United States
Environmental Protection
Agency
National Risk Management
Research Laboratory
Research Triangle Park, NC 27711
Research and Development
EPA/600/SR-95/171 January 1996
EPA Project Summary
Characterization of Mud/Dirt
Carryout onto Paved
Roads from Construction and
Demolition Activities
Michael M. Raile
Several urban areas of the country in
violation of the National Ambient Air
Quality Standard (NAAQS) for particu-
late matter have identified fugitive dust
generated by vehicular traffic on paved
streets and highways resulting from
mud/dirt carryout from unpaved areas
as a primary source of PM-10 (particles
< 10 u,m in aerodynamic diameter).
Since little data are currently available
on the amount of mud/dirt carryout de-
posited on paved roads, this work char-
acterizes the process and evaluates
selected control methods. These con-
trol technologies were evaluated for ef-
fectiveness in controlling mud/dirt
carryout from an unpaved construction
access area onto an adjacent paved
road. The first control used a street
sweeper to mechanically sweep the dirt
and debris from the paved road sur-
face. The second applied a 6- to 12-in.
(15- to 30-cm) layer of woodchip/mulch
material onto the access area of the
construction site to a distance 100 ft
(30 m) from the paved road. The third
applied a 6-in. layer of gravel over the
access area. Street sweeping was found
to be only marginally effective (approxi-
mately 20%) in reducing average silt
loading on the paved road lanes. Treat-
ment of the access area with a buffer
of woodchip/mulch was moderately ef-
fective, reducing average silt loading
by 38 to 46%. The gravel buffer showed
the greatest effectiveness, reducing the
average silt loading by 57 to 68%. These
silt loading reductions result in the fol-
lowing calculated PM-10 reductions:
street sweeping, 14%; woodchips, 27
to 33%; and gravel, 42 to 52%.
This Project Summary was developed
by EPA's National Risk Management
Research Laboratory's Air Pollution
Prevention and Control Division, Re-
search Triangle Park, NC, 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
Several areas of the country that are in
violation of the National Ambient Air Qual-
ity Standard (NAAQS) for PM-10 (particles
<10u.m in aerodynamic diameter) have
conducted studies to identify the sources
of these emissions. A primary source of
PM-10 in many urban areas is the fugitive
dust generated by vehicular traffic on
paved streets and highways.
Road dust emissions occur whenever a
vehicle travels over a paved surface, such
as public and industrial roads and parking
lots. Particulate emissions originate pri-
marily from the road surface material load-
ing (measured as mass of material per
unit area). The surface loading is in turn
replenished by other sources (e.g., pave-
ment wear, deposition of material from
vehicles, deposition from other nearby
sources, carryout from surrounding un-
paved areas, and litter). Because of the
effects of the surface loading, available
control techniques attempt to either (a) pre-
vent material from being deposited on the
surface or (b) remove (from the travel
lanes) any material that has been depos-
ited.
-------
According to the Environmental Protec-
tion Agency (EPA) publication, Compila-
tion of Air Pollutant Emission Factors
(AP-42), the quantity of dust emissions
from vehicle traffic on a paved public road
(per vehicle kilometer traveled or VKT)
may be estimated using the empirical ex-
pression:
E = 4.6 (sL/2)065 (W/3)15
where: E = PM-10 emission factor
(g/VKT),
s = surface silt content (fraction
of particle < 75 urn in
physical diameter),
L = total road surface dust
loading (g/m2), and
W = average weight (tons) of the
vehicle traveling on the
road.
Activities such as construction and
demolition projects can create a tempo-
rary, but substantial, increase in the
amount of fine particles on the surfaces of
adjacent paved roads. This increase in
fine particle loading is the result of mud/
dirt carryout from vehicles leaving the con-
struction/demolition site.
Furthermore, carrying out material onto
a paved road is characterized by substan-
tial spatial variation in loading about the
point of access to the site. This variation
complicates the estimation of emissions
caused by carryout as well as the emis-
sion reductions achievable by control of
carryout. The spatial variations and the
associated difficulties in estimating emis-
sions become less important as the num-
ber of access points in an area increases.
