United States
                    Environmental Protection
                    Agency
Water Engineering Research
Laboratory
Cincinnati OH 45268
                    Research and Development
EPA/600/S2-85/038  June 1985
&EPA          Project Summary
                    Characterizing  and Controlling
                    Urban  Runoff Through
                    Street  and Sewerage  Cleaning
                    Robert Pitt
                      A study was conducted in Bellevue,
                    Washington, to characterize urban run-
                    off and evaluate its control by street and
                    sewerage cleaning. The project was one
                    of a series conducted in the city from
                    1978  through  1983 to investigate
                    Bellevue's urban runoff sources, effects,
                    and potential controls.
                      The  project reported here spanned
                    the period 1980-1983, and it complete-
                    ly monitored more than 300  urban
                    runoff events in two residential areas
                    during that time. Flow-weighted com-
                    posite samples were analyzed for a core
                    list of important constituents. Complete
                    flow monitoring results allowed detailed
                    descriptions of urban runoff quality and
                    quantity, and they permitted estimates
                    concerning the contributions of flows
                    and pollutants from  different source
                    areas.  Street surface and sewerage
                    particulates  were also collected and
                    analyzed to determine the effectiveness
                    of street and sewerage cleaning.
                      Most of the heavy metals were deter-
                    mined to originate from street dirt, but
                    street cleaning improved the quality of
                    urban runoff by a maximum of only 10
                    percent. A specially modified  street
                    cleaner was then tested and found to be
                    much more effective than the conven-
                    tional  model in removing the small
                    particles of street dirt that are washed
                    off the streets  by rains. Catchbasin
                    cleaning twice a year was estimated to
                    improve runoff quality by a maximum of
                    about 25 percent.
                      This Project Summary was developed
                    by EPA's Water Engineering Research
                    Laboratory, Cincinnati, OH, to announce
                    key findings of the research project that
is fully documented in a separate report
of the same title (see Project Report
ordering information at back).

Introduction
  A 2-year  monitoring  program was
conducted to examine the sources of
urban runoff flows  and pollutants in
Bellevue, Washington. The study of these
elements enabled a comprehensive in-
vestigation of the direct effects of street
and sewerage cleaning on runoff quality.
A large number of storm events were
monitored under two extreme street-
cleaning frequencies to investigate pos-
sible  improvements in  urban runoff
quality. Urban runoff was studied in two
residential areas using automatic, flow-
weighted  samplers  and  sonic depth
gauges to monitor runoff during storms
and dry weather.  Street  dirt samples
were obtained in conjunction with specific
street cleaning programs using special
vacuum collection procedures. Storm
sewer inlet  and sewerage particulate
samples were also obtained periodically.
  Significant decreases in street surface
loadings occurred with intensive, three-
times-a-week street cleaning, but large
improvements in the quailty of urban
runoff were not detected. The improve-
ments averaged  only about 10 percent,
possibly because the light Bellevue rains
removed only some  15 percent of the
street loadings.  The particulates that
washed off the streets were of the finer
particle sizes that are not effectively
removed by conventional street cleaners.
These particulates constituted only  a
small portion of the total urban yields of
many pollutants, but street dirt washoff is

-------
a major contributor of many heavy metals
and organic priority pollutants. In addition,
the amounts of rainfall at the two loca-
tions differed by more than 20 percent at
least half of the time. These rain differ-
ences made basin calibrations difficult
and tended to  mask variations in runoff
concentrations or yields that may have
been caused by street cleaning.


