the United States. Kentucky bluegrass has smooth, upright stems
and very fine blades. It is commonly found in the North and Midwest parts of the United States. The grass
heights of both Zoysia and Kentucky bluegrass were maintained between 1.5 inches (3.8 cm) to 2 inches (5
cm) during the tests. The synthetic turf was obtained from a local warehouse store. The height of the blades
of the synthetic turf was approximately 0.25 inches (0.63 cm) which was much shorter than the other grass
types. The blades are made of thin plastic and are uniformly dense.
Slopes: 1 %, 3 %, and 5 % channel slopes were tested.
Flow rates: Flow rates tested were 10 gallons per minute (GPM) (0.038 m3/min), 15 GPM (0.064 m3/min),
and 20 GPM (0.076 m3/min), in each of the 2 ft (0.6 m) wide swale sections.
Swale length: Samples were collected at the head works (0 ft), and 2 ft (0.6 m), 3 ft (0.9 m), and 6 ft (1.8
m) from the head works.
Results and Discussions
During the second test series, 108 samples were collected and analyzed for the following analytical parameters:
1.
2.
3.
4.
5.
6.
Total solids (Standard Methods 2540B)
Total solids after screening with a 106 µm sieve
Total suspended solids (solids retained on a 0.45 µm filter) (Standard Methods 2540D)
Total Dissolved Solids (solids passing through a 0.45 µm filter) (Standard Methods 2540C)
Turbidity using a HACH 2100N Turbidimeter
Particle Size Distribution by Coulter Counter (Beckman® Multi-Sizer IIITM), composite of several different
aperture tube measurements
50
It was difficult to introduce consistent inlet concentrations during all of the experiments. Large particles in the 106
to 425 µm size range had the largest overall variability due to the difficulty of consistently keeping the large
particles suspended in the inlet water. Because of this variability, all concentration data were normalized against the
initial sediment concentrations. Analysis of Variance (ANOVA) tests were performed to determine the effects of the
experimental variables on these normalized concentration changes. However, residuals of the ANOVA tests were
not normally distributed as required (normality confirmed using the Anderson-Darling test). Therefore, for each
swale length, the normalized data were ranked. The ANOVA test was then used on these ranked normalized data to
determine the significance of the variables.
Significant sediment reductions were observed at 2 ft (0.6 m), 3 ft (0.9 m), and 6 ft (1.8 m) for total solids, total
solids after screening with a 106 µm sieve, total suspended solids, and turbidity. There were no significant changes
in total dissolved solids as a function of swale length. Sediments were rapidly reduced between the head works (0 ft)
and 2ft (0.6 m) due to large particle settlement at the beginning of the swale. After each experiment, sands were
visually observed up to 1 ft (0.3 m) from the head works. Unlike solids, turbidity reductions were relatively constant
with swale length since turbidity was not as affected by large particles.
Total suspended solids (mg/L)
1000
p < 0.001 (0 vs 6 ft)
800
600
400
p < 0.001
200
p = 0.002
p < 0.001
0
0 ft
2 ft
3 ft
6 ft
Swale length
Figure 16. Box-and-whisker plots of total suspended solids concentrations vs. swale length (Note
1 ft = 0.30 m)
All variables (swale length, grass type, slope, and flow rate) were found to be significant factors for most of the
particulate constituents. In contrast, all variables were found to be insignificant for total dissolved solids. Among the
three grass types, synthetic turf was found to be the least effective in trapping the sediment, and Zoysia and
Kentucky blue grass had similar sediment reduction rates. The effects of channel slope and flow rate were marginal
for total solids and total suspended solids. However, these effects were clearly significant for turbidity. 1 % slope
was found to be much more efficient in trapping the particulates than the 3 % and 5 % slopes, and the low flow rate
of 10 GPM (0.038 m3/min) (was more effective in reducing turbidity than the higher flow rates of 15 GPM (0.064
m3/min) and 20 GPM (0.076 m3/min). Some of the interactions between the factors were also important and need to
be considered when explaining sediment transport in grass swales. Table 25 shows the significant interaction terms
and associated probabilities (p-values) for each constituent.
51
Table 25. Significant Interaction Factors and Associated P-values
Constituent
Interaction factor
P-value
Total solids
Total solids (< 106 µm)
Grass type * Slope
Grass type * Flow rate
Flow rate * Slope
Grass type * Flow rate
Flow rate * Slope
Grass type * Flow rate
Grass type * Slope
Grass type * Flow rate
0.023
< 0.001
< 0.001
0.005
0.013
0.044
0.001
< 0.001
Total suspended solids
Total dissolved solids
Turbidity
The swale length was found to be a significant factor affecting particle size distributions, while the three other
factors (flow rate, slope, and grass type) were found to be insignificant. Figure 17 shows the median particle sizes of
runoff particulates for each swale length. The median particle sizes consistently decreased by swale length, as
expected, as the large particles were preferentially removed in the swale.
22.5
Median particle size (µm)
20.0
17.5
15.0
p = 0.002
12.5
p = 0.257
p = 0.001
10.0
7.5
p < 0.001 (0 vs 6 ft)
5.0
0ft
2ft
3ft
6ft
Swale length
Figure 17. Box-and-whisker plots of median particle sizes vs. swale length
(Note 1 ft = 0.30 m)
Outdoor Swale Observations
The indoor experiments were conducted to identify the significant factors affecting the transport of sediment in
the grass swales, and to develop an associated model. The confirmation of the model obtained from the indoor
experiments used sampling of stormwater at a full-size outdoor grass swale located adjacent to the Tuscaloosa,
Alabama, City Hall during actual storm events. Sixty-seven samples were collected at various locations along
the swale during 13 storm events from August to December 2004. These samples were analyzed for the same
constituents as analyzed during the second indoor tests.
Descriptions of the Site
This full-sized swale has a length of 116 ft (35.3 m) and is covered with Zoysia grass. Although this is a full-sized
swale, its drainage area is very small, only comprising 0.1 acres (4,200 ft 2 or 390 m2) of paved roads. Table 26
shows the channel slopes at various swale lengths. The slopes are steeper at the beginning of the swale and are
flatter at the end.
52
Table 26. Channel Slopes over Various Swale Regions
Swale region
0 to 5 ft (0 to 1.5 m)
5 to 10 ft (1.5 to 3.0 m)
10 to 70 ft (3 to 21.3 m)
70 to 116 ft (21.3 to 35.3 m)
Mean channel slope
5.2 %
4.8 %
3.0 %
1.4 %
A soil survey conducted at the outdoor swale found the soil to be compacted loamy sand in accordance with U.S.
Department of Agriculture (USDA) Classification methods. In addition to surveying the slope and topsoil of the
grass swale, infiltration rates of the swale were also measured using small double-ring infiltrometers (Turf-Tec, Inc).
The infiltration tests were conducted during both dry and wet conditions. Most of the infiltration rates were much
less than 1 inch/hour (2.54 cm/hour), resulting in little runoff infiltration during the storm events. The detailed soil
survey also found sediment accumulation at the head of the swale, with grass growing through the top of the
accumulated sediments. During storm events, the accumulated sediment created a small puddle at the head of the
swale preventing large particles from entering the swale due to initial sedimentation.
Sample Collection and Analytical Methods
A total of 67 samples were collected at 0 ft, 2 ft (0.6 m), 3 ft (0.9 m), 6 ft (1.8 m), 25 ft (7.6 m), 75 ft (22.8 m), and
116 ft (35.3 m) during 13 storm events from August 22, 2004 to December 8, 2004. However, not all events have a
complete set of samples. During some events, runoff was insufficient at the time of sampling at specific locations for
collecting a runoff sample.
The samples were analyzed for the following parameters:
1.
2.
3.
4.
5.
6.
Total solids (Standard Methods 2540B)
Total solids after sieving with a 106 µm sieve
Total suspended solids (solids retained on a 0.45 µm filter) (Standard Methods 2540C)
Total dissolved solids (solids passing through a 0.45 µm filter) (Standard Methods 2540D)
Turbidity using a HACH 2100N Turbidimeter
Particle Size Distribution by Coulter Counter (Beckman® Multi-Sizer IIITM), composite of several different
aperture tube measurements
All samples were split using a USGS/Dekaport cone splitter to ensure accurate divisions of the particulate-laden
water before analyses.
Results and Discussions
While collecting samples, sediment reduction with increasing swale length was visually observed during most of the
events. Figure 18 shows runoff samples and sediments captured on glass fiber filters at various swale lengths,
collected on October 11, 2004. The sediments captured on filters, for samples obtained near the swale entrance, were
solid black suggesting dusts from paved roads, while the color of sediment captured at 116 ft (35.3 m) was light
brown, primary caused by top soil erosion from the swale. It was clear that runoff sediments were captured as the
stormwater passed through the grass swale, but soil erosion also occurred at longer swale distances.
Total Suspended Solids. Initial total suspended solids at the entrance of the swale varied during different rain
events, ranging from 4 mg/L to 157 mg/L. High sediment reductions were normally observed between 0 ft and 25 ft
(7.6 m). Beyond 25 ft (7.6 m), there was much less sediment reduction in the grass swale. During two events
(10/23/2004 and 11/11/2004), total suspended solids increased between 0 ft and 6 ft (1.8 m) instead of decreasing,
likely due to scouring of previously deposited sediments at the entrance of the swale. An unusual sediment increase
of 51 mg/L between 25 ft (7.6 m) and 75 ft (22.8 m) was observed on 12/08/2004. During this sampling period, the
rain intensity and runoff flow rate significantly increased after collecting the upgradient samples. The higher flow
rate likely scoured the top soil of the swale, resulting in much higher TSS concentrations at 75 ft (22.8 m) than at 25
ft (7.6 m) during that event.
53
Date: 10/11/2004
116 ft
TSS: 10 mg/L
75 ft
TSS: 20 mg/L
25 ft
TSS: 30 mg/L
6 ft
TSS: 35 mg/L
3 ft
2 ft
TSS: 63 mg/L
Head (0ft)
TSS: 84 mg/L
*TSS = Total Suspended Solids
TSS: 102 mg/L
Figure 18. Sampling locations at the outdoor swale monitoring site
(Example sediment samples from 10/11/2004)
180
Total Suspended Solids (mg/L)
160
Legend: Events
8/22/2004
10/10/2004
10/19/2004
11/1/2004
11/21/2004
12/6/2004
140
120
100
10/9/2004
10/11/2004
10/23/2004
11/11/2004
11/22/2004
12/8/2004
80
60
40
20
0
0
20
40
60
80
100
120
Swale length (ft)
Figure 19. Total suspended solids concentrations vs. swale length, observed at the outdoor grass
swale (Note 1 ft = 0.30 m)
54
160
Total suspended solids (mg/L)
140
120
100
80
60
40
20
0
0
2
3
6
25
Swale length (ft)
75
116
Figure 20. Box-and-whisker plots of total suspended solids concentrations vs. swale length
observed at the outdoor grass swale (Note 1 ft = 0.30 m)
Figure 20 indicates that the concentrations were highly variable during the first three feet of the swale, significantly
decreased between 3 ft (0.9 m) and 25 ft (7.6 m), and then decreased only slightly more to the end of the 116 ft (35.3
m) swale.
Turbidity. Significant reductions in turbidity were observed at the outdoor swale. Although initial turbidity at the
entrance ranged from 2 NTU to 137 NTU, all turbidity values (except on 12/08/2004) were reduced below 20 NTU
at 116 ft (35.3 m). Increased turbidity at 75 ft (22.8 m) on 12/08/2004 was due to scouring of the top soil during an
intermittent period of high flows, as mentioned previously.
Particle Size Distribution. Observations at the outdoor swale confirmed that grass swales preferentially remove
larger particles, as expected. Figure 22 shows that most of the median particle sizes are significantly reduced
between 3 ft (0.9 m) and 25 ft (7.6 m). However, there was no noticeable change in particle size between the swale
entrance and 6 ft (1.8 m) due to possible scouring, and beyond 25 ft (7.6 m) due to equilibrium concentrations being
reached.