A prior field study specifically addressed
mud/dirt carryout onto urban paved roads.
It was conducted in 1982 as part of a
national demonstration study of construc-
tion-related dust emissions.
This report describes a field study un-
dertaken to better understand the mecha-
nisms of mud/dirt carryout as well as the
effectiveness of measures used to control
carryout. The study collected and ana-
lyzed surface material samples taken from
a paved road adjacent to a construction
site in the Brush Creek Flood Control
Project in the metropolitan area of Kansas
City, MO. The effects of mud/dirt carryout
control were evaluated by monitoring the
changes in paved road surface dust load-
ing. Both preventive and mitigative mea-
sures for controlling carryout were consid-
ered. Preventive measures attempted to
keep material from being deposited on
roadways, while mitigative measures at-
tempted to remove the material after be-
ing deposited. The mitigative control mea-
sure of interest in this study was com-
bined water flushing and broom sweep-
ing. The two preventive measures that
were studied involved covering the ac-
cess area with a coarse material (gravel
and woodchips/mulch).
Site Description
The paved roadway segment that was
selected for this study was the north-south
section of Elmwood Avenue between Blue
Parkway and Brush Creek Boulevard in
east central Kansas City, MO. The road-
way segment is approximately 1200 ft (365
m) long and 40 ft (12 m) wide. It carries
an annual average daily traffic volume of
-10,000 vehicles and is classified as a
minor arterial roadway. At the time of this
study, a pocket of construction activity as-
sociated with the Brush Creek Flood Con-
trol Project was located on the east side
of Elmwood Avenue. This activity notice-
ably impacted Elmwood Avenue in the
mud/dirt carryover.
With the construction of a dam and other
earthmoving activities, the Elmwood site
was expected to provide enough truck traf-
fic to support a field sampling program
throughout the summer of 1994. Ten-wheel
dump trucks carried earth from the site,
south on Elmwood, and then on to their
final destination. On days that it rained,
the trucks could not enter the site due to
an incline near the site entrance that be-
came too muddy to support vehicles safely.
On those days, the trucks were directed
to other sites where they could work.
Road Surface Sampling
This field sampling program was de-
signed to efficiently collect paved road
surface material samples at various dis-
tances between 50 and 550 ft (15 and 170
m) from the construction site entrance on
Elmwood Avenue over an extended time
period. From previous studies of silt load-
ing on paved roads, it was known that the
loadings could vary from one lane to the
next. For this reason, a sampling scheme
allowed for segregation of the two south-
bound lanes. If necessary, at the end of
the data reduction process, the data from
the two southbound lanes could be inte-
grated with each other to represent just
the southbound portion of the roadway.
One area of southbound Elmwood Av-
enue, just north of the access point, was
designated for the collection of background
silt loading samples that ideally would not
be impacted by carryout from the con-
struction site. Samples of the material on
the road surface were collected by dry
vacuuming and then analyzed for silt con-
tent according to the procedures outlined
in AP-42.
Source Activity Monitoring
Source extent and activity data were
also collected in the sampling program.
Vehicle-related parameters were acquired
using a combination of manual and auto-
matic recording techniques. Pneumatic
tube axle counters were used to obtain
traffic volume data. However, because
these counters recorded only the number
of passing axles, it was necessary to ob-
tain traffic mix information (e.g., number
of axles per vehicle) to convert axle counts
to the number of vehicle passes. Vehicle
mixes were observed visually.
Daily weather data were obtained from
a local newspaper, and rainfall measure-
ments were made on site with a rain
gauge. A daily log was also maintained,
noting any activities that were observed at
the site or any communications that were
pertinent to the outcome of the project.
Study Conditions
In addition to the uncontrolled study con-
dition, three different mud/dirt carryout con-
trols were evaluated in the program: street
sweeping, installation of a woodchip/mulch
apron (buffer) at the site access point,
and installation of a gravel buffer at the
same point. These controls were evalu-
ated sequentially as described below.
First, the paved road adjacent to the
site (Elmwood Avenue) was cleaned to
the extent practical using a combination
of broom sweeping and flushing. This
cleaning represented a "baseline" silt load-
ing value for future reference. (In this con-
text, "baseline" refers to as clean a road
surface as possible.)
