United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-91-005b
September 1990
Air
x>EPA
FEASIBILITY OF INCLUDING
FUGITIVE PM-lO
EMISSIONS ESTIMATES IN THE
EPA EMISSIONS TRENDS REPORT
-------
EPA-450/4-91-005b
FEASIBILITY OF INCLUDING
FUGITIVE PM-10
EMISSIONS ESTIMATES IN THE
EPA EMISSIONS TRENDS REPORT
By
William Barnard And Patricia Carlson
E. H. Pechan & Associates, Inc.
Durham, NC 27707
EPA Contract No. 68-02-4400
EPA Project Officer: E. L. Martinez
Office Of Air Quality Planning And Standards
Office Of Air And Radiation
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
September 1990
-------
This report has been reviewed by the Office Of Air Quality Planning And Standards, U. S.
Environmental Protection Agency, and has been approved for publication as received from the
contractor. Approval does not signify that the contents necessarily reflect the views and policies of the
Agency, neither does mention of trade names or commercial products constitute endorsement or
recommendation for use.
FOREWORD
This report describes the results of Part 2 of a two part study. Part 2 evaluates the feasibility of
developing regional emission trends for PM-10. Part 1, presented in a separate report, evaluates the
feasibility of developing regional emission trends for VOC. These studies are part of the effort
underway to improve national emission trends.
EPA-450/4-91-005b
11
-------
TABLE OF CONTENTS
INTRODUCTION 1
FEASIBILITY ANALYSIS 4
UNPAVED ROADS 4
Introduction 4
Data Requirements 4
Summary 6
PAVED ROADS 6
Introduction 6
Data Requirements 7
Summary 9
WIND EROSION 9
Introduction 9
Data Requirements 11
Summary 11
AGRICULTURAL TILLING 11
Introduction 11
Data Requirements 12
Summary 12
CONSTRUCTION ACTIVITIES 13
Introduction 13
Data Requirements 13
Summary 14
FEEDLOTS 14
Introduction 14
Data Requirements 14
Summary 14
BURNING 15
Introduction 15
Data Requirements 15
Summary 15
LANDFILLS 15
Introduction 15
Data Requirements 15
Summary 16
MINING AND QUARRYING OPERATIONS 16
Introduction 16
Data Requirements 16
Summary 16
UNPAVED PARKING LOTS 17
in
-------
Introduction 17
Data Requirements 17
Summary 17
UNPAVED AIRSTRIPS 18
Introduction 18
Data Requirements 18
Summary 18
STORAGE PILES 18
Introduction 18
Data Requirements 19
Summary 19
PRELIMINARY EMISSIONS ESTIMATES 20
INTRODUCTION 20
UNPAVED ROADS 20
Methodology 20
Results 22
PAVED ROADS 25
Methodology 25
Results 26
WIND EROSION 31
RESULTS AND DISCUSSION 32
RESULTS 32
Feasibility Study 32
Preliminary Emissions Estimates 32
Problem Areas 34
RECOMMENDATIONS 35
REFERENCES 38
IV
-------
TABLES
1. TSP Emissions Inventories Estimates 3
2. Regional PM10 Emissions Estimates from Unpaved Roads 23
3. Regional PM10 Emissions Estimates from Paved Road
Resuspension . . 27
4. Regional PM10 Emissions Estimates from Paved Road Sanding and
Salting 29
5. 1985 NAPAP Wind Erosion Emissions Estimates 31
6. Feasibility of Developing Regional PM10 Emissions Estimates 33
-------
FIGURES
Page
1. Regional Contributions to Unpaved Road PM10 Emissions 24
2. Regional Contributions to Paved Road Resuspension PM10
Emissions 28
3. Regional Contributions to Paved Road Sanding/Salting PM10
Emissions 30
VI
-------
SECTION 1
INTRODUCTION
Each year the National Air Data Branch (NADB) produces a publication on
trends for the emissions of criteria air pollutants (particulate, sulfur dioxide, nitrogen
oxides, volatile organic compounds, carbon monoxide and lead). These data are
compiled into reports (Office of Air Quality Planning and Standards Data Files of
Nationwide Emissions) used internally by OAQPS for the evaluation of observed
trends in ambient air quality measurements. The emissions data are also needed for
reports by the EPA Administrator to Congress and for reports to the general public on
EPA progress in air quality management activities. The emission trends data are
developed using a large number of reference materials including data from EPA's
National Emissions Data System (NEDS), standard air pollutant emission factors from
the EPA publication Compilation of Air Pollutant Emission Factors. AP-42, Fourth
Edition, September 1985, as supplemented October 1986 and September 1988, and
data published by other Federal agencies and private sector statistical reporting
organizations.