180
160
140
Turbidity (NTU)
Legend: Events
8/22/2004
10/10/2004
10/19/2004
11/1/2004
11/21/2004
12/6/2004
120
100
80
10/9/2004
10/11/2004
10/23/2004
11/11/2004
11/22/2004
12/8/2004
60
40
20
0
0
20
40
60
80
100
120
Swale length (ft)
Figure 21. Turbidity vs. swale length, observed at the outdoor grass swale (Note 1 ft = 0.30 m)
55
Median Particle Size (micro meter)
50
45
40
35
30
25
20
15
10
5
0
0
20
40
60
80
100
120
Swale length (ft)
Figure 22. Median particle sizes vs. swale length observed at the outdoor grass swale (Note 1 ft =
0.30 m)
100
90
Cumulative vomule (%)
80
70
60
50
0 ft
2 ft
3 ft
6 ft
25 ft
116 ft
40
30
20
10
0
0.1
1
10
100
1000
Particle diameter (micro meter)
Figure 23. Example particle size distributions for different swale lengths observed on December
6, 2004
Sediment Trapping Model Development and Verification
The primary purpose of conducting the controlled indoor experiments was to develop a model that predicts sediment
reduction of stormwater in actual grass swales. A model was developed using the analytical results (total solids, total
solids less than 106 µm, total suspended solids, total dissolved solids, and particle size distribution analysis)
obtained during the indoor experiment.
Concepts
During both the indoor experiments and the outdoor measurements, it was observed that sediment reductions were
much higher at the beginning of the grass swales and then leveled off with flow distance. Thus, first order decay was
applied to predict the reduction of runoff sediment as a function of swale length in order to normalize the effects of
the different entrance sediment concentrations. The following is the first-order decay as applied to sediment
transport in grass swales:
56
C 
Ln out   kt
 Cin 
(1)
Where:
Cout = Sediment concentration at down-gradient sampling locations
Cin = Initial sediment concentration at the swale entrance
k = First order constant
t=
Swale length in feet from the swale entrance
First-order constants (and associated percent reductions) were computed for each experimental test and observation.
To understand sediment removal in grass swales, the constants and percent reductions were calculated for eight
different particle size ranges:
1.
2.
3.
4.
5.
6.
7.
8.
< 0.45 µm (total dissolved solids)
0.45 to 2 µm
2 to 5 µm
5 to 10 µm
10 to 30 µm
30 to 60 µm
60 to 106 µm
106 to 425 µm
Settling Frequency
To incorporate the effects of particle sizes, swale length, flow depth and flow velocity on sediment transport, the
number of chances particulates may theoretically settle in the grass swale was added to the sediment transport
model. The number of chances particulates settle in a certain swale length is called “settling frequency.” Settling
frequency is computed using Stokes’ law, considering length of swale and velocity, flow depth, and settling velocity
of particles, is shown below.
Stokes’ Law:
Vs 

2
2 R * g * ( Pp  Pf )
*
9
U

(2)
Where:
Vs
R
g
Pp
Pf
=
=
=
=
=
Settling velocity of a particle (cm/s)
Radius of a particle (cm)
Gravitational constant = 980 cm/s2
Density of a particle = 2.65 g/cm3 (assuming silica)
Density of fluid = 1.0 g/cm3 (assuming water at standard
temperature conditions)
U = Dynamic viscosity = 0.01 g/(cm*s) (assuming water at
standard temperature conditions)
The time a certain particle requires to settle in a grass swale is therefore:
Settling _ duration 
flow _ depth
settling _ velocity
(3)
And the traveling time a certain particle takes from the beginning to the end of a swale is:
swale _ length
(4)
Traveling _ time 
flow _ velocity
Therefore:
57
Settling _ frequency 
traveling _ time
settling _ duration
(5)
Predictive Model
The relationship between flow depth and grass height was found to be very important since it considers flow rate,
slope, and runoff contact area with the grass. To determine the effect of the flow depth/grass height ratio, averaged
flow depth/grass height ratios were computed for each experimental condition. Percent reduction-settling frequency
plots were created for three distinct flow depth/grass height ratio categories: 0 to 1.0, 1.0 to 1.5, and 1.5 to 4. Ratios
less than 1 means that the grass height is higher than the flow depth. Kirby (2003) determined Manning’s roughness
factors for these low flow conditions so the flow depth can be accurately calculated for these low flows. In addition
to flow rates and the flow depth/grass height ratio, shear stress and slope were also examined to determine their
significance on sediment trapping. It was found that shear stress and slope were insignificant factors. Therefore,
settling frequency and the flow depth/grass height ratio were utilized for developing equations for modeling
sediment reductions in grass swales. Initial sediment concentrations of the indoor swale experiments were much
higher than the outdoor swale. Thus, the following sediment reduction equations are for high initial sediment
concentrations (about 500 mg/L) for different flow depth/grass height ratio ranges:
Flow depth/grass height ratio: 0 to 1.0
Y  2.101 * log( X ) 2  6.498 * log( X )  76.82
(6)
Where:
Y = Percent reduction
X = Settling frequency
Analysis of Variance
Source
DF
SS
Regression
2 10765.9
Error
142 23177.0
Total
144 33942.8
MS
5382.93
163.22
F
32.98
P
0.000
Sequential Analysis of Variance
Source
DF
SS
F
P
Linear
1 7728.35 42.16 0.000
Quadratic
1 3037.51 18.61 0.000
Normal Probability Plot of Residuals [Test: Anderson-Darling]
[ ( Flow depth) /( Gr ass height) Ratio: 0 to 1 .0 ]
99.9
M ean -2 .6 5 5 9 6 E-1 4
StDev
1 2 .6 9
N
145
AD
1 .1 4 7
P -V alue
0 .0 0 5
P er cent
99
90
50
10
1
0.1
-40
-30
-20
-10
0
Residual
10
20
30
40
Residuals V er sus the Fitted V alues [ ( Flow depth) /( Gr ass height) Ratio: 0 to 1 .0 ]
( r esponse is P er cent r eduction ( % ) )
Residual
20
0
-20
-40
70
75
80
85
Fitted V alue
90
95
100
Figure 24. Normal probability plot and residual plot of the residuals vs. fitted values for the (flow
depth)/(grass height) ratio between 0 to 1.0
58
Flow depth/grass height ratio: 1.0 to 1.5
Y  8.692 * log( X )  80.94
(7)
Where:
Y = Percent reduction
X = Settling frequency
Analysis of
Source
Regression
Error
Total
Variance
DF
SS
1
9986.4
62 13957.0
63 23943.5
MS
9986.44
225.11
F
44.36
P
0.000
Normal Probability Plot of Residuals [Test:Anderson-Darling]
[( Flow depth)/(Gr ass height) Ratio: 1 .0 to 1 .5 ]
99.9
M ean -8 .7 0 4 1 5 E-1 4
StDev
1 4 .8 8
N
64
AD
0 .5 1 5
P -V alue
0 .1 8 4
P er cent
99
90
50
10
1
0.1
-50
-25
0
Residual
25
50
Residuals V er sus the Fitted V alues [(Flow depth)/(Gr ass height) Ratio: 1 .0 to 1 .5 ]
(r esponse is P er cent r eduction (% ))
40
Residual
20
0
-20
-40
50
60
70
80
90
100
Fitted V alue
Figure 25. Normal probability plot and residual plot of the residuals vs. fitted values for the (flow
depth)/(grass height) ratio between 1.0 to 1.5
Flow depth/grass height ratio: 1.5 to 4.0
Y  2.382 * log( X ) 2  15.47 * log( X )  67.46
(8)
Where:
Y = Percent reduction
X = Settling frequency
Analysis of Variance
Source
DF
SS
Regression
2 48358.8
Error
131 21893.5
Total
133 70252.3
MS
24179.4
167.1
Sequential Analysis of Variance
Source
DF
SS
F
F
144.68
P
0.000
P
59
Linear
Quadratic
1
1
45055.2
3303.5
236.03
19.77
0.000
0.000
Normal Probability Plot of Residuals[Test:Anderson-Darling]
[ (Flow depth)/( Gr ass height) Ratio: 1 .5 to 4 ]
99.9
M ean 3 .6 9 0 5 8 0 E-1 4
StDev
1 2 .8 3
N
134
AD
0 .5 5 4
P -V alue
0 .1 5 1
P er cent
99
90
50
10
1
0.1
-40
-30
-20
-10
0
Residual
10
20
30
40
Residuals V er sus the Fitted V alues [( Flow depth) /(Gr ass height) Ratio: 1 .5 to 4 ]
(r esponse is P er cent r eduction (% ))
40
Residual
20
0
-20
-40
40
50
60
70
80
Fitted V alue
90
100
110
Figure 26. Normal probability plot and residual plot of the residuals vs. fitted values for the (flow
depth)/(grass height) ratio between 1.5 to 4.0
Figures 24, 25, and 26 also show the residual plots. The residuals were found to be normally distributed, except for
the flow depth/grass height ratio 0 to 1.0 due to several larger residuals at the extreme condition. Logtransformations on the percentage trapping values on the y-axis was attempted, however, they did not improve the
distribution of residuals. Also, the variability of the residuals was smaller at higher percent reduction rates than for
smaller rates. This is because the percent reduction values can not be greater than 100 % although the settling
frequency increases. Figure 27 shows the regression lines and 95 % confidence intervals of the three different flow
depth/grass height ratios, and the percent reductions for total dissolved solids (particles < 0.45 µm). The percent
reduction increases as the flow depth/grass height ratio decreases. This is especially true for lower settling
frequencies, while the differences in the regression lines become negligible when the settling frequency is above 10.
60
Particulate Transport in Grass Swale
Comparison of regression lines with 95%Confidence Intervals
by different (Flow depth)/(Grass depth) Ratios
100
90
Percent reduction (%)
80
Ratio: 0 - 1.0
70
60
Ratio: 1.0 - 1.5
50
Ratio: 1.5 - 4
40
30
Total Dissolved Solids
(<0.45 um)
20
10
0
0.00001
0.0001
0.001
0.01
0.1
1
10
100
Settling frequency
Figure 27. Comparison of regression lines with 95 % confidence intervals for different (flow
depth)/(grass height) ratios
Comparisons of Indoor and Outdoor Swale Observations
Since the drainage area of the outdoor swale was very small, most of the events had shallow flows below the
average grass height of 1 inch, which falls into the 0 to 1.0 flow depth/grass height ratio. Thus, the regression line
and 95 % confidence intervals of the indoor swale for 0 to 1.0 flow depth/grass height ratio were compared with the
outdoor swale observations shown in Figure 28. There was a significant difference between the indoor and outdoor
swale performances. The percentage reductions of particulates at the outdoor swale were much lower than for the
indoor swale, especially for low settling frequencies. This was assumed to be primary due to the lower initial
sediment concentrations at the outdoor swales (ranging from 4 to 157 mg/L TSS), compared to the initial sediment
concentrations during the indoor swale experiments (ranging from 193 mg/L to 1,021 mg/L TSS).
Particulate Transport in Grass Swale
Comparison of regression lines with 95%Confidence Intervals
between high and low initial sediment concentrations
100
90
High concetration
200 mg/L to 1000 mg/L (TSS)
Ratio: 0 to 1.0
Percent reduction (%)
80
70
60
50
40
30
20
Low concetration
40 mg/L to 160 mg/ L (TSS)
Ratio: 0 to 1.0
10
0
0.00001
0.0001
0.001
0.01
0.1
1
10
100
1000
Settling frequency
Figure 28. Regression lines with 95 % confidence intervals for the low and high initial sediment
concentrations (high concentrations from the second indoor experiment, average of 500 mg/L,
range of 200 to 1,000 mg/L; low concentrations from the outdoor swale observations, average of
60 mg/L, range of 10 to 160 mg/L)
61
Summary of Sediment Transport in Swales
Compared with conventional concrete gutters, grass swales offer several advantages in stormwater quality
management as well as cost efficiency. Grass swales are excellent for reducing stormwater pollutants by both
infiltration and filtration, depending on the soil, grass, and flow conditions. In this chapter, sediment concentration
reductions during shallow flows in grass swales were examined.