Study Area  Description
  Bellevue is a middle to upper middle
class suburban community near Seattle,
Washington. The city is decentralized,
with residential areas served by shopping
malls and numerous businesses along
arterial streets. No heavy industry exists
in Bellevue. The population growth has
been rapid in the past decade, and the
population was about 74,000 in 1980.
Though the growth of Bellevue has mostly
been in residential areas, recent develop-
ment has included construction of addi-
tional office buildings and hotels. The
area is  within commuting distance of
Seattle.  Bellevue receives about 1  m/yr
of rain, but substantially more rain falls
on the Olympic Peninsula to the west,
and much smaller amounts occur to the
east in Washington.
  Two residential areas were studied in
this project—Surrey Downs and Lake Hills
(Figure 1). The communities are about 5
km apart, and each covers  an area of
about 40 ha. Both are fully developed,
mainly as single family residences.
  The Surrey Downs basin is about 38 ha
in size and includes the Bellevue Senior
High School in addition to single family
homes that were built in the late 1950's.
Most  of the  slopes in the basin are
moderate, with some steeper slopes on
the west side.  The Surrey  Downs basin
ranges in elevation from about 3 to 55 m,
and about 60  percent of it is pervious.
Back and front  yards make up most of the
land surface area of the basin, and streets
make up about 10 percent. The streets
are generally  in  good condition, with
smooth  to  intermediate  textures.  The
curbs need repairing in a few locations.
Westwood Homes Road and 108th Street
have no curbs. The Surrey Downs basin
has little traffic, and the on-street parking
density is low. The storm drainage system
discharges into an artificial pond located
in an adjacent development. This pond
discharges  into  Mercer Slough,  which
eventually drains to Lake Washington
and Puget Sound.
  The Lake Hills catchment covers about
41 ha and contains the St. Louise parish
church and school in addition to single
                                                                 Redmond
                                                  Larsen Lake
                                                   Phantom Lake
                                                                         N
Figure 1.   City of Bellevue, Washington, study sites.
family  homes,  also  developed in the
1950's. Lake  Hills has  slightly  more
pervious area than Surrey Downs, but its
lots are typically smaller. With  a few
exceptions, the slopes in Lake Hills are
more  moderate than those  found in
Surrey Downs. The elevation of the Lake
Hills study area ranges from 80 to 125 m.
The street surfaces and  gutter systems
are similar to those  in  Surrey Downs.
Most of the streets in Lake H ills also carry
low volumes of traffic  and  have low
parking densities,  except for  two busy
roads that cross through the  area. The
Lake Hills storm drainage system outfalls
into a short open channel that joins Kelsey
Creek just downstream from Larsen Lake.
Kelsey Creek also discharges into Mercer
Slough.
Sources of Runoff and Pollutants
  Sources and amounts of runoff, flows,
and receiving  water conditions are all
affected by site-specific rain conditions.
Bellevue rains are quite  different from
those in most other U.S. locations. They
occurred every 2 or 3 days during the
study period and were on the average less
than 6 mm each. Fewer than 10 percent
had volumes greater than 25 mm, and the
largest rain monitored was 100 mm. Dry
periods of more than a week are rare, but
they did occur. Storms during the wet
season generally yield twice the amount
of rain and last twice as long as those
during the dry season.
  This project monitored about 400 rains
and all base flow volumes between eventsi
that occurred at the two main study"