Because of regional differences in emission sources, control programs, climatic
effects, etc., there is a need to investigate and possibly publish information on
pollutant trends on a regional basis. The purpose of this study was to 1) evaluate the
feasibility of developing regional emission trends for volatile organic compounds (VOC)
and particulate matter (PM) less than or equal to 10 microns (PM10) and 2) produce
preliminary estimates of these emissions for several representative source categories.
The feasibility analysis and development of regional emission trends for VOC is
presented in a separate report. This report addresses the feasibility of developing
regional PM10 emissions estimates. The focus of this study was on large potential
contributors to PM10 emissions which have been excluded from past analyses of PM.
Twelve categories of natural source PM10 emissions were examined. These
categories were determined based upon previous inventory estimates .(which have
been TSP inventories). The categories evaluated for the feasibility of developing
regional emissions estimates were:
Unpaved Roads
Paved Roads
Wind Erosion
Agricultural Tilling
Construction Activities
Feedlots
Burning
Landfills
Mining and Quarrying
-------
Unpaved Parking Lots
Unpaved Airstrips
Storage Piles
This report presents the results of the feasibility analysis and regional emissions
estimates for paved roads, unpaved roads and wind erosion.
In previous emissions inventories (which have typically represented TSP
inventories), the first eight categories have comprised more than 95% of all natural
area source particulate emissions. Table 1 below lists the estimated TSP emissions
from several of the above sources for two of these inventories. The inventories
represented here are ones determined by Evans and Cooper (1980) and by EPRI
(1988). The value in the Average column represents either the average emissions
from the two inventories or the single value for that source when only one inventory
value was available. Table 1 gives an indication of the magnitude of each of the
sources considered above, although it is possible that some of the sources not
inventoried in the past are of equal or larger magnitude than those reported here. All
values are in short tons.
Table 1 indicates that wind erosion and unpaved roads are the major
contributors to fugitive particulate emissions. In the inventories examined, these
sources have emissions that are an order of magnitude higher than any other source.
-------
Table 1. TSP Emissions Inventories Estimates
Source
Wind Erosion
Unpaved Roads
Agricultural Tilling
Construction
Mining and Quarrying
Paved Roads
Burning
Feedlots
Unpaved Airstrips
Landfills
Unpaved Parking Lots
Storage Piles
Evans and Cooper
229,655,000
272,050,000
39,012,000
21,547,000
9,588,800
8,763,000
1,939,000
EPRI (Heisler)
956,300,000
120,400,000
5,950,000
9,432,000
5,089,000
848,000
440,000
9,000
Average
592,977,500
196,225,000
22,481,000
15,489,500
9,588,800
6,926,000
1,393,500
440,000
9,000
-------
SECTION 2
FEASIBILITY ANALYSIS
In order for regional emissions estimates of PM10 to be feasible, two primary
requirements must be met. These requirements are:
1. A PM10 emission factor for the source must be available and,
2. Regional estimates of the source extent (activity factor) must
either be available or estimable from national data.
A recently published EPA document (Midwest Research Institute, 1988)
contains PM10 emission factors for several of the sources listed above, however, these
emission factors are typically based upon engineering judgement or a very limited
number of tests.
UNPAVED ROADS
Introduction
Unpaved roads are perhaps one of the easiest sources to develop regional
PM10 emissions estimates for, because of the availability of a PM10 emission factor and
source extent estimates. There is an established PM10 emission factor available in
AP-42. The background behind the development of that emission factor, other
unpaved road TSP emission factor development studies and the work carried out as
part of the 1985 NAPAP effort to develop a PM10 emission factor for this source
indicate that there is a large degree of uncertainty associated with an emission factor
for this source.
Total U.S. emissions of total paniculate (TSP) have been estimated previously
(Heisler, 1988; Evans and Cooper, 1980). These estimates indicated that 272.1 and
120.4 million tons of particulate were produced by this source for the years 1982 and
1976 respectively.
Data Requirements
Calculation of PM10 emissions from unpaved roads requires information
regarding an emission factor and the source extent. For unpaved roads, the source
extent is vehicle miles of travel (VMT) on unpaved roads. As indicated above, EPA
publication AP-42 contains an established PM10 emission factor for this source.