The indoor swale experiments revealed the significant factors affecting sediment transport in grass swales (swale
length, grass type, channel slope, and runoff flow rate). Interactions between the factors were found to be important
as well. Large particles, such as sands, were easily trapped at the beginning of the swales under most flow
conditions, while smaller particles were more likely to be carried greater distances along the swales. Models for
predicting sediment reductions were developed based on the particle settling frequency, and the flow depth/grass
height ratio.
Significant sediment reductions were observed in the grass swales. However, the outdoor swale monitoring revealed
that sediment reductions also depend on initial sediment concentrations. Greater sediment reduction rates were
observed for higher initial sediment concentrations.
Sample Processing and Particulate Analyses in Water Samples
In order to evaluate the ability of a stormwater treatment device to remove pollutants, it is necessary to have a
repeatable and comprehensive protocol for sample preparation and analysis. This section focuses on the important,
and often overlooked, area of the protocols used to ensure representative sample splitting for quantification of
pollutant mass by size fraction in the samples. The large sample sets discussed in this paper were generated using
simulant solids mixtures used by two teams of researchers (at Penn State Harrisburg [PSH] and the University of
Alabama [UA]) as they verified the solids’ removal performance of two prototype and full-scale devices. Solids
were the focus of this evaluation because of their relative ease and low cost of measurement and, as described later,
the relationship between particle size and pollutant loading in stormwater runoff. The result of this paper is a
detailed protocol for stormwater runoff sample preparation, whether the analysis is for solids only or also includes
particulate-associated chemical parameters.
Sample Preparation – Churn and Cone Splitters
Regardless of whether the samples are collected using grab- or composite-sample techniques, preparation of the
final samples for analysis is crucial. For solids analysis, many device verification protocols require each collected
sample be analyzed for both total suspended solids (TSS) and suspended sediment concentration (SSC). This
requires that each sample be split into multiple fractions, where each fraction is representative of the original
sample.
Two devices are currently accepted as producing an adequate split of a sample into representative subsamples –
churn splitters and cone splitters. The cone splitter is an excellent device for splitting samples that have been
returned to the laboratory for analysis. However, it is not normally used in the field. Instead, sampling splitting in
the field is often performed using a churn splitter, which is also very useful when manually collecting samples that
need to be composited. The churn splitter also has been used in laboratories as a means of subdividing samples into
analysis aliquots.
Gray, et al. (2000) documented the USGS conclusions regarding churn splitters. Churn splitters were determined to
be reliable for SSC determinations for solids concentrations of up to 1,000 mg/L when the median particle diameter
was less than 250 m. However, above 10,000 mg/L, or when the sample had greater than 25% sand (>62 m)
particles, churn splitters are considered by the USGS to be unreliable. Cone splitters are considered acceptable up to
10,000 mg/L and adequate up to 100,000 mg/L. For stormwater runoff samples (non-construction erosion samples),
where TSS/SSC concentrations are typically less than 2,000 mg/L and seldom have large fractions of sand-sized
particles even on the dirtiest sites, churn splitters and cone splitters should produce similar and repeatable results.
These upper-limit concentrations are highly unusual for stormwater runoff and it is crucial to determine how well
62
these sample-splitting devices function under more typical stormwater-solids concentration, especially when
periodic large particles may be present.
A recent research project investigating the performance of a proprietary filtration device (Pitt and Khambhammettu
2006, Khambhammettu 2006) included replicate solids and particle size analyses of samples split using a
commercial churn splitter and the USGS/Dekaport cone splitter. The 14-L churn splitter was developed by the
USGS in the late 1970s to accommodate manual sample compositing and splitting into separate sample bottles in the
field. It is available from Bel-Art Products (http://service.belart.com/cat/378050014.html).
The USGS in Quality of Water Branch Technical Memorandum No. 78.03 (January 1978)
(http://water.usgs.gov/admin/memo/QW/qw78.03.html) states that the churn sample splitter was designed to
facilitate the withdrawal of a representative subsample from a large composite sample of a water-sediment mixture.
Portions of the larger sample then can be withdrawn from the spigot into appropriately-sized sample bottles. The
USGS also described limitations of the churn sample splitter. They found that when relatively high concentrations of
sand-size particles (>62 µm) are present, subsamples for suspended-sediment concentration or particle size
determinations should not be taken from the churn. Those samples should be collected in separate bottles directly
from the stream. These results were confirmed in Gray et al. (2000).
During the last few years, researchers at the University of Alabama have tested these different methods to split
samples containing stormwater particulates in preparation for laboratory analyses. During the field evaluation of a
proprietary filtration device, Pitt and Khambhammettu (2006) and Khambhammettu (2006) obtained many replicate
samples using both a churn splitter and a cone splitter. These data provided an opportunity to measure the
performance of these common sample splitters under actual field conditions, using a mixture containing a wide
range of particle sizes and concentrations. Previously, Nara and Pitt (2005), Nara (2005) and Morquecho (2005)
conducted many laboratory tests to determine the performance of the cone splitter under controlled conditions, for a
wide variety of source area and outfall stormwater samples. This section briefly summarizes these results.
Initial Laboratory Evaluations of the USGS/Dekaport Cone Splitter
Morquecho (2005) studied stormwater characteristics affecting treatment and fate of stormwater pollutants in the
natural environment. One of the major elements of her work was developing appropriate sample processing methods
that would precede the detailed laboratory analyses (particle size analyses, anodic stripping voltammetry, pollutant
associations for different particle sizes, etc.). Since each sampling event generated large runoff volumes, the runoff
had to be split into smaller fractions for the different tests. Morquecho conducted performance evaluations of the
USGS/Dekaport Teflon™ cone sample splitter (Rickly Hydrological Company). This splitter was developed by the
USGS to more accurately divide large samples into the smaller volumes needed for the individual analyses (10
outlet ports on the cone splitter). Standard USGS procedures recommend the cone splitter for suspended sediment
concentrations up to 10,000 mg/L. According to the Rickly Hydrological Company, the Dekaport Teflon™ Sample
Splitter is a pour-through device machined from solid Teflon™ used for splitting water samples for a wide range of
particle sizes and water volumes. The USGS analysis found that this method resulted in superior sample replication,
especially compared to prior methods (such as shaking the sample bottle and pouring sample portions into graduated
cylinders). A review of the USGS data on cone splitters is summarized in Gray, et al. (2000).
The cone splitter was tested for its ability to equally split samples by first using a water blank and measuring the
resultant volumes. Since Morquecho was interested in producing five homogeneous fractions, two outlets were
inserted into each 1L bottle. The total 5L water sample was poured through the top of the cone splitter and the
resultant volumes collected in each bottle were measured. This trial was performed three times. The results of these
tests are show in Table 27. The trials produced errors, measured as coefficients of variation (COV), of 3.4 to 6.2%.
63
Table 27. USGS/Dekaport® cone splitter trials
Trial
USGS/Dekaport® Outlet Number
1,2
3,4
5,6
7,8
9,10
Average
Standard Deviation
COV
1
930 mL
985
910
940
1060
965 mL
59.79
0.062
2
930 mL
950
1050
950
950
966 mL
47.75
0.049
3
940 mL
980
1030
995
970
983 mL
33.09
0.034
Each of the five subsamples was further split into smaller fractions with the cone splitter, depending on the
laboratory tests. As an example, this work examined the chemical characteristics of stormwater particulates in
narrow particle size ranges (0 to 0.45 µm, 0.45 to 2 µm, 2 to 5 µm, 5 to 10 µm, 10 to 45 µm, 45 to 106 µm, 106 to
250 µm, 250 to 1200 µm, and >1200 µm sizes), requiring accurate sample splits for each sample being analyzed.
Nara and Pitt (2005) and Nara (2005) also conducted cone splitter performance tests. Nara’s research investigated
the transport of stormwater particulates in grass-lined drainage swales, both in controlled laboratory tests and with
full-scale swales during actual rains. The USGS/Dekaport cone sample splitter was used during this research to
accurately prepare the samples prior to detailed particle size analyses. The cone splitter was used to separate each
sample into 6 portions for separate chemical and size analyses, including duplicates. During the experimental design
phase of this research, Nara prepared sets of challenge samples consisting of Sil-Co-Sil and sieved sand mixture as
follows: 15% Sil-Co-Sil® 106, 50% Sil-Co-Sil® 250, 25% sand (sieved between 90 and 250 μm), and 10% sand
(sieved between 300 to 425 μm). Two separate runs were conducted for each of two mixes of this sediment recipe.
In addition to the above mixture, Sil-Co-Sil 250 and 90-to-250-µm sieved sand were both tested individually.
Approximately 500 mg of the sediment mixtures was mixed with 1 L of tap water, resulting in particulate
concentrations of approximately 500 mg/L. The test solutions were poured into the top of the USGS/Dekaport cone
sample splitter to produce ten identical sub-samples. Total solids analyses were conducted on all of the sub-samples
for the three sediment mixtures. The average total solids concentration for each sediment mixture was approximately
560 mg/L, due to the presence of about 60 mg/L dissolved solids in the tap water used.
Tables 28 and 29 show the performance of the USGS/Dekaport cone sample splitter for the different sediment
mixtures (analyzed for total solids [TS]) and for volume. The cone splitter was very efficient in splitting a sample
into very similar sub-samples. Very little variability was found between the sub-samples for both sample volumes
and sediment concentrations. The coefficients of variation (COV) of all the sub-sample sets for the three different
sediment types were less than 5%, although the splits for the larger particles had slightly greater variabilities than the
splits for the smaller particles. The variabilities for the water volume splits were very similar to those found during
the earlier (Morquecho 2005) tests. Also, the repeatability from subsample port to port was very good, with similar
variabilities (less than 5%) between the same port during different tests.
Table 28. Total Solids (TS) Test Results for the Dekaport® Cone Splitter
Tube ID
1
2
3
4
5
6
7
8
9
10
Avg.
Std. Dev
COV
Mix Sediments
First run
Second run
TS (mg/L)
TS (mg/L)
547.4
561.9
549.5
572.6
560.6
556.0
550.0
561.5
565.0
552.0
576.2
563.4
573.8
572.9
556.8
587.5
560.0
561.0
563.3
572.4
560.26
566.12
9.83
10.33
0.018
0.018
SIL-CO-SIL® 250
First run
Second run
TS (mg/L)
TS (mg/L)
573.1
563.2
556.0
559.8
563.7
547.6
553.5
558.8
558.2
560.6
564.3
565.3
577.4
523.1
565.3
571.9
563.6
559.0
574.5
570.4
564.95
557.96
7.98
14.01
0.014
0.025
Sieved Sand (9
First run
TS (mg/L)
573.7
578.4
558.8
565.0
586.7
598.0
587.9
576.3
581.0
569.7
577.55
11.58
0.02
Second run
TS (mg/L)
554.7
536.9
575.7
565.0
576.5
627.6
602.8
592.7
563.0
537.4
573.24
28.57
0.05
64
Table 29. Volumetric Test for the Dekaport® Cone Splitter
Tube ID
1
2
3
4
5
6
7
8
9
10
Average
Std. Dev
COV
First run
(mL)
97
95
109
104
101
101
107
97
101
99
101.1
4.48
0.04
Second
run (mL)
97
95
109
103
99
99
107
95
100
98
100.2
4.76
0.05
Third run
(mL)
97
96
108
103
100
99
107
96
100
97
100.3
4.37
0.04
Fourth
run (mL)
97
96
108
103
99
100
108
94
101
98
100.4
4.74
0.05
Fifth run
(mL)
97
95
109
104
100
101
107
95
100
98
100.6
4.79
0.05
Sixth run
(mL)
97
95
109
104
100
101
107
96
100
98
100.7
4.67
0.05
Avg.