Once baseline levels were reached, the
surface loading was allowed to increase
to its "steady state" condition, with sam-
pling conducted before and after precipi-
tation throughout the "conditioning" period.
Post-precipitation sampling was performed
once the road surface became dry enough
to collect surface samples. The data from
these samples established both the mag-
nitude and extent of the uncontrolled mud/
dirt carryout from the site and provided a
time history of the overall carryout pro-
cess starting from an essentially clean
surface.
When the uncontrolled tests were com-
pleted, the paved road was thoroughly
cleaned again (sweeping and flushing) to
baseline condition prior to evaluation of
street sweeping as the first control method.
-------
Thereafter, the road was swept periodi-
cally, using the fleet of street sweepers
that were already being used to control
carryout in the immediate vicinity. Sweep-
ing occurred every other workday, except
on days that it rained. Surface sampling
was conducted at approximately the same
time periods before and after precipita-
tion, as was done for the uncontrolled
sampling, to determine the overall reduc-
tion in silt loading.
Prior to evaluating the second control
method, the paved road was again ag-
gressively cleaned (sweeping and flush-
ing) to reestablish the baseline condition.
Coordinated with the cleaning, the
woodchip/mulch material was applied in a
6- to 12-in. layer to the site access point
and adjacent areas to provide a 100-ft
buffer between the paved and unpaved
surfaces. The buffer allowed the mud/dirt
carried on the truck tires and underbodies
to deposit in the buffer area, rather than
on the paved road. Reductions in silt load-
ing were then quantified by appropriate
surface sampling at comparable time peri-
ods before and after precipitation.
Finally, after the paved road was
cleaned again (sweeping and flushing) to
the baseline condition, the previously in-
stalled woodchip/mulch buffer area was
replaced with a gravel buffer (100 ft long
and 6 in. deep). Comparable surface sam-
pling was performed before and after
precipitation in a manner similar to that
described above.
Note that the original test plans had
called for the evaluation of street sweep-
ing, the gravel buffer, and an asphalt
buffer. However, the construction site su-
pervisor responded to the city's request
for controlling on-site fugitive dust by us-
ing a woodchip/mulch buffer. Because the
woodchip/mulch buffer was an actual con-
trol measure chosen by the contractor, it
was decided to evaluate this buffer's ef-
fectiveness.
By the time that the gravel buffer was to
be evaluated, traffic into and out of the
site had been reduced and was not ex-
pected to pick up until concrete pouring
began in earnest at the dam site. Be-
cause of schedule constraints, it was jointly
decided by the EPA work assignment man-
ager and the contractor to generate "cap-
tive" traffic to complete the evaluation of
the gravel buffer. The "captive" traffic that
was used was a 10-wheeled truck that
was identical to the type of trucks that
were originally hauling material from the
site. The captive truck was half-loaded to
represent the average of a loaded and an
unloaded condition.
Time History of the Project
The field sampling portion of the pro-
gram spanned from the end of May until
mid-September 1994. The long time span
was the result of many delays caused by
the weather. There were consecutive days
when no sampling activity occurred. Dur-
ing those days, there was either no haul-
ing activity at the site or it had rained the
evening before. This prevented any trucks
from entering the site and also prevented
surface samples from being taken.
The sampling program was also affected
by the trucks' originally leaving the site
going south and later began exiting to the
north. This change in direction caused
problems in maintaining a constant flow of
traffic from the test site over the south-
bound lanes of Elmwood. Also, because
of interference from a power pole, trucks
exiting north from the construction site
were forced to swing into the southbound
lane, thereby impacting the silt loading
background area.
Traffic into and out of the site slowed
after the earthmoving activities were largely
completed and concrete pouring for the
dam had not yet begun. Therefore, supple-
mental "captive" traffic was generated by
hiring a truck and driver to drive into and
out of the test site for the remaining por-
tion of the field sampling program; i.e., the
evaluation of the gravel buffer.