-------
locations during 2 years of data collection.
Bellevue received about 1 m of rain during
each project year. Base flows represented
relatively large portions of the total annual
urban flows.
  Important differences between the two
study sites occurred in quantity of rainfall,
base flow, and runoff yields. Overall, the
base plus stormwater urban flows from
Lake Hills were about 18 percent greater
than those from Surrey Downs  when
normalized by area (figured on an equal
area basis).
  For both study years and test basins,
only about 25 percent of the rain that fell
left the areas as  runoff. Typically, the
small rains had the smallest runoff factors
(Rv: ratio of runoff volume to rain volume)
and the large rains had the largest factors.
  Multiple regression  analyses  were
conducted to relate the Rv values to total
rain, average and peak intensity, and days
since  last  rain. Results showed that
rainfall volumes alone accounted for
about  95  percent  of the individual Rv
values. The  season  of the year was
extremely important in determining actual
runoff and  rainfall  relationships. The
winter wet months of November through
February had Rv values some 35 percent
higher than those for the drier months of
March through October for similar rains.
Thus, there was no real need to adjust the
calculated Rv values for rain intensity or
length of preceding dry period. All that
needed to be considered was total rainfall
and season.
  A model was  developed to determine
the sources of  runoff and their propor-
tionate contributions. The model was
based  on the variations of Rv values for
different rain volumes. Source of runoff
considered included vacant  lots,  parks,
front and backyards, rooftops, driveways,
parking lots, and streets. The amount of
runoff contributed by each  source de-
pended on its distance from the drainage
system, its size, and the type of surface
cover.
  For all rains greater than about 2.5mm,
impervious surfaces contributed  more
than 60 percent of  the total urban runoff
flows. The remainder of the flows were
approximately evenly divided  between
front and  back yards. Vacant  lots and
parks contributed very little flow because
of their limited presence in the test areas.
Street  surfaces contributed about 25
percent of the total urban flows for most
rains causing runoff.
Quality of Stormwater Runoff
  Collecting data on the quality of storm-
water runoff was a major aspect of this
project. Most of the analytical effort was
associated with a core list of important
constituents. The sampling  procedures
involved  collecting  total storm  flow-
weighted composite samples throughout
most of the events that occurred during
the 2-year sampling period at the Surrey
Downs and Lake Hills sites. Variations in
stormwater quality with total  storm char-
acteristics were analyzed. Very few vari-
ations were observed in the  total solids
concentrations for various storm char-
acteristics, and they were  statistically
insignificant.
  The runoff water quality  at Bellevue
was much better than that at most other
U.S. locations, but the base flow quality
was worse than  expected. The  reason
was probably that the study basins were
completely urbanized and the base flows
consisted of percolated urban sheet flow
waters from previous storms that were
draining out of the surface soils. In basins
with undeveloped upstream areas,  the
base flow would  originate mostly from
the nonurbanized upper  reaches and
would be of much better quality.
  The mass yields for annual urban runoff
(Table 1) indicated an apparent difference
between  the runoff  in  Lake Hills and
Surrey Downs when expressed on a unit
            area basis, but the total annual storm
            runoff plus base flow discharges from the
            two basins were quite similar. A much
            larger fraction of the total urban runoff in
            Surrey Downs occurred as base  flow
            between rain events. The runoff events in
            Lake Hills were more sharply defined, and
            the base flows made up a much smaller
            fraction of the mass yields for urban
            runoff.
              The relative contributions of pollutants
            from various source areas (Table 2)
            differed from the contributions  of runoff
            flows. During very small rains, most of
            the runoff and pollutant discharges were
            associated with the  directly connected
            impervious areas. As the  rain  total in-
            creased (to greater than about 2.5 mm),
            the pervious areas became much more
            important. These patterns varied signifi-
            cantly,  depending on specific rain char-
            acteristics and land uses. For most  rain
            events, total solids originated mostly from
            the back and front yards. Street surfaces,
            however,  were expected to account for
            most of the lead,  zinc, and COD  dis-
            charges. Phosphates and total  Kjeldahl
            nitrogen were mostly contributed from
            street  surfaces, driveways, and parking
            lots combined. Front and back yards
            contributed slightly less than half of these
Table 1.    A nnual Mass Yields for Baseflow and Stormwater Runoff (kg/ha)
                                  Surrey Downs
                                 Lake Hills
Constituent
Base
Flow
Storm
Runoff
Total
Base
Flow
Storm
Runoff
Total
Total solids
COD
Total Kjeldahl nitrogen
Total phosphorus
Lead
Zinc
110    205    315
 11      90    100

  0.60     1.8     2.4

  0.11     0.40    0.51

  0.03     0.26    0.29
  0.06     0.24    0.30
76     280     360

 9.9    110     120
 0.20    2.7     2.9

 0.04    0.69    0.73

 0.02    0.45    0.47
 0.027   0.31    0.34
Table 2.   Percentage Distribution of Urban Runoff Pollutants from Various Source Areas*

                              Percent of Pollutant Contributed from Source Area*
Source Area
Streets
Driveways and parking lots
Rooftops
Front yards
Back yards
Vacant lots and parks
Total
Solids
9
6
<1
44
39
2
COD
45
27
3
13
12
<1
Phosphates
32
21
5
22
20
<1
Total
Kjeldahl
Nitrogen Lead
31 60
20 37
10 <1
19 <1
20 <1
<1 <1
Zinc
44
28
24
2
2
<1
"For 2.5- to 65-mm rains.
^Approximate.

-------
nutrients. Zinc contributions from roof-
tops (galvanized gutters) made up about a
fourth of the total zinc discharges.