97.0
95.3
108.7
103.5
99.8
100.2
107.2
95.5
100.3
98.0
Std. Dev.
0
0.52
0.52
0.55
0.75
0.98
0.41
1.05
0.52
0.63
COV
0
0.005
0.005
0.005
0.008
0.010
0.004
0.011
0.005
0.006
Replicate Analyses of Solids Analyses using Churn and Cone Sample Splitters
During a portion of these full-scale device tests, several controlled tests were conducted using challenge solutions
that contained specific particle size distributions and concentrations. Figure 29 shows the particle size distributions
of the influent and effluent samples used during some of these tests. The distributions for the other controlled tests
were similar and are given in the final evaluation report by Pitt and Khambhammettu (2006). The influent
concentrations ranged from about 50 to 500 mg/L solids measured as TSS. For almost all tests, the effluent samples
contained less than 5% sand-sized particles (by mass), similar to what is seen in typical stormwater samples.
Figure 29. Performance plot of particle size distribution for mixed media.
65
The samples described in this section were manually collected using a 0.5-L dipper grab sampler from the effluent
flow exiting the device. The dipper sampler was placed in the well-mixed cascading flow so that the complete flow
stream entered the dipper sampler container. An effluent sample was manually collected every 1 minute for each 30minute test period. These samples were composited in the 14-L churn sample splitter. After each test was completed,
the churn splitter was mixed continuously and three samples of 1L each were collected for the laboratory analyses.
Once at the laboratory, each of the samples was split again using the USGS/Dekaport® sample splitter to obtain the
analytical subsamples (in triplicate), as described in the previous subsection.
A total of 21 separate controlled experiments were conducted and total solids, total suspended solids, total dissolved
solids (by difference), and particle size distribution (PSD) analyses were carried out for each sample and replicates.
The total number of samples analyzed during the controlled tests was 168, including 40 blanks. The following
discussion presents the results of the duplicate tests, showing how the churn sample splitter and the cone sample
splitter performed with these samples that were similar to typical stormwater.
As noted above, three samples were collected for each effluent sample during the controlled experiments. Each of
these three samples had duplicate analyses conducted. Therefore, 126 analyses were conducted during these tests.
The results of the replicate tests for suspended solids using the churn sample splitter are shown in Figure 30. As
noted above, the laboratory tests used the complete subsamples, after splitting, and did not pipette or withdraw a
sample portion from the container for each test. The error bar plots show that the standard deviations were typically
between 5 to 7 mg/L, with periodic values being much larger. The analytical variabilities for the test results are
shown to be greater for the lower concentrations (especially for samples having <100 mg/L TSS) compared to the
samples having higher concentrations, when expressed as the COV (standard deviation/average). The median COV
value was 27% for the total suspended solids test results, while the median COV value was 8% for the total solids
test results.
Each of the churn samples were further split in the laboratory using the USGS/Dekaport cone sample splitter. In
addition, the sample blanks were also split using the cone splitter. Therefore, a total of 84 paired sets of samples
were available for comparison, representing 168 analyses. Figure 31 shows plots for these replicate analyses. With
duplicate analyses, a simple scatterplot was prepared comparing the two results for each subsample. Statistically,
there were an insufficient number of samples to show a significant difference at the 0.05 level between the two
paired analyses for each sample (even with 84 sample pairs). The use of 84 sample pairs should have detected a
significant difference of 10%, or greater, at the 95% confidence level and 80% power (assuming a COV of 25%).
The 95% confidence interval for the slope coefficient was very narrow and contained 1.0, again indicating similar
responses. For the churn splitter, the variations (expressed as COV, or relative standard deviations) are much larger
for small concentrations than for large concentrations. However, the median COV values are much lower for the
cone splitter than for the churn splitter, being about 11% for total suspended solids, and 5% for total solids.
General Laboratory Processing for Particle Size Fractionation of Samples
The UA research team has developed a general laboratory method for processing stormwater samples for particulate
analyses and particle size determinations, which is briefly outlined below. This procedure has been adopted by both
the UA and PSH laboratories for their solids analysis in order to ensure the repeatability of the sample analysis. This
procedure is used for samples generated both during the laboratory-verification testing and during traditional field
sampling.

Nylon screening material with 1200 µm openings is used at the top of the cone splitter to capture large
particles such as leaves, twigs and insects. This screening material is washed/soaked in hot soapy water for
one hour, and then thoroughly rinsed in DI water before use. Pouring the complete sample through this
coarse screening as it enters the cone splitter prevents these large materials from entering and clogging the
splitter. Removal of these large particles ensures a more-accurate subsample division. In addition, these
larger materials are usually in very low abundance in flowing stormwaters and collecting them from the
complete several-liter sample usually provides the researchers with sufficient mass to conduct selected
chemical analyses, after their removal from the screen (by rinsing with DI water), drying and weighing.

The sample volumes needed for each analytical series are determined, and the cone splitter tubes are
adjusted to provide the needed sample volumes for each split subsample. As an example, 100 mL is usually
used for particulate analyses, using Total Suspended Solids (TSS) methods. If 1L of sample is being split
66
and 7 particulate analyses are to be conducted, one tube is placed in each of 7 subsample containers
(usually graduated cylinders so the subsample volume can be directly determined) below the cone splitter,
and the remaining three tubes are placed in a single larger subsample container to collect the excess sample
volume.

In this approach to solids and particle size analyses, samples are separated into the following subsamples
for individual analyses:
1) Sample mass >1200 µm (from the coarse nylon screening before the cone splitter). This step is
optional for laboratory testing where the test solids mixture has been pre-screened to remove all
particles greater than 1000 μm.
2) Unfiltered subsample (2 replicates);
3) Sieved by 250 µm stainless steel 3 inch sieve (Tyler #60) (3 replicates); and
4) Filtered by 0.45 µm membrane filter (2 replicates)
This split generates 7 subsamples that are used for the particulate analyses, including the duplicate
analyses. A description of each subsample is given below.

o
The material captured on the 1200 µm coarse screening material is carefully rinsed from the
screen using DI water into a pre-dried and weighed oven-proof evaporating dish. This then is
placed in the drying oven at 103o to 105oC to dry. The sample is then weighed to determine the dry
weight of the large debris. It is important to recall that this weight is for the complete sample
volume, not for the smaller subsample volume. Therefore, accounting for its mass as part of the
original sample must apply to the entire volume of sample, not to the volume of a single
subsample. The collected debris can be appropriately digested and analyzed for chemical
constituents, after drying and weighing.
o
The unfiltered water subsamples are used for total solids analyses, so the total sample particulate
content less than 1200 µm can be determined. Duplicate analyses are conducted using the two
subsamples.
o
The next three subsamples are separately poured through a 3-inch stainless steel Tyler #60 sieve,
leaving subsamples having particulates smaller than 250 µm in the subsample. Two of these are
used for duplicate total solids analyses, while one will be used for the Coulter Counter particlesize-distribution analyses.
o
The last two subsamples are separately filtered through 0.45-µm membrane filters. The filtrate is
used for standard dissolved solids analyses, while the filters are dried and weighed for suspended
solids analyses. Membrane filters with a nominal pore diameter of 0.45 μm are used here for two
reasons: (1) many researchers consider the division between colloidal and suspended particles to
be approximately 0.45 μm (and the TSS methodology does not specify the filter’s pore size,
although some TSS guidance lists a glass-sintered filter of 2 µm, or smaller, especially if volatile
TSS analyses are to be measured), and (2) many of these samples undergo metals analysis for both
total and dissolved metals. Dissolved metals samples where the metals’ concentration is at the
concentrations typically seen in stormwater must be filtered through a membrane, as opposed to a
glass-fiber, filter. Past research (Clark 1996) has shown that glass-fiber filters have the potential to
contaminate the filtrate, especially by zinc.
The remaining 250-µm-sieved subsample is used for the Coulter Counter particle size distribution (PSD)
analyses. The samples usually collected by these research teams have the vast majority of the particulate
mass between 1 and 250 µm in size, the optimal range for Coulter Counter analyses. With the solids
analyses described above, particulate masses in the following size ranges are directly measured: <0.45 µm,
0.45 to 250 µm, 250 to 1200 µm, and >1200 µm. Therefore, the overall particulate size distribution can be
created by integrating the high resolution Coulter Counter size distribution results (reported on a volume
basis which is correlated to mass through particle density – which typically varies little among the particle
size fractions analyzed by the Coulter Counter), with the associated mass values for the coarser ranges.
Three aperture tubes (30 μm, 140 μm, 400 μm) typically are used with the UA Coulter Counter (Beckman®
67
Multi-Sizer III™) to create a composite distribution. The PSH methodology differs only in that the 100-μm
aperture tube is used instead of the 140-μm aperture tube which creates a slightly different overlap in the
data reduction step.

This composite distribution is created using the software provided by Coulter that overlaps the different
tube results. Each aperture tube can quantify particles in the range of approximately 2% to 60% of the
aperture size (e.g., 30 μm tube – 0.6 μm to 18 μm; 140 μm tube – 2.8 μm to 84 μm; 400 μm tube – 8 μm to
240 μm). As can be seen, each of the tubes substantially overlaps the other tube, providing sufficient
duplication of particle diameters for the software overlap. Three-inch stainless steel Tyler sieves are used to
presieve the subsamples before analyses by each aperture to minimize clogging; the sieve size selected is
the smallest commercially-available sieve that exceeds the maximum analytical range of each tube, while
still being smaller than the tube aperture itself (e.g., 30-μm tube – 20-μm sieve [Tyler #625], 140-μm tube –
106-μm sieve [Tyler #150], 400-μm tube – 250-μm sieve [Tyler #60]). The sample is pipetted through the
sieve and directly into the Coulter Counter vessel. Each aperture tube analysis usually is repeated three
times and the results averaged. This range of aperture tubes allows for very high resolution results from
about 0.6 to 240 µm.
Summary of Particulate Sampling and Analyses Issues
As research continues on the performance (and improvements to performance) of stormwater treatment devices, the
ability to document performance by particle size range will become more important. In addition, this size
fractionation information is crucial in predicting the fate and transport of many pollutants in urban runoff. The
ability of the research laboratories to document a standardized protocol for sample preparation and testing is vital to
being able to adequately compare performance results between devices and between vendors for manufactured
treatment devices. Sample preparation for particle size fractionation and analysis is not adequately addressed in the
standard protocols commonly used in the stormwater profession. Therefore, the protocol presented here addresses
those gaps in guidance to produce a repeatable sample preparation technique. The sample preparation protocol is
summarized in the steps below.
1.
2.
3.
4.
5.
The cone splitter is set up for the desired number of subsamples, up to a maximum of 10, considering the
needed analytical volumes for each subsample.
A 1200-m cleaned mesh screen is placed on top of the cone splitter and the entire sample is poured
through the mesh. The material retained on the mesh is dried, weighed and saved for further analyses.
The remaining samples are split according to the required analyses with approximately half of the samples
screened through a 250-m sieve in order to characterize the fraction in the larger range (below the mesh
size) that cannot be quantified as easily by the Coulter Counter Multisizer 3.
For samples requiring both TSS and SSC, the appropriate subsamples are set aside. For verification
protocol testing, this requires four aliquots – TSS unsieved, TSS sieved to less than 250 m, SSC unsieved,
and SSC sieved to less than 250 m.
TSS/SSC samples are filtered through the appropriate filter – either glass-fiber (if metals analyses are not
required) or membrane (if metals analyses are required).