However, in order to utilize that emission factor for the calculation of regional
emissions estimates, regional information on the "correction parameters" that are
-------
included in the emission factor equation are necessary. The emission factor for PM10 .
emissions from unpaved roads is given as:
e = k(5.9) (s/12) (S/30) (W/3)°-7 (w/4)°-5 ((365-p)/365) (1)
where: e = emission factor (Ib/VMD
k = particle size multiplier (dimensionless; 0.36 for PM10)
s = silt content of road surface material (% - fraction of surface
material less than 75 microns in diameter)
S = mean vehicle speed (mph)
W = mean vehicle weight (tons)
, w = mean number of wheels
p = number of days with at least 0.01 inch precipitation per year
Equation 1 indicates that regional emission factors can be developed in order to
calculate regional emissions estimates from unpaved roads, if regional values for s, S,
W, w and p are available or can be estimated.
Reasonable estimates of the mean vehicle weight and the mean number of
wheels for vehicles traveling on unpaved roads can be determined using data
available from the U.S. Department of Transportation (DOT). This data is not
published, but is available upon request from DOT (Jeff Haugh, personal
communication). Reasonable estimates of vehicle speed based upon the functional
classification of the various unpaved roads can also be made. Although this is not
truly regional information, there is no reason to assume that vehicle weights, number
of wheels or vehicle speed on similar functional class roads should change
significantly from region to region.
Information on the silt content of unpaved roads and the number of days with at
least 0.01 inch of precipitation per year would be expected to vary from region to
region. Estimates of the number of days with a least 0.01 inches of precipitation can
be readily obtained from monthly summary information for meteorological stations
located within each state.
Silt content is much more difficult to obtain on a regional basis. AP-42 contains
some information on the silt content of various unpaved roads. The information
contained in AP-42 represents 103 samples. However, the majority of these road
samples (89 of 103) represent industrial unpaved roads, not normal highway traffic
unpaved roads. In many cases, industrial unpaved roads utilize process byproducts
(such as slag) for surfacing material rather than materials that would be utilized for
normal highway traffic unpaved roads. Since the surface material utilized may have a
significant impact on the silt content, the information in AP-42 may not be
representative of the silt content found in a particular region. Additionally, the
information given in AP-42 is not presented on a regional basis.
-------
There is, however, a significant database of normal highway traffic unpaved
road silt content information available. Researchers at the Illinois State Water Survey
(ISWS) have sampled and analyzed the silt content of over 200 unpaved road surface
material samples from over 30 states (Gary Stensland and Allen Williams, personal
communication) as part of the National Acid Precipitation Assessment Program
(NAPAP) alkaline particulate inventory effort. Thus, regional information on the silt
content of unpaved roads is available for developing regional unpaved road PM10
emissions estimates. This assumes that the average regional silt content does not
change over time. This assumption is probably adequate, unless large scale changes
occur in the type of surface material utilized by the states in a region.
The final data requirement necessary for calculating regional unpaved road
PM10 emissions estimates is regional source extent information. For unpaved roads,
this would be regional VMT on unpaved roads. DOT publishes information on the
mileage of unpaved roads for several road functional classes in the Annual Highway
Statistics, but does not provide VMT information for these roads. National VMT
information for unpaved roads is presented in the Highway Statistics report, and this
information reveals that the majority of unpaved road travel is on local functional class
unpaved roads. Unfortunately, local functional class unpaved road mileage by state is
not reported in the Highway Statistics, so developing an algorithm to assign regional
unpaved VMT based on unpaved road mileage from the information presented in that
document is not possible. Conversations with DOT personnel (Don Kestyn, personal
communication) indicated that most states do provide information on the number of
miles of local functional class unpaved roads broken down by Average Daily Traffic
Volume (ADTV). This information is compiled annually in a spreadsheet, and can be
obtained from DOT. This information can be utilized to calculate estimates of state-
level VMT on local functional class unpaved roads.
Summary
Regional estimates of unpaved road PM10 emissions are feasible with currently
available data. State (Allen Williams, ISWS, personal communication) and county
(Barnard, 1990) level PM10 emissions for the year 1985 have already been made as
part of the NAPAP research effort. Estimates of PM10 emissions from this source are
presented in the next section for a five year period.
PAVED ROADS
Introduction
Emissions of particulate matter from vehicle travel on paved roads arise from a
variety of processes. Among these processes are tailpipe emissions, tire wear, brake
wear, road sanding and salting and resuspension of dust tracked or deposited onto
-------
the road surface by vehicle traffic itself. The majority of reentrained traffic dust has
been found to consist primarily of mineral matter similar to sand or soil. Additionally, a
small component of the reentrained dust is derived from direct emissions from vehicles
(tailpipe, brake or tire wear).