Data Analysis/Results
The silt loading data were plotted as a
function of distance from the site access
point. Because earlier studies found a rapid
decrease in silt loading with distance, the
data were plotted on a semi-logarithmic
graph. The data were grouped by control
technology (e.g., uncontrolled, woodchip/
mulch) to construct a single plot for each
control. The data were then regressed for
each lane and for each control method.
The uncontrolled silt loadings did not ex-
hibit any discernible trend to increase with
time. This is probably due to the fact that,
because of the steep slope of the access
road, no truck haulage occurred for 1 or 2
days after rainfall. Thus, at the study site,
precipitation did not enhance carryout onto
Elmwood, but rather rainfall at least par-
tially cleaned the road surface. The trucks
were diverted to haul from other sites in
the area during these periods.
The area under the silt loading distribu-
tion curve was determined using the
trapezoidal rule for each sampling event.
When the resulting area was divided by
500 ft, the average silt loading (sL) was
found. The average silt loadings measured
for the uncontrolled condition ranged from
2.6 to 8.9 g/m2 for the "A" lane and from
1.0 to 5.8 g/m2 for the "B" lane. These
ranges correspond approximately to the
upper 20th-percentile of the silt loading
data base presented in AP-42. Thus,
carryout clearly resulted in heavy silt load-
ings on Elmwood Avenue. In addition, the
curb or "A" lane was roughly twice as
heavily loaded as the other southbound
lane ("B"). This was expected because
the loaded trucks tended to travel almost
exclusively in the "A" lane in preparing to
turn west onto Blue Parkway.
Because of problems encountered in
defining an appropriate "background value"
of silt loading (as a result of the impacts
of construction vehicles occasionally exit-
ing to the north on Elmwood), the control
efficiency of reduction in sL (presented in
Table 1) is in terms of a range between a
lower bound and an upper bound. The
lower bound was obtained by assuming a
background value of zero for silt loading.
The upper bound was obtained by sub-
tracting the relatively high background
value of 0.5 g/m2. This high background
value was determined from an average
daily traffic value of 10,000 using the rela-
tionship between silt loading and traffic
volume.
The overall control efficiency for street
sweeping was found to be 19 and 27% for
the "A" and "B" lanes^respectively, based
on the reduction in ~sL Street sweeping
was found to be much more effective in
reducing total surface loadings. Part of
the poor performance for removal of silt
loading can be attributed to the abrasion
of coarse material left on the roadway
after sweeping; i.e., the sweeper gener-
ated additional material in the silt fraction
by breaking large particles into smaller
ones. In addition, the same sweeper al-
ready was being used to control carryout
from construction sites in the immediate
vicinity.
The woodchip/mulch buffer proved to
be more effective than the street sweeper
in reducing average silt loading. Over the
two sampling events, an average control
efficiency of 33 to 37% for the "A" lane
and 49 to 64% for the "B" lane was found
for the woodchip/mulch buffer. This con-
trol measure is of considerable interest
because it represents a "real world" solu-
tion to the problems of carryout in that the
contractor constructed a buffer from waste
material collected on-site. This resulted in
a far more cost-effective (i.e., reduction in
silt loading per unit cost) control than street
sweeping.
An important potential drawback was
observed during the use of woodchip/
-------
Table 1. Silt Loading Control Efficiencies
Control
A Lane
B Lane
Both Lanes
(lower limit/upper limit)
Street sweeping
Woodchip/mulch application
Gravel application
19/22%
33/37%
56/64%
21/27%
49/64%
58/76%
20/24%
38/46%
57/68%
mulch. Because woodchip/mulch is soft
and easily compressed by vehicles, the
weight of a passing vehicle will displace
the air contained in the buffer. This effect
caused substantial fugitive dust clouds that
could be seen when a vehicle traveled
over the buffer. Although the buffer was
effective in controlling the carryout of ma-
terials from the site, on-site reentrained
fugitive dust vehicular emissions may have
increased due to the woodchip/mulch
buffer.
The gravel buffer was found to be the
most effective control studied in this re-
port, reducing average silt loadings by 56
to 64% in the "A" lane and 58 to 76% in
the "B" lane. In addition, unlike the
woodchip/mulch material, the gravel buffer
probably reduced on-site reentrained fugi-
tive dust vehicle emissions by covering
the travel surfaces with a coarser mate-
rial.