The above protocol can easily be modified to reflect specific project requirements. If a Coulter Counter is not
available, these methods still result in important particle size data. Additional mechanical sieves can also be added to
increase the number of discrete particle sizes. This is commonly used in our laboratories during tests of chemical
analyses of the different particle sizes. The most important features in the described method is the use of the cone
splitter and analyzing the complete subsample, and further separating the sample into a few broad categories
(especially >250 µm, for example).
Summary of the Significance of Stormwater Particulates
The significance of stormwater particulates, including bedload and floatables, is becoming better understood. The
use of TSS alone may not be adequate in investigating problems, and when conducting treatability tests. Particle size
distributions, covering the complete range of the particulates, are needed. The location of the sample collection is
68
also very important. Samples collected near the original source of the particulates likely have many large particles
that strongly influence the distribution of particles, by mass. However, most sheetflows, and even concentrated
flows, are not capable of transporting the larger particles great distances. This is especially true if grass channels are
used with shallow flows. While source area samples may have median particle sizes of several hundred
micrometers, outfall samples rarely have any particles greater than a few hundred micrograms. Typical outfall
sample median particle sizes are usually much less than 50 µm, for example. The larger particles are preferentially
removed and trapped in the drainage system. If an inlet is located directly adjacent to a receiving water, then large
particles may be discharged, but this is uncommon.
Sample collection must include bedload and floatable solids captures if a complete description of the sediment is
needed. Automatic samplers cannot collect these sample fractions. Manual sampling, if conducted in a cascading
water flow, should capture bedload, but the mass fraction of these components are relatively small and large sample
volumes must be collected to capture enough sample to conduct chemical tests on the larger material. When
evaluating control practices, it is necessary to capture samples to describe a complete mass balance (all influent and
all effluent locations, plus captured material). This will enable sample collection inaccuracies to be determined,
along with direct measurements of the device performance.
Sample handling and analysis techniques must also be carefully conducted. SSC analyses are preferred, where
sample cone/funnel splitters are used to separate the sample into subsample fractions.
References
Ackers, J.C., D. Butler, and R.W.P. May. Design of Sewers to Control Sediment Problems. Construction Industry
Industry Research and Information Association (CIRIA), Report 141, 1996.
Akeley, R.P. “Retention ponds for control of urban stormwater quality.” In Proceedings National Conference on
Urban Erosion and Sediment Control: Institutions and Technology, EPA-905/9-80-002, Chicago, January 1980.
American Society of Civil Engineers. Design and Construction of Sanitary and Storm Sewers, ASCE Manuals and
Reports on Engineering Practice, No. 37, 1970.
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 Environ. Res., 71, 4,
398.
Armitage, N., and Rooseboom, A. (2000a) The Removal of Urban Litter from Stormwater Conduits and Streams:
Paper 1 - The Quantities Involved and Catchment Litter Management Options. Water SA. 26, 181.
Armitage, N., and Rooseboom, A. (2000b) The Removal of Urban Litter from Stormwater Conduits and Streams:
Paper 2 - Model Studies of Potential Trapping Structures. Water SA. 26, 189.
Arthur, S., and Ashley, R.M. (1998). The influence of Near Bed Solids Transport on First Foul Flush in Combined
Sewers. Wat. Science Technol., 37, 1, 131.
Arthur, S.; Ashley, R.; Tait, S.; and Nalluri, C. (1999) Sediment Transport in Sewers - A Step Towards the Design
of Sewers to Control Sediment Problems. Proc. the Inst. Civ. Eng., Water Maritime and Energy 136, 1, 9.
Aryal, R.K.; Furumai, H.; Nakajima, F.; Boller, M. (2005). Dynamic behavior of fractional suspended solids and
particle-bound polycyclic aromatic hydrocarbons in highway runoff. Water Res. 39(20):5126-5134.
Aryal, R.K.; Jinadasa, H.K.P.K.; Furumai, H.; Nakajima, F. (2005). A long-term suspended solids runoff simulation
in a highway drainage system. Water Sci. Technol. 52(5):159-167.
Ashley, R.M. and R.W. Crabtree. “Sediment origins, deposition and build-up in combined sewer systems.” Water
Science Technology, Vol. 25, No. 8, 1992.
Ashley R.M.; Hvitved-Jacobsen, T.; Vollertsen, J.; McIlhatton, T.; and Arthur S. Sewer Solids Erosion, Washout,
and a New Paradigm to Control Solids Impacts on Receiving Waters. Proc. the Eighth International Conference
on Urban Storm Drainage. August 30 – September 3, 1999, Sydney, Australia. Edited by IB Joliffe and JE Ball.
The Institution of Engineers Australia, The International Association for Hydraulic Research, and The
International Association on Water Quality, 171. 1999.
Ashley, R.; Crabtree, B.; and Fraser, A. Recent European Research into the Behavior of Sewer Sediments and
Associated Pollutants and Processes. 2000 Joint Conference on Water Resources Engineering and Water
Resources Planning and Management, July 2000, Minneapolis, MN. American Society of Civil Engineers, CDROM. 2000.
Ashley, R.M.; Dudley, J.; Vollertsen, J.; Saul, A.J.; Blanksby, J.R.; Jack, A. The effect of extended in-sewer storage
on wastewater treatment plant performance. Water Science and Technology, 45(3), 239-246. 2002.
69
ASTM (American Society for Testing and Materials). (1997) Method D3977-97(B): Standard Methods for
Determining Sediment Concentration in Water. ASTM International, West Conshohocken, PA.
Babaeyan-Koopaei, K.; Butler, D.; and Davies, J. (1999) Modelling of Gross Solids Transport In Sewers. Proc. the
Eighth International Conference on Urban Storm Drainage. August 30 – September 3, 1999, Sydney, Australia.
Edited by IB Joliffe and JE Ball. The Institution of Engineers Australia, The International Association for
Hydraulic Research, and The International Association on Water Quality, 507.
Bannerman, R., K. Baun, M, Bohn, P.E. Hughes, and D.A. Graczyk. Evaluation of Urban Nonpoint Source
Pollution Management in Milwaukee County, Wisconsin. U.S. Environmental Protection Agency, PB 84-114164,
Chicago, Ill., 1983.
Barry, S.C.; Taylor, S.E.; and Birch, G.F. Heavy Metals in Urban Stormwater Canals Entering Port Jackson,
Australia and Their Impact on the Estuarine Environment. Proc. the Eighth International Conference on Urban
Storm Drainage. August 30 – September, 3 1999, Sydney, Australia. Edited by IB Joliffe and JE Ball. The
Institution of Engineers Australia, The International Association for Hydraulic Research, and The International
Association on Water Quality, 1825. 1999.
Becker, F.A., P.D. Hedges, and R.P.M. Smisson. “The relationship between pollutants and specific sewage settling
velocity fractions.” NOVATECH 95, 2nd International Conference on Innovative Technologies in Urban Storm
Drainage. May 30 - June 1, 1995. Lyon, France. pp. 607-610. Organized by Eurydice 92 and Graie. 1995.
Benoit, G.; Rozan, T.F.; Patton, P.C.; and Arnold, C.L. (1999) Trace Metals and Radionuclides Reveal Sediment
Sources and Accumulation Rates in Jordan Cove, Connecticut. Estuaries. 22, 1, 65.
Benoit, G.; Wang, EX.; Nieder, W.C.; Levandowsky, M.; and Breslin, V.T. (1999) Sources and History of Heavy
Metal Contamination and Sediment Deposition in Tivoli South Bay, Hudson River, New York. Estuaries. 22, 2A,
167.
Bertrand-Krajewski, J-L; Chebbo, G; and Saget, A. (1998) Distribution of Pollutant Mass vs. Volume in Stormwater
Discharges and the First Flush Phenomenon. Water Res. (G.B.), 32, 8, 2341.
Blazier, A.; Sansalone, J.J.; Kim, J.-Y. (2003). Aggregation of suspended rainfall-runoff particles: Evaluation of pH
and velocity gradients. WEFTEC 2003 Conf. Proc. Water Environment Federation. CD-ROM.
Bouteligier, R., Vaes, Guido, Berlamont, J. (2002). In Sewer Sediment and Pollutant Transport Models. Proc. 9th
Int. Conf. Urban Drainage – Global Solutions for Urban Drainage. CD-ROM.
British Standard Institute. Sewerage – Guide to New Sewerage Construction, BS 8005, Part 1. 1987
Buffleben, M.S.; Zayeed, K.; Kimbrough, D.; Stenstrom, M.K.; and Suffet, I.H. (2001). Evaluation of Urban NonPoint Source Runoff of Hazardous Metals that Enters Santa Monica Bay, California. 5th International Conf.:
Diffuse/Nonpoint Pollution and Watershed Management. CD-ROM.
Burgess, E.H.; Ray, C.; and Farnham, R. (1999) Operational Experience With CSO Control Facilities in
Indianapolis, Indiana. Proc. Water Environ. Fed. 72nd Annu. Conf. Exposition, [CD-ROM], New Orleans, LA.
Burton, G.A. and R. Pitt. Stormwater Effects Handbook. CRC Press. Boca Raton, FL. 2001.
Butcher, J.B.; Pollman, C.; and Thirolle, E.A. (2000) Mercury TMDLs for Arivaca and Penna Blanca Lakes, AZ.
Watershed 2000 Management Conference, July 2000, Vancouver, British Columbia. Water Environment
Federation, CD-ROM.
Butler, D. and S.H.P.G. Karunaratne. “The suspended solids trap efficiency of the roadside gully pot.” Wat. Res.
Vol. 29, No. 2. pp. 719-729. 1995.
Butler, D., Y. Xiao, S.H.P.G. Karunaratne, and S. Thedchanamoorthy. “The gully pot as a physical and biological
reactor.” Wat. Sci. Tech. Vol. 31, No. 7, pp. 219-228. 1995.
Buxton, A.P.; Tait, S.; Stovin, V.; Saul, A. (2002). Developments in a methodology for the design of engineered
invert traps in combined sewer systems. Water Science and Technology, 45(7), 133-142.
Byars, M.S., and Kelly, M. (2001). Sediment Transport in Urban Stream Restoration. Proc. ASCE EWRI Conf. Bridging the Gap: Meeting the World’s Water and Environmental Resources Challenges. CD-ROM.
Bzdusek, P.A.; Lu, J.; Christensen, E.R. (2005). Evidence of fine-grained sediment transport and deposition in
Sheboygan River, Wisconsin, based on sediment core chemical tracer profiles. Water Resour. Res. 41(4):W04016.
Campisano, A.; Creaco, E.; Modica, C. (2004). Experimental and numerical analysis of the scouring effects of
flushing waves on sediment deposits. J. Hydrol. (Amst.). 299(3-4):324-334.
Characklis, G.W., and Wiesner, M.R. (1997) Particles, Metals, and Water Quality in Runoff from Large Urban
Watershed. J. Environ. Eng., 123, 8, 753.
Chebbo, G.; Ashley, R.; Gromaire, M.-C. (2003). The nature and pollutant role of solids at the water-sediment
interface in combined sewer networks. Water Sci. Technol. 47(4):1-10.
Chen, G.H., and Leung, D.H.W. (2000) Utilization of Oxygen in a Sanitary Gravity Sewer. Water Res. 34, 3813.
70
Chen, S.-H.; Hsu, M.-H.; Chen, T.-S. (2003). Simulation for interactions between storm sewer and overland flows.
Proc. ASCE Internat. Conf. Pipeline Eng. Const.: New Pipeline Technologies, Security, and Safety. 1:437-446.
Chow, Ven Te. Open Channel Hydraulics. McGraw-Hill. 1959.
Cigana, J.; Couture, M.; Lefebvre, G.; and Marche, C. Evidence of a Critical Velocity in Underflow Baffle Design
for Floatables Control in Combined Sewer Overflows (CSOs). Proc. Adv. in Urban Wet Weather Pollut.