Contributions of Street Dirt to
Urban Runoff Discharges
  About  600 samples of street surface
accumulations were collected  from  the
test areas during the 2-year project. The
particulate loads for each sample were
plotted  to observe changes  in street
surface loadings with time and to deter-
mine the initial rates of deposition and
long-term accumulation. The deposition
rate is the amount of dirt that accumulates
over the first several days after a signifi-
cant rain or street cleaning. This rate is a
function  of various characteristics of the
area; especially climate, land use, traffic
and street surfacetexture. The accumula-
tion  rate equals  the  amount of  dirt
deposited minus the amount removed by
rain, street cleaning, traffic-induced tur-
bulence or wind. Material blownfromthe
street can remain suspended in the  air,
but most of it settles to the ground within
about 10 m of the roadway.
  Each of the street surface samples was
separated  into  eight different particle
sizes. These size distributions showed
that the  smallest particle sizes account
for  only  a small fraction of  the total
material, especially during the wet season
when the rains were most effective in
removing the smallest  particles. During
the dry season, the larger particle sizes
also accounted for relatively small frac-
tions of the total solids weight. Most of
the street surface particulates were asso-
ciated with particles in the middle size
ranges of 0.125 to 1.0 mm.
  The initial accumulation rates (assumed
to be equal to the deposition rates) in the
test areas were estimated to vary between
1 and 6  g/curb-meter  per day, with an
average  rate of about 3 g. This rate
compares with  accumulation rates  ob-
served in  other locations for smooth
streets in good condition. The frequent
rains do  not remove all of this material
from the  streets. The texture of the street
traps  and protects particulates so that
typical street cleaning  equipment and
rains cannot remove them. About 50 to
100 g/curb-meter of street surface par-
ticulates  remain  on the  streets after
storms of about 6 mm  or greater. Infre-
quent large rains may remove much more
of the street surface particulates than the
smaller rains common in Bellevue.
  Chemical characteristics of street  dirt
in the different study areas varied most
with respect to  lead,  especially when
comparing streets with varying traffic
levels. Particle size also had a significant
effect on concentrations of chemical
pollutants in street dirt. Chemical oxygen
demand, Kjeldahl nitrogen, and phospho-
rus concentrations all showed high con-
centrations associated with the smallest
particle  sizes, small concentrations with
the intermediate sizes, and high concen-
trations with the large sizes. Lead and
zinc concentrations were the highest with
the smallest particle sizes which is typical
(based on other studies for heavy metals).
  The contribution of the street surface
particulates to runoff water depends on
the ability of the rain to loosen and wash
these particulates from the street surface.
During the 2-year project about 25 pairs
of street  surface loading  values were
obtained within 2 days of  rain. Figure 2
shows the percent and size  distribution of
surface  particulates washed from the
street in Lake Hills during both the dry
and wet seasons. The  initial loadings
were significantly greater than the resid-
ual loadings  for  particle  sizes  smaller
than about 500 microns.  The smallest
particle  sizes had the greatest significant
washoffs, whereas particles greater than
about 500 microns had lower significant
washoff values. With the smallest particle
sizes, the washoffs varied from about 40
to 50 percent, whereas increases were
found in street  loadings for the larger
                             particle sizes. The overall reduction in net
                             loading averaged about 16 percent. Even
                             more material may have been removed
                             during the rains and replaced at the same
                             time by erosion material.
                               Most of the material washed from the
                             street surfaces had particle  sizes of less
                             than about 125  microns. Only about 10
                             percent  of  the  washoff  material  was
                             greater than about 500  microns.  The
                             largest particle sizes were notably absent
                             from the washoff material. A  total of
                             about 8.5 to 10 g/curb-meter was re-
                             moved  from the street surfaces during
                             the rains, with about 4 to 6 g/curb-meter
                             having  particle  sizes smaller than  125
                             microns.
                               Table 3 summarizes the  approximate
                             annual street dirt accumulations, washoff
                             values, and fugitive losses to air for the
                             Surrey  Downs and Lake  Hills sites. In
                             many cases, the amounts involved were
                             substantially greater for Lake Hills than
                             for Surrey Downs. About 15 percent of
                             the annual street dirt accumulation  was
                             washed from  the streets  and either
                             discharged or accumulated in the sewer-
                             age systems. About  10 percent of  the
                             annual accumulation was lost to the air,
                             with much  of this material  settling  out
                             near the roadway. The remaining street
                             dirt would build up over time  on the street
                             surface or be removed by street cleaning
                             operations.
   -10
 Figure 2.
          6350
                               1000   2000   6350
                   Panicle Sizes (microns}
                                                                        Total
Percent and size distribution of surface particulates washed from the street in Lake
Hills during the wet and dry seasons.