Total U.S. emissions of total particulate (JSP) have been estimated previously
(Heisler, 1988; Evans and Cooper, 1980). The total resuspended dust estimates from
these inventories were 5.089 and 8.763 million tons for the years 1982 and 1976
respectively. These estimates included only the resuspended component of paved
road emissions and did not include tailpipe emissions, tire wear or brake wear or
emissions from road sanding and salting for snow removal. The estimates produced
by Evans and Cooper were also produced on a state-by-state basis and thus regional
emissions estimates for total particulate emissions from paved roads for the year 1976
can be derived.
In evaluating the feasibility of preparing PM10 emissions estimates for paved
roads, only two sources of particulate emissions were considered: resuspended dust
and sanding and salting of paved roads for snow removal.
Data Requirements
Resuspended Dust
In order to calculate PM10 emissions on a regional basis for paved roads,
certain data are necessary. At a minimum, an emission factor and the source extent
to which that emission factor applies are required. For paved roads, the source extent
is the VMT. EPA publication AP-42 contains an emission factor for urban paved roads
that can be applied to develop PM10 emissions estimates. However, in order to
determine regional emissions estimates, regional information on the parameters
necessary to calculate emissions are necessary. The emission factor for PM10
emissions for paved road dust resuspension is given as:
e = k (sL/0.7)p (2)
where: e = emission factor (Ib/VMT)
s = surface silt content (fraction of particles below 75 microns
diameter)
L = total road surface dust loading (grains/ft2)
k = base emission factor (Ib/VMT)
p = exponent (dimensionless)
For PM10 emissions, k = 0.0081 Ib/VMT and p = 0.8. The product of s and L yields
the silt loading. Thus, to develop regional emission factors for PM10, regional
information on s and L (or the product sL) must be available, since the other
-------
parameters are not region specific.
AP-42 Table 11.2.5-3 contains a summary of silt loadings (sL) for a variety of
roadway categories. However, the data base utilized for this table represents only 44
samples from 5 major cities. Thus, if this information was utilized to develop regional
emission factors, the spatial coverage would be limited. Additionally, this emission
factor is designed to cover emissions from paved urban roads, and there is no
emission factor listed in AP-42 for paved rural roads.
Information contained in a report by Cowherd et al. (1988) indicates that the
urban paved road emission factor can be utilized for estimating emissions from all
public paved roads (both urban and rural) and that the silt loading (sL) term can be
estimated by the following equation:
sL = 30.54/(V°41) (3)
where: sL = surface silt loading (grains/ft2)
V = average daily traffic volume (vehicles/day)
Thus, if region specific values can be obtained for V, then regional silt loading values
can be calculated.
In addition to regional emission factors, the second component necessary to
determine regional emissions estimates is development of regional estimates of the
source extent. For paved road emissions, this would be VMT.
Each year, DOT publishes Highway Statistics. Included among the information
in this publication are data on the number of vehicle miles of travel for each state
broken down by highway functional classification for both rural and urban VMT. Thus,
regional estimates of the source extent are available from DOT on an annual basis.
Also included in the data provided in the Highway Statistics publication are the
number of miles of road in each highway functional classification. As a consequence,
the average daily traffic volume in each state can be calculated. Thus, the silt loading
term in equation 3 necessary to calculate region specific emission factors (i.e. state-
level) can be calculated.
Road Sanding and Salting
There is a "gap filling" PM10 emission factor (MRI, 1988) available for sanding
and salting paved roads, but it is based on a very limited amount of data. The
recommended PM10 emission factor for sand application is:
e = 2,000 f (s/100) (4)
8
-------
where: e = emissions (Ib/ton of sand applied)
f = proportion of PM10 in the silt fraction of sand
s = silt content of the sand (percent)
The recommended PM10 emission factor for salt application is given as a single
value of 10 Ib/ton of salt applied.
In order to utilize the emission factor in equation 4, a regional database for f
and s would be required. Currently, no such database exists. However, default
values are given that cause the emission factor to evaluate to 0.018 Ib/ton of sand
applied. This value could be utilized in conjunction with an understanding of either the
amount of sand applied in each region or the number of miles of road treated in each
region along with the application rate per mile to determine the PM10 emissions.
Typically, the application rate (along with the salt/sand ratio) are available from state
highway departments (Dennis Carter, North Carolina DOT, personal communication).