Reduction, Cleveland, Ohio, WEF (CP3805), 275. 1998a.
Cigana, J.; Lefebrve, G.; Marche, C.; and Couture, M. Design Criteria of Underflow Baffles for Control of
Floatables. IAWQ 19th Biennial Int. Conf., Vancouver, Can., 8, 58. 1998b.
Cigana, J.; Lefebrve, G.; Marche, C.; and Couture, M. Design Criteria of Underflow Baffles for Control of
Floatables. Water Sci. Technol. (GB), 38, 10, 57. 1998c.
Cigana, J.; Couture, M.; Meunier, C.; and Comeau, Y. Determination of the Vertical Velocity Distribution of
Floatables in CSOs. Water Sci. Technol. (G.B.). 39, 2, 69. 1999.
Cigana, J.F.; Lefebvre, G.; and Marche, C. Experimental Capture Efficiency of Floatables using Underflow Baffles.
Collection Systems Wet Weather Pollution Control: Looking into Public, Private, and Industrial Issues, May
2000, Rochester, NY. Water Environment Federation, CD-ROM. 2000.
Cigana, J.; Lefebvre, G.; and Marche, C. Critical Velocity of Floatables in Combined Sewer Overflow (CSO)
Chambers. Water Sci. Tech. 44:287. 2001.
Clark, S. Evaluation of Filtration Media for Stormwater Runoff Treatment. MSCE thesis prepared for the
Department of Civil and Environmental Engineering, the University of Alabama at Birmingham, Birmingham,
AL. 442 pgs. 1996.
COE. U.S. Army Corps of Engineers. Engineering and Design: Channel Stability Assessment for Flood Control
Projects. EM 1110-2-1418. U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS. 1994.
Corsi, R.; Greb, S.R.; Bannerman, R.T.; and Pitt, R.E. (1999) Evaluation of the Multi-Chambered Treatment Train,
a Retrofit Water-Quality Management Device. U.S. Geological Survey Open-File Report 99-270. Middleton, WI.
Cristina, C.M.; Sansalone, J.J. (2003). "First flush," power law and particle separation diagrams for urban stormwater suspended particulates. J. Environ. Eng. 129(4):298-307.
Dally, L.K., D.P. Lettenmaier, S.J. Burges, and M.M. Benjamin. Operation of Facilities for Urban Stream Quality
Enhancement. University of Washington, Department of Civil Engineering, Water Resources Series Technical
Report No. 79, Seattle, Washington, July 1983.
De Martino, F.; Gisonni, C.; Hager, W.H. (2002). Drop in combined sewer manhole for supercritical flow. Journal
of Irrigation and Drainage Engineering, 128(6), 397-400.
Del Giudice, G.; Gisonni, C.; and Hager, W.H. (2000) Supercritical Flow in Bend Manhole. J. Irrigation Drainage
Eng. 126, 48.
Deletic, A. The First Flush Load of Urban Surface Runoff. (1998) Water Res. (U.K.), 32, 8, 2462.
Dettmar, J.; Rietsch, B.; Lorenz, U. (2002). Performance and operation of flushing devices: Results of a field and
laboratory study. Global Solutions for Urban Drainage, Proc. of the Ninth Int. Conf. on Urban Drainage, Sept 813 2002, Portland, OR, CD-ROM.
Digman, C.J., Littlewood, K., Butler, D., Spence, K., Balmforth, D.J., Davies, J., and Schütze, M. (2002). A Model
to Predict the Temporal Distribution of Gross Solids Loading in Combined Sewerage Systems. Proc. 9th Int. Conf.
Urban Drainage – Global Solutions for Urban Drainage. CD-ROM.
Doll, J.C.; Bryson, H.; Pizzi, D.V. (2004). Modeling indicators of channel condition for local watershed planning to
target stream restoration and management efforts. Watershed 2004 Conf. Proc. Water Environment Federation.
CD-ROM.
Driscoll, E. D. “Detention and retention controls for urban stormwater.” Engineering Foundation Conference:
Urban Runoff Quality - Impact and Quality Enhancement Technology, Henniker, New Hampshire, edited by B.
Urbonas and L. A. Roesner, published by the American Society of Civil Engineers, New York, June 1986.
EPA (U.S. Environmental Protection Agency). Results of the Nationwide Urban Runoff Program. Water Planning
Division, PB 84-185552, Washington, D.C., December 1983.
Erbe, V., Frehmann, T., Geiger, W.F., Krebs, P., Londong, J., Rosenwinkel, K.-H., and Seggelke, K. (2002).
Integrated Modelling as an Analytical and Optimisation Tool for Urban Watershed Management. Water Science
and Technology 46:141.
Estèbe, A.; Mouchel, J-M; and Thévenot, D.R. (1998) Urban Runoff Impacts on Particulate Metal Concentrations in
River Seine. Water, Air, Soil Pollut., 108, 1-2, 83.
Eyles, N.; Doughty, M.; Boyce, J.I.; Meriano, M.; Chow-Fraser, P. (2003). Geophysical and sedimentological
assessment of urban impacts in a Lake Ontario watershed and Lagoon: Frenchman's Bay, Pickering, Ontario.
Geoscience Canada. 30(3):115-128.
71
Fan, C.Y.; Field, R.; Pisano, W.C.; Barsanti, J.; Joyce, J.J.; and Sorenson, H. (2001a). Sewer and Tank Flushing for
Sediment, Corrosion, and Pollution Control. J. Water Resour. Plan. Man. 127:194.
Fan, C.-Y.; Field, R.; Sullivan, D.; and Lai, F.-h. (2001b). Toxic Pollutants in Urban Wet-Weather Flows: An
Overview of the Multi-Media Transport, Impacts, and Control Measures. Proc. ASCE EWRI Conf. - Bridging the
Gap: Meeting the World’s Water and Environmental Resources Challenges. CD-ROM.
Fan, C.-Y.; Field, R. (2002). Flushing for sewer corrosion and pollution control. Linking Stormwater BMP Designs
and Performance to Receiving Water Impact Mitigation, Proc. of an Engineering Foundation Conf., August 19 –
24, 2001, Snowmass, CO. 509-513.
Fan, C.-Y.; Field, R.; Lai, F.-h. (2003). Sewer-sediment control: Overview of an environmental protection. J.
Hydraul. Eng. 129(4):253-259.
Fischer, C.F. (2002). GASB-34: An Engineer’s perspective. WEFTEC 2002 Conf. Proc., September 2002. Water
Environment Federation. CD-ROM.
Fortier, S. and F.C. Scobey. “Permissible canal velocities.” Trans. ASCE, Vol. 89, paper No. 1588, pp. 940-984.
1926.
Frederick, R, and Corrigan, M. B. (1998) Urbanization and Streams: Studies of Hydrologic Impacts. Proc.
Watershed Manage. - Moving from Theory to Implementation, Denver, Colo., WEF (CP3804), 671.
Furumai, H.; Balmer, H.; Boller, M. (2002b). Dynamic behavior of suspended pollutants and particle size
distribution in highway runoff. Water Science and Technology, 46(11-12), 413-418.
Garnaud, S.; Mouchel, J.M.; Chebbo, G.; and Thevenot, D.R. (1999) Heavy Metal Concentrations In Dry And Wet
Atmospheric Deposits In Paris District: Comparison With Urban Runoff. Sco. Total Environ.. 235, 1-3, 235.
Garnaud, S.; Mouchel, J-M.; and Thevenot, D.R. (1999) Mobility Evolution of Particulate Trace Metals in Urban
Runoff: From Street Deposits and Road Runoff to Combined Sewer Deposits and Catchment Outlet. Proc. the
Eighth International Conference on Urban Storm Drainage. August 30 – September 3, 1999, Sydney, Australia.
Edited by IB Joliffe and JE Ball. The Institution of Engineers Australia, The International Association for
Hydraulic Research, and The International Association on Water Quality, 1511.
Ghani, A.; Salem, A.M.; Abdullah, R.;Yahaya, A.S.; Zakaria, N.A. (1999) Incipient Motion of Sediment Particles
Over Loose Deposited Beds in a Rigid Rectangular Channel. Proc. the Eighth International Conference on Urban
Storm Drainage. August 30 – September 3, 1999, Sydney, Australia. Edited by IB Joliffe and JE Ball. The
Institution of Engineers Australia, The International Association for Hydraulic Research, and The International
Association on Water Quality, 157.
Gietz, R. J. Urban Runoff Treatment in the Kennedy-Burnett Settling Pond. For the Rideau River Stormwater
Management Study, Pollution Control Division, Works Department, Regional Municipality of Ottawa-Carleton,
Ottawa, Ontario, March 1983.
Glenn, D.W.; Tribouillard, T.; and Sansalone, J.J. (2001a). Temporal Deposition of Anthropogenic Pollutants in
Urban Snow as Generated by Traffic. WEFTEC 2001 Conf. Proc. CD-ROM.
Glenn, D.W.; Tribouillard, T.; and Sansalone, J.J. (2001b). Temporal Heavy Metal Delivery and Partitioning in
Urban Snow Exposed to Traffic. WEFTEC 2001 Conf. Proc. CD-ROM.
Gray, P.; Ward, C.; Chan, S.; and Kharadi, F. (2000) CSO Planning: Integrated Collection System and Wastewater
Treatment Plant Hydrodynamic Simulation. Collection Systems Wet Weather Pollution Control: Looking into
Public, Private, and Industrial Issues, May 2000, Rochester, NY. Water Environment Federation, CD-ROM.
Greb, S.R., and Bannerman, R.T. (1997) Influence of Particle Size on Wet Pond Effectiveness. Water Environ. Res.,
69, 6, 1134.
Grizzard, T. J., C. W. Randall, B. L. Weand, and K. L. Ellis. “Effectiveness of extended detention ponds.”
Engineering Foundation Conference: Urban Runoff Quality - Impact and Quality Enhancement Technology,
Henniker, New Hampshire, edited by B. Urbonas and L. A. Roesner, published by the American Society of Civil
Engineers, New York, June 1986.
Grout, H.; Wiesner, M.R.; and Bottero, J.Y. (1999) Analysis of Colloidal Phases in Urban Stormwater Runoff.
Environ. Sci. Technol. (G.B.), 33, 6, 831.
Hijioka, Y.; Frurumai, H.; and Nakajima, F. (2000) Runoff Pollutant Load of Suspended Solids with Different
Particle Sizes in a Separate Sewer System. Proc. 3rd JSWE Symposium, September 2000, Neyagawa, Japan.
Japanese Society of Water Environment (JSWE).
Hittman Assoc. Methods to Control Fine-Grained Sediments Resulting from Construction Activity. U.S.
Environmental Protection Agency, Pb-279 092, Washington, D.C., December 1976.
Hodgins, D.O.; Hodgins, S.L.M.; and Corbett, R.E. (2000) Modeling Sewage Solids Deposition Patterns for the Five
Fingers Island Outfall, Nanaimo, British Columbia. Watershed 2000 Management Conference, July 2000,
Vancouver, British Columbia. Water Environment Federation, CD-ROM.
72
House, L.B., R.J. Waschbusch, and P.E. Hughes. Water Quality of an Urban Wet Detention Pond in Madison,
Wisconsin, 1987-1988. U. S. Geological Survey, in cooperation with the Wisconsin Department of Natural
Resources. USGS Open File Report 93-172, 1993.
Hunt, C.L., and Grow, J.K. (2001). Impacts of Sedimentation to a Stream by a Non-Compliant Construction Site.
Proc. ASCE EWRI Conf. - Bridging the Gap: Meeting the World’s Water and Environmental Resources
Challenges. CD-ROM.
Johnson, P.D., R. Pitt., S.D. Durrans., M. Urrutia., and S. Clark. Innovative Metal Removal Technologies for Urban
Stormwater. Alexandria, VA: Water Environmental Research Foundation, WERF-97-IRM-2. 2003.