-------
Table 3.   Approximate Annual Street Dirt Accumulation, Washoff, and Fugitive Losses to Air
      "   (kg/ha)
Surrey Downs
Constituent
Total solids
COD
Total Kjeldahl nitrogen
Total phosphorus
Lead
Zinc
Accumu-
lation*
200
25
0.2
0.1
0.1
0.03
Washoff
30
3
0.04
0.02
0.02
0.006
Loss
to Air*
20
2
0.02
0.01
0.01
0.003
Accumu-
lation*
350
70
0.8
0.2
0.4
0.08
Lake Hills
Washoff
60
10
0.16
0.05
0.1
0.02

Loss
to Air*
20
3
0.05
0.01
0.02
0.005
*Using an average 2- to 5-day accumulation period.
^Calculated based on the deposition minus the accumulation rates times the average interevent
 time period.
Sediment Accumulations in
Sewerage Systems and
Catchbasins
  Sewerage system sediment loadings
were periodically observed in the Surrey
Downs and Lake Hills study areas. The
drainage systems were cleaned before
the project  began,  and  the sediment
volumes in inlets and catchbasins were
observed nine times during 2 years. The
first observations  in  December 1979
showed light accumulations. The next
observations were made in August 1980.
Beginning in January 1981, observations
were made every 1 or 2 months until the
end of  the project.  The first  year of
observations indicated steady accumula-
tions of sediment, but the  loading re-
mained about the same during the second
year. Typically, about twice as  much
polluted sediment was observed in  the
storm drainage systems at any given time
as was noted on the streets. The flushing
of the sewerage sediments out of  the
drainage systems and into the receiving
waters was not analyzed, but such an
event would probably not occur except
during large storms. The smaller storms
probably removed a small fraction of the
sewerage sediments.
  The  stable  sediment  volumes that
occurred during the second year were
about 60 percent of the  available sump
volumes of the catchbasins and inlets.
Only about 12 mm of sediment was found
in the manholes with outlets  on  the
structure bottoms, whereas about 150
mm of sediment accumulated in the inlets
and catchbasin sumps. When analyses
were conducted for individual structures,
wide variations were observed. The depth
below the outlet appeared to be the most
important factor, but the larger-capacity
sumps did not always contain the largest
amount of sediment. Large sump capacity
would allow less frequent cleaning before
the stable volume  was  obtained  and
smaller outlet-to-sump bottom distances
would be associated with greater scouring
during storms.
  The chemical quality of the sediment
material  in  the  catchbasins  was also
analyzed.  The chemical quality of  this
catchbasin and inlet sump material was
very similar to that of the  same sized
particles of street dirt. Thus most of the
catchbasin  sediments were probably
street  surface particulates  that  had
washed off the streets during rains but
were not discharged to the outfall.
  About 100 liters/ha per year accumu-
lated in the Surrey Downs storm sewer
inlet structures, whereas only about two-
thirds  of  this amount accumulated in
Lake Hills. About  50 percent more inlet
structures per hectare exist in Lake Hills
as opposed to Surrey Downs, where the
accumulation rate per inlet structure was
generally more than double that of Surrey
Downs. Nine of the ten most  heavily
loaded catchbasins in the first summer
inventory for Surrey Downs were located
on  or  just downstream from the  two
streets in the study area that did not have
curbs and had extensive off-street sedi-
ment sources.
  Very few pipes in either Surrey Downs
or Lake Hills had slopes of  less than 1
percent. Since  both  study areas were
drained by steeply sloping pipe systems,
the accumulation of sediments in the
storm drainage systems was not great.