Indeed, the "gap filling" document lists some typical sand/salt ratios, one of which is
very similar to that obtained for North Carolina. The "gap filling" document also lists
the number of miles of treated road for the majority of states, thus allowing state-level
computations of paved road sanding and salting PM10 emissions. .
Summary
PM10 regional emissions from paved roads can be calculated on an annual
basis with existing data. This would include emissions from road sanding and salting,
although emissions from this component of paved road emissions would require more
assumptions than paved road resuspension estimates.
WIND EROSION
Introduction
Although this source was characterized in the 1985 NAPAP Emission Inventory
and the emission factor and source extent were developed, the estimates developed
are not year-specific (they represent a statistical average using a 30 year wind
record), they are not easily reproduced and they represent emissions of particles < 20
microns. A variety of emission factors have been utilized over the years for estimating
wind erosion TSP emissions. Typically, the early ones (including the estimates
presented in Table 1) involved utilizing the Universal Soil Loss equation in some form
to convert horizontal erosional losses to vertical fluxes. More recent estimation
procedures have involved utilizing an credibility index coupled with some measure of
the threshold friction velocity to evaluate the vertical wind erosion emissions.
The current AP-42 industrial wind erosion emission factor utilizes this second
approach, but is not intended to be utilized to estimate large scale (i.e. regional)
-------
emissions. As a consequence, additional work to develop an emission factor that can
be readily applied to calculate regional level information on an annual basis would be
necessary. It is possible that the same (or similar) approach utilized for the 1985
NAPAP inventory could be used, but with some modifications in the utilization of the
wind data required. Additionally, the developer of the current NAPAP method has
indicated that the current method is computationally time consuming. Modification of
the current NAPAP emission estimation procedure would also require that it be
modified to predict PM10 emissions, since the 1985 NAPAP emissions estimates were
for particles < 20 microns.
The California Air Resources Board (GARB) has recently funded a survey in an
effort to identify the best available emission factors and emission estimation
methodologies for wind erosion from agricultural and desert lands (Dickson et al..
1988). For agricultural lands, the survey concluded that the estimates made using the
Universal Soil Loss equation were the best available given the existing data. The only
suggestions for improvement to this estimation technique were to adjust the climatic
factor used in this equation to account for irrigation practices and to utilize the mean
energy velocity rather than the mean wind velocity in calculating the climatic factor.
Emissions of PM from desert land were determined using the following equation:
F = 1.78x10-16U2782 (5)
where: F = aerosol flux, (g/cm2-s)
U = wind speed at 10 meters (cm/sec)
The study noted that several assumptions were required for utilizing this equation to
estimate PM emissions from desert lands. All improvements to the estimation of wind
erosion estimates from desert lands suggested by the study would require a high level
of effort. It was also noted that activity data (acreage) for disturbed desert land (which
have much higher emission rates than undisturbed desert land) have thus far not been
defined or quantified.
The U.S. Department of Agriculture is in the process of developing a wind
erosion prediction system. This system is intended to be a PC based system
involving a modular structure. The modules will deal with weather, crops, tillage, soil
properties, water and energy balances and wind erosion mechanics (Hagen, 1989).
This computer system is intended to replace the Universal Soil Loss equation.
Several problems exist with trying to incorporate this system into developing regional
emission estimates. First, the system is not scheduled for completion until 1991 and a
user-friendly version is not due until 1993. Second, the system is set up to predict
field scale wind erosion, not regional scale wind erosion. Third, the system is
designed for agricultural applications, and current plans do not include consideration of
wind erosion from non-agricultural land. Finally, although the system does predict a
suspension component, this component generally represents particles < 100 microns.
10
-------
Thus, using this system to predict PM10 emissions will involve making assumptions as
to the fraction of the suspension component that is < 10 microns.
Data Requirements
In order to calculate regional emissions, a regional emission factor is needed.
Although the Universal Soil Loss equation has been and is still being utilized to predict
wind erosion emissions, all researchers involved in determining wind erosion
emissions concede that it is an inadequate method to predict suspended particulate
emissions, especially PM10. Although other potential approaches have been
developed and utilized, these methods typically do not lend themselves to relatively
straight forward application to regional emissions estimates. It may be possible,
however, to develop a reasonable method for predicting wind erosion on a regional
level.
Adequate source extent information is available from a variety of sources,
although some additional research is needed to evaluate the best sources of
information and to ascertain how frequently the source extent information is updated.
The source extent utilized in the development of the 1985 NAPAP inventory was the
National Resource Inventory. This soils database is updated on approximately a five
year basis. As a consequence, development of annual regional emissions estimates
to evaluate trends could be hampered even if an adequate emission factor is
developed if the source extent estimates are invariant over long periods of time.