Johnston, R.J., and Swallow, S.K. (1999) Asymmetries in Ordered Strength of Preference Models: Implications of
Focus Shift for Discrete-Choice Preference Estimation. Land Economics. 75, 2, 295.
Jones, B. H. and Washburn, L. (1998) Storm Water Runoff into Santa Monica Bay: Identification, Impact and
Dispersion. Proc. California and the World Ocean ‘97, San Diego, Calif., ASCE, 888.
Kamedulski, G. E. and R. H. McCuen. “The effect of maintenance on stormwater detention basin efficiency.” Water
Resources Bulletin, Vol. 15, No. 4, pg. 1146, August 1979.
Kepner, W.G.; Semmens, D.J.; Bassett, S.D.; Mouat, D.A.; Goodrich, D.C. (2004). Scenario analysis for the San
Pedro River, analyzing hydrological consequences of a future environment. Environ. Monitor. Assess. 94(13):115-127.
Keshavarzy, A., and Ball, J.E. (1999) Initiation of Sediment Motion at the Bed With Turbulent Shear Stresses in an
Open Channel Flow. Proc. the Eighth International Conference on Urban Storm Drainage. August 30 –
September 3, 1999, Sydney, Australia. Edited by IB Joliffe and JE Ball. The Institution of Engineers Australia,
The International Association for Hydraulic Research, and The International Association on Water Quality, 164.
Khambhammettu, U. (2006) Evaluation of Upflow Filtration for the Treatment of Stormwater. M.S. Environmental
Engineering, The University of Alabama, Tuscaloosa, AL.
Kirby, Jason. Determination of Vegetal Retardance in Grass Swale. M.S. thesis. University of Alabama. 2003.
Kokkonen, T.; Koivusalo, H.; Karvonen, T.; Croke, B.; Jakeman, A. (2004). Exploring streamflow response to
effective rainfall across event magnitude scale. Hydrol. Process. 18(8):1467-1486.
Konrad, C.P.; Booth, D.B.; Burges, S.J. (2005). Effects of urban development in the Puget Lowland, Washington,
on interannual streamflow patterns: Consequences for channel form and streambed disturbance. Water Resour.
Res. 41(7):W07009.
Krein, A., and Schorer, M. (2000) Road Runoff Pollution by Polycyclic Aromatic Hydrocarbons and its
Contribution to River Sediments. Water Res. (G.B.). 34, 4110.
Krishnappan, B.G.; Marsalek, J.; Watt, W.E.; and Anderson, B.C. (1999) Seasonal Size Distributions of Suspended
Solids in a Stormwater Management Pond. Water Sci. Technol. (G.B.), 39, 2, 127.
Lager, J.A., W.G. Smith, W.G. Lynard, R.M. Finn, and E.J. Finnemore. Urban Stormwater Management and
Technology: Update and Users' Guide. U.S. Environ. Protection Agency, EPA-600/8-77-014, 313 p. Cincinnati,
Ohio, September 1977.
Lee, G.F., and Jones-Lee, A. (1996a) Evaluation of the Water Quality Significance of the Chemical Constituents in
Aquatic Sediments: Coupling Sediment Quality Evaluation Results to Significant Water Quality Impacts. Proc.
WEFTEC‘96, 69th Annu. Conf. & Expo., Dallas, TX, Water Environ. Fed., 8-49.
Liang, W.Y.; Henry, D.; Grgic, D. (1998) Evaluation of Ontario’s Storage Sizing Criteria for Wet Pond Design.
Proc. Water Environ. Fed. 71st Annu. Conf. Exposition, Orlando, Fla., 2, 667.
Lindsay, W. L. Chemical Equilibria in Soils. John Wiley and Seno, New Trk. 1979.
Lloyd, S.D., and Wong, T.H.F. (1999) Pariculates, Associated Pollutants and Urban Stormwater Treatment. Proc.
the Eighth International Conference on Urban Storm Drainage. August 30 – September 3, 1999, Sydney,
Australia. Edited by IB Joliffe and JE Ball. The Institution of Engineers Australia, The International Association
for Hydraulic Research, and The International Association on Water Quality, 1833.
Lysne, D.K. “Hydraulic design of self-cleaning sewage tunnels.” Journal of Sanitary Engineering Division of ASCE,
Vol. 95, 1969.
Magnuson, M.L.; Kelty, C.A.; and Kelty, K.C. (2001). Trace Metal Loading on Water-Borne Soil and Dust Particles
Characterized Through the Use of Split-Flow Thin Cell Fractionation. Analytical Chemistry. 73:3492.
Mailhot, A., Rousseau, A.N., Salvano, E., Turcotte, R., and Villeneuve, J.-P. (2002). Evaluation of the Impact of a
Municipal Clean Water Program on Water Quality of the Chaudiere River Watershed Using the Integrated
Modeling System GIBSI. Revue des Sciences de l’Eau 15: 149.
Mark, O., Weesakul, S., and Hung, N.Q. (2002). Modeling the Interaction Between Drainage System, Wastewater
Treatment Plant and Receiving Water in Pattaya Beach. Proc. 9th Int. Conf. Urban Drainage – Global Solutions
for Urban Drainage. CD-ROM.
73
Maurer, D.E.; Groff, C.D.; Karmasin, B.M.; Perrine, G.L.; Kelly, S.D.; Ross, P.E. (2004). Innovative technologies
of Detroit’s large CSO control facility. WEFTEC 2004 Conf. Proc. Water Environment Federation. CD-ROM.
May, R.W.P Sediment Transport in Pipes and Sewers with Deposited Beds. HR Wallingford, LTD., Wallingford,
Report SR320, 1993.
McCuen, R.H. Hydrologic Analysis and Design, 2nd Edition. Prentice Hall. 1998.
McGregor, J.D. (2003). The case for cleaning. Civil Eng. 73(1):60-63.
McIlhatton, T.D.; Sakrabani, R.; Ashley, R.M.; Burrows, R. (2002). Erosion mechanisms in combined sewers and
the potential for pollutant release to receiving waters and water treatment plants. Water Science and Technology,
45(3), 61-69.
Miller, W.; Boulton, A.I. (2005). Managing and rehabilitating ecosystem processes in regional urban streams in
Australia. Hydrobiologia. 552(1):121-133.
Morquecho, R. (2005) Pollutant Associations with Particulates in Stormwater. Ph.D. Dissertation, The University of
Alabama, Tuscaloosa, AL.
Morrill, J. (1999) Sound, Sewer Reach Truce. InTech. 46, 7, 56.
Morrisey, D.J.; Williamson, R.B.; Van Dam, L.; and Lee, D.J. (2000) Stormwater Contamination of Urban
Estuaries. 2. Testing a Predictive Model of the Build-Up of Heavy Metals in Sediments. Estuaries. 23, 67.
Moses, T.; Lower, M. (2005). Reconstructing streams: Conventional designs give way to natural approach. Public
Works. 136(5):69-70.
Mosley, L.M., and Peake, B.M. (2001). Partitioning of Metals (Fe, Pb, Cu, Zn) in Urban Run-Off From the Kaikorai
Valley, Dunedin, New Zealand. New Zealand J. Marine Freshwater Res. 35:615.
Nalluri, C. and A. Ab-Ghani. “Bed load transport with deposition in channels of circular cross section.” Proceedings
of the Sixth International Conference on Urban Storm Drainage. Niagra Falls, Canada. 1993.
Nalluri, C. and E.M. Alvarez. “ The influence of cohesion on sediment behavior.” Water Science Technology 1992
Nara, Y. and Pitt, R. (2005) Alabama Highway Drainage Conservation Design Practices - Particulate Transport in
Nara, Y. Sediment Transport in Grass Swales. Master of Science, Environmental Engineering thesis, Department
of Civil and Environmental Engineering, The University of Alabama, Tuscaloosa, Alabama. 2005.
Grass Swales and Grass Filters. University Transportation Center for Alabama, Project Number 04117. The
University of Alabama. Tuscaloosa, AL.
Nelson, E. (1999) Sediment Budget of a Mixed-use, Urbanizing Watershed. Proc. AWRA’s 1999 Annual Water
Resources Conference - Watershed Management to Protect Declining Species, December 1999, Seattle, WA.
American Water Resources Association, 469.
Nelson, E.J.; Booth, D.B. (2002). Sediment sources in an urbanizing, mixed land-use watershed. Journal of
Hydrology, 264(1-4), 51-68.
Newman, T.L., II; Leo, W.M.; and Gaffoglio, R. (2000a) Characterization of Urban-Source Floatables. Collection
Systems Wet Weather Pollution Control: Looking into Public, Private, and Industrial Issues, May 2000,
Rochester, NY. Water Environment Federation, CD-ROM.
Parker, J.T.C.; Fossum, K.D.; and Ingerssol, T.L. (2000) Chemical Characteristics of Urban Stormwater Sediments
and Implications for Environmental Management, Maricopa County, Arizona. Environ. Manage. 26, 99.
Paul, J.F.; Comeleo, R.L.; Copeland, J. (2002). Landscape metrics and estuarine sediment contamination in the midAtlantic and southern New England regions. Journal of Environmental Quality, 31(3), 836-845.
Persyn, R.A.; Glanville, T.D.; Richard, T.L.; Laflen, J.M.; Dixon, P.M. (2004). Environmental effects of applying
composted organics to new highway embankments: Part 1. Interrill runoff and erosion. Trans. ASAE. 47(2):463469.
Persyn, R.A.; Glanville, T.D.; Richard, T.L.; Laflen, J.M.; Dixon, P.M. (2005). Environ. effects of applying
composted organics to new highway embankments: Part III. Rill erosion. Trans. Am. Society Agricultural
Engineers. 48(5):1765-1772.
Petterson, T.J.R. (1999) The Effects of Variations of Water Quality on the Partitioning of Heavy Metals in a
Stormwater Pond. Proc. the Eighth International Conference on Urban Storm Drainage. August 30 – September
3, 1999, Sydney, Australia. Edited by IB Joliffe and JE Ball. The Institution of Engineers Australia, The
International Association for Hydraulic Research, and The International Association on Water Quality,1943.
Pettersson, T.J.R. (2002). Characteristics of suspended particles in a small stormwater pond. Global Solutions for
Urban Drainage, Proc. of the Ninth Int. Conf. on Urban Drainage, Sept 8-13 2002, Portland, OR, CD-ROM.
Pisano, W.C. (1996) Summary: United States Sewer Solids Settling Characterization Methods, Results, Uses, and
Perspective. Water Sci. and Technol., 33, 9, 109.
Pisano, W.C., and Brombach, H. (1996) Solids Settling Curves: Wastewater Solids Data Can Aid Design of Urban
Runoff Controls. Water Environ. Technol., 8, 4, 27.
74
Pisano, W.C., J. Barsanti, J. Jopyce, and H. Sorensen. Sewer and Tank Sediment Flushing: Case Studies.
EPA/600/R-98/157. Water Supply and Water Resources Division, U.S. Environmental Protection Agency,
Cincinnati, OH Dec. 1998.
Pisano, W.C.; O’Riordan, O.C.; Ayotte, F.J.; Barsanti, J.R.; Carr, D.L. (2003). Automated sewer and drainage
flushing systems in Cambridge, Massachusetts. J. Hydraul. Eng. 129(4):260-266.
Pitt, R. Demonstration of Nonpoint Pollution Abatement Through Improved Street Cleaning Practices. U.S. EPA.
Grant No. S-804432. EPA-600/2-79-161. 270 pages. Cincinnati, August 1979.
Pitt, R. and G. Shawley. A Demonstration of Non-Point Source Pollution Management on Castro Valley Creek.
Alameda County Flood Control and Water Conservation District (Hayward, CA) for the Nationwide Urban
Runoff Program, U.S. Environmental Protection Agency, Water Planning Division, Washington, D.C., June 1982.