Urban Runoff Controls
  The last phase in developing an urban
runoff  control program is to examine
measures that can be used to reduce the
identified  problem pollutants or flows
originating from the various source areas
that discharge to the receiving water. To
meet water quality objectives, a combina-
tion of several control measures may be
necessary. Complex procedures for ana-
lyzing decisions may also be necessary if
multiple objectives are  important. This
part of the study evaluated the effective-
ness of street cleaning in controlling the
urban runoff problems in Bellevue.
  Street cleaning tests were conducted
using two different cleaning frequencies:
No cleaning, and intensive 3-day-a-week
cleaning. For several months, one clean-
ing frequency was used in the Surrey
Downs main basin, and the other was
used in the Lake Hills basin. The frequen-
cies were then rotated.  During another
period of several months, no street clean-
ing was conducted in either basin.
  Street loadings  ranged from  about 40
to 300 g/curb-meter  (with an average
value of about 115) during the  period of
no  street  cleaning. The loadings  were
reduced to about 20 to 200 g/curb-meter
(with an average of about 60) after street
cleaning (from about 650 to 400 microns)
because of the selective removal of the
largest  particles by the street cleaners.
Rains,  on the other hand, increased the
median particles sizes because they were
most  effective in removing the  finer
material.
  This  study collected more than 400
street dirt samples in the two test basins
immediately before and after the streets
were cleaned. Figure 3 compares the
initial and residual street dirt loads for a
wide range of loading  conditions. Street
cleaning equipment cannot remove par-
ticulates from the street surface unless
the loadings are above a certain  level.
This value was about 85 g/curb-meter in
the test basins. If the initial street surface
loading values were smaller than this
value, then the residual loadings typically
were about equal  to the initial loadings.
Statistical analysis showed that the fre-
quent rains in Bellevue were  probably
more effective than the street cleaning in
keeping the streets cleaned. The street
surface loadings after rains were usually
50to 100g/curb-meter. Typical  mechan-
ical street cleaning equipment was quite
ineffective in removing small particles.
  Particle sizes smaller than about 350
microns were not substantially affected
by street  cleaning,  and there  was no
effective removal of street dirt  particles
smaller than about  125 microns. Very
subtantial removals were observed in the
large  particles, however. A decrease
occurred in median particle sizes as the
street  cleaners preferentially  removed
larger particles. These decreases  were
especially important  when  the  initial
median particle sizes were large.

-------
  A series of special tests were conducted
during September and October of 1982 to
compare the effectiveness of a  modified
street cleaner  with  that  of a  standard
mechanical street cleaner. Many modifi-
cations were made to a standard Tymco*
regenerative air street cleaner.  The pur-
pose of these modifications was to reduce
respirable fugitive dust emissions during
street cleaning.  The modifications in-
cluded partial hoods around the gutter
brooms,  a  pressure  controller  to better
regulate the air flows, and a venturi
scrubber with a settling chamber in the
street cleaner hopper. The water  spray
bar was also disconnected. This modified
street cleaner was compared with both a
standard street cleaner that was used in
the previous full-scale tests and with an
unmodified version of itself. The results
of these special tests are  also shown in
Figure 3. Both the modified and unmodi-
fied regenerative air  street  cleaners
showed substantially better performance
than the regular mechanical street clean-
er, especially for finer particle sizes. The
poor  performance  of the  mechanical
street cleaners  was aggravated by the
low loadings of these small particles. The
regenerative air cleaners were  much
more suited to the low loadings of small
particles.
  Bellevue street cleaning  costs  were
about $13/curb-kilometer. About 73 per-
cent of  this  cost was  associated  with
labor and labor overhead,  about 18 per-
cent was  associated with  all of the
maintenance costs, and 18 percent was
associated with the disposal costs. Street
cleaner operating costs (including labor,
depreciation, tires, oil, and gasoline) were
about 64 percent of the total costs.
  A large part of the data  analysis done
during this project attempted to identify
differences in runoff concentrations and
yields caused by different street cleaning
operations.  No  significant  differences
were observed in runoff yields or concen-
trations during periods of intensive street
cleaning. A very few exceptions occurred,
but they were  probably  due  to  other
factors.
  The poor street cleaning effectiveness
is probably the  result  of the specific
Bellevue rain conditions. The rainfall and
resulting runoff volumes  varies greatly
between the two test areas—by at least
25 percent for about half of the rain pairs.
These differences could  have  shielded
the effects of the different street cleaning
operations.
'Mention of trade names  or commercial products
does not constitute endorsement or recommenda-
tion for use.
   800
   700--
             • Mobil (Surrey Downs)
             a Mobil {Lake Hills)                                     -,
             T: Tymco (Surrey Downs) standard regenerative air street cleaner™'
            M: Modified Tymco (Surrey Downs) regenerative air street
               cleaner
   100
       0      100      200      300      400     500
                                Initial Load (Ib/ curb-mile)