Quality Assurance (QA) Results
Eight sets of co-located surface samples
were collected (using the "embedded" sam-
pling approach) and analyzed. The regu-
lar and the QA samples yielded compa-
rable results for total loading, silt loading,
and overall silt content (as determined by
dividing the total loading by the silt load-
ing). The silt loading range percent values
fell well within the +50% guideline set
forth in the test plan, with an overall mean
of 17% in absolute value of range per-
cent.
A total of 24 QA samples were obtained
by riffle splitting of field samples of road
surface loading. Each subsample was
taken through sieve analysis. The QA sta-
tistic (relative value, RV) for 17 of the 24
pairs (71%) fell within the +0.05 guideline
established in the test plan, with an absolute
maximum of 0.154 for Sample 2-B-R-593 (i.e.,
having a "regular" silt content of 10.5% con-
trasted with a QA value of 9.3%). In gen-
eral, larger absolute values of the QA
statistic are associated with lower silt frac-
tions.
Conclusions
The testing and evaluation program that
was conducted led to several conclusions.
• A broad range of paved road silt load-
ings were measured near the con-
struction site access point under the
uncontrolled condition and each con-
trolled condition, but no condition ex-
hibited clearly discernible time trends.
In other words, silt loadings did not
tend to increase with time. This may
be the result of rainfall partially clean-
ing the surface between different sam-
pling events and largely reducing ac-
cess road traffic until the steep slope
had dried. Once access point traffic
was restored to its normal level,
reentrainment and displacement to
nontraveled parts of the road (i.e.,
curb) offset the additional loading from
carryout so that a new "equilibrium"
was established.
• Street sweeping was found to be only
marginally effective (approximately
20%) in lowering average paved road
silt loading values in carryout areas.
In general, total loadings were re-
duced far more effectively, but the
street sweeper appears to have
abraded the remaining material, thus
"creating" additional material in the
silt fraction.
The 6- to 12-in. layer of woodchip/
mulch was moderately effective in
controlling carryout, with average
paved road silt loadings being reduced
38 to 46%. This control measure was
implemented by the construction con-
tractor at the request of Kansas City
officials. Furthermore, the control
made use of material that was avail-
able on-site at no cost. Although the
woodchip/mulch buffer was moder-
ately effective in controlling off-site
emissions, it was noted that this buffer
may have increased on-site PM-10
emissions. The buffer was fairly "soft"
and was readily compressed by ve-
hicles traveling over it. This compres-
sion displaced the trapped air, and
puffs of fugitive dust were observed.
The 6-in. layer of gravel was found to
reduce average paved road silt load-
ings by 57 to 68%. This was the high-
est efficiency found in the present
study. Unlike the other buffer mate-
rial, gravel formed a far stronger sur-
face that did not yield under vehicular
traffic, and no on-site increase in fugi-
tive dust emissions was noted.
Based on these measured reductions
in silt loading and using the PM-10
emission equation, the following cal-
culated PM-10 reductions would re-
sult: street sweeping would reduce
PM-10 by approximately 14%; treat-
ment of the access area with
woodchips/mulch would moderately
reduce PM-10 by 27 to 33%; the
gravel buffer would result in the larg-
est reduction of PM-10, by 42 to 52%;
and the PM-10 control efficiencies are
somewhat lower than the silt loading
reductions, because of the 0.65 power
on silt loading in the PM-10 emission
equation.
-------
Michael M. Raile is with Midwest Research Institute, Kansas City, MO 64111.
Charles C. Masser is the EPA Project Officer (see below).
The complete report, entitled "Characterization of Mud/Dirt Carryout onto Paved
Roads from Construction and Demolition Activities," (Order No. PB96-129028;
Cost: $28.00, 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:
Air Pollution Prevention and Control Division
National Risk Management Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection Agency
National Risk Management Research Laboratory (G-72)
Cincinnati, OH 45268
Official Business
Penalty for Private Use $300
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
EPA/600/SR-95/171
------- |