Summary
If EPA is willing to use the Universal Soil Loss equation and if the source extent
information utilized is updated on an annual basis, then estimation of regional PM10
emissions would be feasible. However, if those conditions are not adequate, then
estimation of emissions from this source would not be feasible at this time. Regional
estimates of PM10 emissions from wind erosion for 1985 developed from the 1985
NAPAP inventory are presented in the next section.
AGRICULTURAL TILLING
Introduction .
Agricultural tilling emissions were estimated for the 1985 NAPAP emission
inventory by using a previously developed TSP emission inventory (Evans and
Cooper, 1980) coupled with information from AP-42 on the fraction of particles < 10
microns. In addition, a plume depletion factor of 0.1 was utilized to represent the
potential for particles traveling long distances from their source (and thus having the
potential to neutralize acids in precipitation since this was the emphasis of the NAPAP
research program). Although this information was reported as part of the 1985
. 11
-------
inventory, the actual data represents 1976.
The emission factor in AP-42 for this source category (upon which the 1985
NAPAP emission inventory was based) was developed based on a very limited data
set and the development work was primarily carried out in the mid-70's to early 80's.
The emission factor available in AP-42 for agricultural tillage is:
e = k (4.80) (s)°-6 (6)
where: e = emission (Ib/acre)
s = silt content (%)
k = particle size multiplier (= 0.21 for PM10)
Data Requirements
A regional PM10 emission factor is necessary to calculate regional emissions
estimates for this source. From equation 6, it is clear that in order to develop regional
emission factors, data on the soil silt content for the region must be available.
Information on this parameter may be available in certain soil databases such as the
National Resources Inventory or it may be possible to estimate this parameter using
soil type maps and generalized soil characteristics information. Although this
information may not be updated on an annual basis (i.e. the time frame over which
trends estimates would be developed), that may be less important for evaluating the.
silt content than for estimating the source extent (as was indicated in the Wind Erosion
section above). It is possible that the silt content does not vary significantly from year
to year.
The source extent information is much more likely to be available on a yearly
basis for agricultural tilling, since tilling activity is crop related and the annual acreage
of the various types of crops is estimated on an annual basis.
Summary
Since an existing emission factor is available and since adequate source extent
information could be obtained, regional PM10 emissions estimates are feasible for this
source. One potential problem for estimation of emissions for this source, however, is
that no climatic parameters are considered in the emission factor. For instance, if the
seasons just prior to tilling are very dry, it seems likely that emissions for tillage
operations would be greater than those carried out for a year with greater precipitation
in the preceding seasons. The current emission factor does not account for this type
scenario.
12
-------
CONSTRUCTION ACTIVITIES
Introduction
Construction activities can be divided into two broad categories, building
construction and road construction. A general estimate for the emission factor for
building construction was made in the early 70's. That value (1.2 tons/acre of
construction/month), is based on field measurements of TSP emissions from
apartment and shopping center construction projects. Some additional work has been
carried out more recently, that looked at PM10 emission factors for topsoil removal,
earthmoving and aggregate hauling (MRI, 1988). Based on that information, the
following emission factors for PM10 were derived:
For topsoil removal 20 IbA/MT
For earthmoving (cut and fill) 4.3 IbA/MT
For truck haulage 10 IbA/MT
These emission factors are based on TSP emission factors determined using
dispersion modeling multiplied by the PM10/TSP ratio to obtain the PM10 emission
factor. No more than four tests were carried out to develop these emission factors,
however.
Data Requirements
Again, a PM10 specific emission factor for the region is necessary to calculate
regional PM10 emissions. As noted above, there is a general TSP emission factor for
total construction activity and PM10 specific emission factors for three processes
involved in construction activity. Thus, a PM10 emission factor is available.
However, in order to calculate regional PM10 emissions estimates from this
source, the source extent must also be available on a regional basis. With the new
PM10 emission factors developed for this source, the required information is the
number of vehicle miles traveled during construction activities. This data is simply not
available and would have to be estimated. With the old TSP emission factor, only the
number of acres and months that construction occurred were necessary to calculate
PM emissions estimates. Previous large scale TSP inventories (Evans and Cooper,
1980; Heisler, 1988) utilized a conversion factor to evaluate the number of acres
under construction from annual data on construction receipts. Smaller scale
inventories (Cowherd and Guenther, 1976) have utilized the actual acreage of
construction for estimating emissions. In addition, the number of months that each
type of construction activity is performed is required. Estimation procedures for this
have already been developed (Cowherd and Guenther, 1976). Thus, the source
extent estimation methodology is already in place for TSP emissions estimation, but
would have to be completely developed for the PM10 emission factors.