Pitt, R. and R. Sutherland. Washoe County Urban Stormwater Management Program; Volume 2, Street Particulate
Data Collection and Analyses. Washoe Council of Governments, Reno, Nevada, August 1982.
Pitt, R. Characterizing and Controlling Urban Runoff through Street and Sewerage Cleaning. U.S. EPA. Contract
No. R-805929012. EPA/2-85/038. PB 85-186500/AS. 467 pages. Cincinnati, June 1985.
Pitt, R., and J. McLean. Toronto Area Watershed Management Strategy Study: Humber River Pilot Watershed
Project. Ontario Ministry of the Environment, Toronto, Ontario, 1986.
Pitt, R., R. Field, M. Lalor, and M. Brown. “Urban stormwater toxic pollutants: Assessment, sources and
treatability.” Water Environment Research. Vol. 67, No. 3, pp. 260-275. May/June 1995.
Pitt, R., B. Robertson, P. Barron, A. Ayyoubi, and S. Clark. Stormwater Treatment at Critical Areas: The MultiChambered Treatment Train (MCTT). U.S. Environmental Protection Agency, Wet Weather Flow Management
Program, National Risk Management Research Laboratory. EPA/600/R-99/017. Cincinnati, Ohio. 505 pgs. March
1999.
Pitt, R.; Harrison, R.; Lantrip, J.; and Henry, C.L. (1999) Infiltration Through Disturbed Urban Soils and CompostAmended Soil Effects on Runoff Quality And Quantity. Proc. Water Environ. Fed. 72nd Annu.Conf. Exposition,
[CD-ROM], New Orleans, LA.
Pitt. R. S. Clark, J. Lantrip, and J. Day. Telecommunication Manhole Water and Sediment Study;
Vol. 1: Evaluation of Field Test Kits (483 pgs);
Vol. 2: Water and Sediment Characteristics (1290 pgs);
Vol. 3: Discharge Evaluation Report (218 pgs);
Vol. 4: Treatment of Pumped Water (104 pgs).
Bellcore, Inc., Special Report SR-3841. Morriston, NJ. With further support from NYNEX, BellSouth,
Bell Atlantic, GTE, SNET, Pacific Bell, US West, Ameritech and AT&T. December 1998.
Pitt, R. and Khambhammettu, U. Field Verification Tests of the UpFlowTM Filter. Small Business Innovative
Research, Phase 2 (SBIR2) Report. U.S. Environmental Protection Agency, Edison, NJ. 2006
Pollert, J.; Stansky, D. (2002). Combination of computational techniques - Evaluation of CSO efficiency for
suspended solids separation. Global Solutions for Urban Drainage, Proc. of the Ninth Int. Conf. on Urban
Drainage, Sept 8-13 2002, Portland, OR, CD-ROM.
Pollert, J.; Stransky, D. (2003). Combination of computational techniques - evaluation of CSO efficiency for
suspended solids separation. Water Sci. Technol. 47(4):157-166.
Prakash, A. (2005). Impact of urbanization in watersheds on stream stability and flooding. Proc. 2005 Watershed
Management Conf. - Managing Watersheds for Human and Natural Impacts: Engineering, Ecological, and
Economic Challenges. Am. Society of Civil Engineers. 1681-1688.
Randall, C. W., K. Ellis, T. J. Grizzard, and W. R. Knocke. “Urban runoff pollutant removal by sedimentation.”
Urban Runoff Pollutant Removal by Sedimentation", Proceedings of the Conference on Stormwater Detention
Facilities, Planning, Design, Operation, and Maintenance, Henniker, New Hampshire, Edited by W. DeGroot,
published by the American Society of Civil Engineers, New York, August 1982.
Rate, A.W.; Robertson, A.E.; and Borg, A.T. (2000). Distribution of Heavy Metals in Near-Shore Sediments of the
Swan River Estuary, Western Australia. Water, Air, Soil Pollut. 124, 155.
Rhoads, B.L., and Cahill, R.A. (1999) Geomorphological Assessment of Sediment Contamination in an Urban
Stream System. Applied Geochemistry. 14, 4, 459.
Ristenpart, E. and M. Uhl. “Behavior and pollution of sewer sediments.” Proceedings of the Sixth International
Conference on Urban Storm Drainage. Niagra Falls, Canada. 1993.
Rochfort, Q.; Grapentine, L.; Marsalek, J.; Brownlee, B.; Reynoldson, T.; Thompson, S.; Milani, D.; and Logan, C.
(2000) Using Benthic Assessment Techniques To Determine Combined Sewer Overflow and Stormwater Impacts
in the Aquatic Ecosystem. Water Qual. Res. J. Can. 35, 365.
75
Rose, S., and Peters, N.E. (2001). Effects of Urbanization on Streamflow in the Atlanta Area (Georgia, USA): a
Comparative Hydrological Approach. Hydrol. Processes. 15:1441.
Rossi, L.; Krejci, V.; Rauch, W.; Kreikenbaum, S.; Fankhauser, R.; Gujer, W. (2005). Stochastic modeling of total
suspended solids (TSS) in urban areas during rain events. Water Res. 39(17):4188-4196.
Rushforth, P.J.; Tait, S.J.; and Saul, A.J. (1999) Importance of Sediment Composition on Erosion of Organic
Material from In-sewer Deposits. Proc. the Eighth International Conference on Urban Storm Drainage. August
30 – September 3, 1999, Sydney, Australia. Edited by IB Joliffe and JE Ball. The Institution of Engineers
Australia, The International Association for Hydraulic Research, and The International Association on Water
Quality, 515.
Sakrabani, R.; McIlhatton, T.; Chebbo, G.; Ashley, R.; Oms, C. (2002). Near Bed Solids in combined sewers.
Global Solutions for Urban Drainage, Proc. of the Ninth Int. Conf. on Urban Drainage, Sept 8-13 2002, Portland,
OR, CD-ROM.
Sansalone, J. (1996) Immobilization of Metals and Solids Transported in Urban Pavement Runoff. North American
Water and Environ. Congress ‘96, Anaheim, CA, ASCE, NY.
Sansalone, J. J., and Buchberger, S.G. (1997) Partitioning and First Flush of Metals in Urban Roadway Storm
Water. J. Environ. Eng., 123, 2, 134.
Sansalone, J.J., and Hird, J.P. (1999) A Passive Source Control Prototype For Heavy Metals Transported by Urban
Storm Water. Proc. Water Environ. Fed. 72nd Annu. Conf. Exposition, [CD-ROM], New Orleans, LA.
Sansalone, J.; Mishra, S.K.; and Singh, V.P. (2001). The Role of Anthropogenic Hydrology on the Transport,
Partitioning and Solid-Phase Distribution of Heavy Metals in Urban Storm Water – Implications for Treatment.
WEFTEC 2001 Conf. Proc. CD-ROM.
Sedlak, D. L., Phinney, J. T. and Bedsworth, W. W. (1997) Strongly Complexed Cu and Ni in Wastewater Effluents
and Surface Runoff. Environ. Sci. Technol., 31, 10, 3010.
Shafer, M.M.; Overdier, J.T.; Phillips, H.; Webb, D.; Sullivan, J.R.; and Armstrong, D.E. (1999) Trace Metal Levels
and Partitioning in Wisconsin rivers. Water, Air, Soil Pollut.. 110, 3-4, 273.
Shakoor, A.; Smithmyer, A.J. (2005). An analysis of storm-induced landslides in colluvial soils overlying mudrock
sequences, southeastern Ohio, USA. Engineering Geology. 78(3-4):257-274.
Skipworth, P.S.; Tait, S.J.; and Saul, A.J. (1999) Erosion of Sediment Beds in Sewers: Model Development. J.
Environ. Eng.. 125, 6, 566
Smith, W. G. “Water quality enhancement through stormwater detention.” Proceedings of the Conference on
Stormwater Detention Facilities, Planning, Design, Operation, and Maintenance, Henniker, New Hampshire,
Edited by W. DeGroot, published by the American Society of Civil Engineers, New York, August 1982.
Sutherland, R.A.; Tack, F.M.G.; Tolosa, C.A.; and Verloo, M.G. (2000). Operationally Defined Metal Fractions in
Road Deposited Sediment, Honolulu, Hawaii. J. Environ. Qual. 29, 1431.
Terstriep, M.L., G.M. Bender, D.C. Noel. Nationwide Urban Runoff Project, Champaign, Illinois: Evaluation of the
Effectiveness of Municipal Street Sweeping in the Control of Urban Storm Runoff Pollution. Contract No. 1-539600, U.S. Environmental Protection Agency, Illinois Environmental Protection Agency, and the State Water
Survey Division, University of Illinois, December 1982.
Tetzlaff, D.; Grottker, M.; Leibundgut, C. (2005). Hydrological criteria to assess changes of flow dynamic in urban
impacted catchments. Physics Chemistry Earth. 30(6-7 SPEC.):426-431.
Vignoles, M., and Herremans, L. (1966) Motorway Pollution: Total Suspended Solids in Runoff Water and Settling
Tank Design Approach. Proc. WEFTEC‘96, 69th Annu. Conf. & Expo., Dallas, TX, Water Environ. Fed., 4-167.
Vollersten, J.; Hvitved-Jacobsen, T.; McGregor, I.; and Ashley, R. (1998) Aerobic Microbial Transformations of
Pipe and Silt Trap Sediments from Combined Sewers. IAWQ 19th Biennial Int. Conf., Vancouver, Can., 8, 244.
Weng, Q.H. (2001). Modeling Urban Growth Effects on Surface Runoff with the Integration of Remote Sensing and
GIS. Environ. Man. 28:737.
Whipple, W. and J. V. Hunter. “Settleability of urban runoff pollution.” Journal WPCF, Vol. 53, No. 12, pg. 1726,
1981.
Willems, P. and Berlamont, J. (2002b). Probabilistic Emission and Immission Modeling: Case-Study of the
Combined Sewer – WWTP-Receiving Water System at Dessel (Belgium). Water Science and Technology 45:117.
Williams, A.T., and Simmons, S.L. (1997a) Estuaries Litter at the River/beach Interface in the Bristol Channel,
United Kingdom. J. Coastal Res. (U.K.), 13, 4, 1159.
Williams, A.T., and Simmons, S.L. (1997b) Movement Patterns of Riverine Litter. Water, Air, Soil Pollut. (Neth.),
98, 1-2, 119.
Williams, A.T., and Simmons, S.L. (1999) Sources of Riverine Litter: The River Taff, South Wales, UK. Water, Air,
& Soil Pollution, 112, 1-2, 197.
76
Wissman, R.C.; Timm, R.K.; Logsdon, M.G. (2004). Effects of changing forest and impervious land covers on
discharge characteristics of watersheds. Environ. Manage. 34(1):91-98.
Wolff, T., Benton, S., Cheung, P. (2002). Integrating Collection System and Wastewater Treatment Plant Hydraulic
Modeling for Wet Weather Control. WEF/CWEA Collection Systems 2002 Conf. Proc., CD-ROM.
Yao, K.M. “ Sewer line design based on critical shear stress.” Journal Environmental Engineering Division of
ASCE, Vol. 100. 1974.
Yeager, M.A.; Wise, D.D.; Suffet, I.H. (2005). Understanding and managing urbanization-induced stream erosion in
Southern California. California and the World Ocean '02: Revisiting and Revising California's Ocean Agenda Proc. Conf., CWO '02, 2005. Am. Society of Civil Engineers. 1407-1421.
Zug, M., Gommery, L., Engel, N., Pawlowski-Ruesing, E. (2002). Integrated Sewerage System Methodology and
Modeling: from Grand Couronne to Berlin. Proc. 9th Int. Conf. Urban Drainage – Global Solutions for Urban
Drainage. CD-ROM.
77