Figure 3.   Street dirt loads before and after street cleaning.
                                                          600
                                                                   700
                                                                            800
  Though street surface particulates con-
tributed less than 25 percent to runoff
yields in nearly all cases, they contributed
about 50 percent of the total runoff lead
yield. If the street cleaning operations
could control a substantial fraction of the
street surface particulates, then reducing
runoff particulates by  street cleaning
might  be important.  Rains were most
effective in removing  particles  smaller
than several  hundred microns  in size.
These particle sizes are not abundant, but
they do contain the largest concentrations
of heavy metals and relatively  large
concentrations of many  nutrients.  As
noted, however, mechanical street clean-
ing equipment is not  very effective  in
removing small particles. The regenera-
tive air street cleaners were more effec-
tive in this area.
  The  coordination of street  surface
sampling, street cleaning operations, and
runoff monitoring during this  project
allowed many data analysis procedures
to be used to investigate possible effects
of street cleaning on runoff water quality.
The use of two test basins and the  rotation
of the street  cleaning operations also
allowed one basin to be compared with
the other, along with internal basin
comparisons. No  significant differences
were noted in the runoff concentrations
over the ranges of data that were common
to the various data sets.
  Intensive street cleaning resulted  in
about a 25- to 50-percent reduction  in
                                         street  surface loadings.  If the  street
                                         surface contributes about half of the total
                                         runoff yield for a specific pollutant, then
                                         intensive street cleaning may remove 10
                                         to 20 percent of the pollutant discharge.
                                         Precise runoff measurements  and con-
                                         sistent rainfalls over the test and control
                                         basins would  therefore be  required to
                                         detect  these relatively  small  improve-
                                         ments. Intensive street cleaning signifi-
                                         cantly  reduced only the  large particle
                                         sizes,  and those  particle  sizes  most
                                         subject to washoff by rains were not
                                         effectively reduced. This may result in
                                         less  than  a  6 percent improvement in
                                         runoff water quality for intensive street
                                         cleaning.  The  regenerative air  street
                                         cleaner is expected to be about 1.25 times
                                         more effective in reducing runoff yields.

                                         Conclusions
                                           Direct  receiving  water effects from
                                         urban runoff pollutants were not signifi-
                                         cant for most storms, but potential long-
                                         term problems associated  with  urban
                                         runoff may be associated with settleable
                                         solids, lead, and  zinc.  These settled
                                         materials may have silted up spawning
                                         beds and introduced high concentrations
                                         of potentially toxic materials directly to
                                         the stream sediments. The oxygen deple-
                                         tion  observed in the interstitial waters
                                         was probably caused by organic sediment
                                         buildup from runoff events.
                                           Flooding in the  receiving  waters has
                                         increased significantly with urbanization.