13
-------
Summary
Unless a method for estimating the source extent for use with the PM10
emission factors can be found, development of regional PM10 emissions estimates
cannot be made at this point.
FEEDLOTS
Introduction
Very little work has been done to develop an emission factor for feedlot
particulate emissions. Again, the work that has been done was completed in the late
70' s and early 80's. The TSP emission factor that was developed at that time
depended upon either the feedlot capacity or throughput. PM10 emission factors have
been developed from the TSP emission factor by using the PM10/TSP ratio for
agricultural tilling (MRI, 1988).
Data Requirements
Although the PM10 emission factor for this source is very crude (due to the way
it was developed), one does exist. Current PM10 emission factors are:
e = 180 lb/day/1,000 head capacity, or
e = 17 tons/1,000 head throughput
Source extent information could be extracted from the Census of Agriculture
developed by the Bureau of the Census. Additional information may be available from
the various state agricultural agencies. If information on the source extent was
developed from the Census of Agriculture, it is possible that that source of information
is not updated on an annual basis, thus estimates for trends purposes would not truly
reflect the trend in emissions.
Summary
Both a PM10 emission factor and source extent information are available, so
regional emissions estimates could be calculated. However, several potential
problems exist with calculating PM10 regional emissions estimates based on the
existing information. First, the emission factor contains no climate correction factor,
even though the "gap filling" document where it was derived states that emissions are
related to climate, soil texture, season, cattle density, natural mitigation of cattle in
holding pens and pen cleaning cycles. None of these parameters are incorporated
into the emission factor. Additionally, it is not clear whether or not the source extent is
available on an annual basis as would be required of trends calculations.
14
-------
BURNING
Introduction
This category includes burning from both forest wildfires and prescribed
burning. There is an available PM10 emission factor from AP-42. Included in these
emission factors are regional emission factors for prescribed burning.
Data Requirements
Regional PM10 emission factors are available in AP-42 for prescribed burning
and many of these could be applied to wildfire burning. AP-42 does indicate,
however, that these regional emission factors should not be utilized for emission
inventories or for planning purposes.
Source extent estimates for wildfires for a given year could be obtained from
the U.S. Forest Service regional offices. Prescribed burning source extent and spatial
distributions should be fairly easy to obtain, since these are set fires and the locations,
number of acres, fuel type, etc. should be known. The current trends method uses
data from 1978 and assumes that it has remained constant each year since then.
Summary
Since the emission factor is available and since the source extent should be
relatively easy to obtain, regional PM10 emissions estimates are feasible.
LANDFILLS
Introduction
Particulate emissions from landfill operations are caused by traffic, materials
handling, and covering wastes with soil. There is not a single value PM10 emission
factor available for landfill operations, however, the processes performed in landfill
operations are covered by emission factors in AP-42. Some recent work on two
landfills in Chicago indicated that the major contributor to landfill PM10 emissions was
due to traffic on unpaved areas. A suggested PM10 emission factor of 1 Ib/yd3/mi of
travel to the disposal site was developed for that study (MRI, 1988). Additional
emission factor development work would be required to validate this for landfills in
other areas of the country.
Data Requirements
While there is a suggested PM10 emission factor, it requires a knowledge of the
source extent consisting of both the number of cubic yards of waste disposed of and
15
-------
the distance from the entrance to the disposal site. While information on the number
of cubic yards of material disposed of in landfills can probably be obtained EPA's
Office of Solid Waste (OSW), it is doubtful that the average distance from the entrance
to the disposal area is available.
Summary
While a "gap filling" PM10 emission factor exists, information on the two
parameters of the source extent required to use this emission factor is probably
available for only one of these parameters. Assumptions would have to be made for
the other parameter in order for regional PM10 emissions estimates to be made. Thus,
regional emissions estimates are feasible but would require some degree of effort in
order to develop estimation procedures.
MINING AND QUARRYING OPERATIONS
Introduction
A wide variety of activities contribute to PM10 emissions from this source.
Included among these are vehicle loading and unloading, blasting, crushing, drilling,
overburden removal and unpaved road travel. PM10 emission factors have been
developed for the majority of these activities.
Data Requirements
A mentioned above, PM10 emission factors exist for the majority of the activities
that comprise the source of emissions at mining and quarrying operations. However,
many of these emission factors require correction parameters such as silt or moisture
content. Thus, in addition to estimates of the source extent, regional estimates of
these correction parameters would also be required.