-------
This flooding has affected several benefi-
cial  uses of these waters (aquatic life
habitat  and water conveyance, for ex-
ample).
  For all rains greater than about 2.5 mm,
the  impervious surfaces (streets, side-
walks, driveways, parking lots, and roof-
tops) contributed more than 60 percent of
the total urban runoff flows. The remain-
der  of  the flows were approximately
evenly divided between front and back
yards. Vacant lots and parks contributed
very little to the flows because of their
limited  presence in the test  areas.  For
most of the rain events monitored, the
street surfaces  contributed  about  25
percent of the total urban runoff flows.
  Most  of  the total solids in  the  urban
runoff originated from  front and back
yards in the test areas; the street surfaces
contributed only a small fraction. Lead,
zinc,  and COD,  however, were mostly
contributed from street  surfaces.  Nutri-
ents  (phosphorus and total Kjeldahl
nitrogen) originated mostly from  street
surfaces, driveways, and parking lots.
  Motor vehicle activity was expected to
be the primary contributor of most of the
toxic pollutants. Gasoline and  diesel fuel
combustion products,  lubricant and fuel
leakages, and wear of the vehicles  af-
fected the street dirt material most signif-
icantly.
  The maximum observed runoff event
discharged about 25 percent of all pollu-
tants that were on the street surfaces, the
catchbasins, and sewerage  combined.
Thus most urban runoff pollutants were
not source-limited. However,  those par-
ticulates that  were most  available  for
washoff (the smallest particles) might be
source-limited. About half of the total
particulates from annual  urban runoff
discharge might be residing on the street
surfaces and tied up on catchbasins and
storm drainage sediments at  any given
time. If  the Bellevue rain  events could
remove  more of this material, the  urban
runoff discharges would be much greater
than observed.
  Rains removed only a small  fraction of
the  total  particulate  loadings on  the
impervious surfaces (about 15 percent).
Large particles were not effectively re-
moved, and only about half of the smallest
particles (less than 50 microns) were
washed off during  rains.  These  small
particles were not very abundant,  but
they  had very high concentration of heavy
metals and nutrients. Most of the settled
particulates in the storm drainage inlets
and sewerage pipes also remained after
the observed storms.
  Intensive street cleaning three times a
week produced significant decreases in
street surface loadings—from about 115
g/curb-meter down to about 60 g/curb-
meter. The median  particle  sizes  also
decreased significantly with  intensive
street cleaning. A regenerative air street
cleaner performed substantially better in
removing the  finer street surface mate-
rials than did the regular mechanical
street cleaner.
  Extensive data  analysis showed no
significant improvements in runoff water
quality during periods of intensive street
cleaning. The  street cleaning operations
tested are expected to improve runoff
quality by a maximum of only 10 percent.
The street cleaning equipment preferen-
tially removed the larger particles,  and
the rain events removed the finer mate-
rials. Street cleaning did  not very effec-
tively remove  the available particulates.
Mechanical-broom street cleaning effec-
tively removed the larger litter from the
streets. Infrequent street cleaning may
result in significant increases in fugitive
dust losses to  the atmosphere.
  After an initial cleaning,  nearly a full
year was needed for sediment to reach a
stable volume  in the inlet structures. Only
about 60 percent of the  total available
sump volumes in inlets and catchbasins
were used to  detain particulates at the
stable  volume.  At any larger  storage
levels, the rains effectively controlled the
volumes. Cleaning the inlets and catch-
basin sumps about twice a year is expect-
ed  to reduce  the  lead and total solids
concentrations in urban runoff by 10 to
25 percent. COD, the nutrients, and zinc
might be reduced by 5 to 10 percent.
  Based  on this project, many recom-
mendations can be made about public
works practices in the Bellevue area, but
their  effects  on  improving the urban
runoff  quality would probably be quite
small. If  intensive  street cleaning were
implemented  along  with  semiannual
catchbasin sediment cleaning, most pol-
lutants in urban runoff discharges would
be reduced by about 10 percent. Some of
the heavy metal discharges  might be
reduced by as  much as 25 percent. Even
though these reductions are quite small,
they could contribute significantly to
reducing the accumulation of these highly
polluted sediments in the smaller creek
systems, especially if the receiving water
flows were reduced.
  Peak runoff flows could be reduced by
requiring the use of more pervious areas
in developed  areas,  or by the use of
appropriately sized and located detention
basins.
  The full report was submitted in fulfill-
ment of Cooperative Agreement No. CR-
805929 by the City of Bellevue under the
sponsorship of the  U.S.  Environmental
Protection Agency.
                                                                                        * U.S GOVERNMENTPRINTINOOFFICE: 19M- 559-111/10858

-------
     Robert Pitt is a Consulting Engineer, Blue Mounds, Wl 53517.
     Richard Field is the EPA Project Officer (see below}.
     The complete report,  entitled  "Characterizing and Controlling Urban Runoff
       Through Street and Sewerage Cleaning, "(Order No. PB 85-186 500/AS; Cost:
       $35.50, subject to change) will be available only from:
            National Technical Information Service
            5285 Port Royal Road
            Springfield, VA 22161
            Telephone: 703-487-4650
     The EPA Project Officer can be contacted at:
            Wastewater Research Division
            Water Engineering Research Laboratory—Cincinnati
            U.S. Environmental Protection Agency
            Edison, NJ 08837
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45^68
     BULK RATE
POSTAGE & FEES PAID
        EPA
  PERMIT No. G-35
Official Business
Penalty for Private Use $300
                                      OC63240   N£RL
                                      LCU  W  TILIEY
                                      REGION  V
                                      LIBRARIAN
                                                            ST
                                                               It-    60604

-------