An additional complication to calculating regional PM10 emissions estimates
from this source is that not all of the emission factors require the same source extent
information. For instance, the PM10 emission factors for crushed stone processing all
require information on the number of tons of stone processed. However, for western
surface coal mining operations, some emission factors require information on the
number of tons of coal, the number of blasts (for blasting), the number of cubic meters
of overburden removed, and the vehicle miles of travel. As a consequence, several
different regional source extent estimates must be derived in order to utilize these
emission factors to develop regional PM10 emissions estimates.
Summary
Although PM10 emission factors are available for this source for the variety of
16
-------
emission generating activities, several of these emission factors require knowledge of
correction parameters that would need to be known on a regional basis. In addition,
several of the emission factors needed to calculate emissions from this source require
different source extent estimates. Several of the source extent estimates required
would be very difficult to develop on a regional basis (i.e. the number of vehicle miles
of travel at western surface coal mines). As a consequence, although it may be
feasible to produce regional PM10 emissions estimates, it would require a large
development effort to produce the source extent estimates necessary to calculate
these emissions.
UNPAVED PARKING LOTS
Introduction
Since unpaved parking lots are similar to unpaved roads, some of the initial
work to develop an emission factor for PM10 emissions for this source has already
been done. Additional work has been carried out recently to evaluate emissions from
this source, and has come up with a method for estimating emissions (MRI, 1988).
However, no field validation of the method was carried out.
Data Requirements
As indicated above, a PM10 emission factor applicable to this source has
already been developed. Correction parameter information could be developed
relatively easily, even the silt content term since the surface material utilized for
unpaved parking lots would probably have the same source as that for unpaved roads
in the same region.
Source extent evaluation for this source would be extremely difficult. The
authors know of no source for determining the number and size of unpaved parking
lots. Some type of computational algorithm would need to be developed, in order to
estimate the source extent, if regional PM10 emissions estimates were to be
developed.
Summary
Although a PM10 emission factor exists for this source, development of regional
PM10 emissions estimates is probably not possible at this time, given the difficulty of
developing adequate source extent information.
17
-------
UNPAVED AIRSTRIPS
Introduction
Unpaved airstrips are similar to unpaved roads and the PM10 emission factor
applicable to unpaved roads has been suggested for use with this source (MRI, 1988).
Addition of a wind erosion multiplier (equal to 2) to account for dust produced by the
prop wake has been suggested. Since the unpaved road emission factor is the basis
for emissions from this source, the same correction parameter requirements are
applicable. Source extent for this source is the number of landing and takeoff (LTD)
cycles performed per airport.
Data Requirements
A regional PM10 emission factor is available for this source. Adequate
information on the correction parameters required for utilizing this emission factor can
also probably be developed on a regional basis, since the majority of the information is
the same as that required for unpaved roads.
In previous TSP emissions inventories that have been developed for this source
(Cowherd and Guenther, 1976), a computer tape has been obtained from the Federal
Aviation Administration (FAA) which was utilized to determine the number of airports
utilizing dirt, turf or gravel runways. Personal communications with regional FAA
personnel would be required to estimate the number of LTOs for these airports. By
determining the number and the LTOs at these airports, the source extent can be
determined.
Summary
Regional emissions estimates from this source are feasible. However,
considering the amount of time that would be required to develop the source extent
term, especially in light of the magnitude of emissions from this source, this source is
probably not worth including in a regional PM10 emissions estimate. Indeed, EPA's
own "gap filling" PM10 emission factor document (MRI, 1988) indicates that this source
is a minor source of PM10.
STORAGE PILES
Introduction
AP-42 has a PM10 specific emission factor for this source category. Since this
emission factor uses the silt content of the stored material, additional work may be
needed to improve the database on silt contents of various materials on a regional
basis.
18
-------
Data Requirements
As indicated above, a PM10 emission factor is available, although information on
the correction parameters utilized by this emission factor to calculate emissions may
not be available on a regional basis.
Source extent information on this source category is virtually non-existent and
would have to be developed from scratch.
Summary
Development of regional PM10 emissions estimates for this source is probably
not feasible at this time since an adequate source extent database does not exist.
19
-------
SECTION-3
PRELIMINARY EMISSIONS ESTIMATES
INTRODUCTION
In order to demonstrate the feasibility of developing regional PM10 emissions
estimates, preliminary estimates were made for two fugitive dust sources. Regional
PM10 emissions estimates for unpaved and paved roads for the period 1984 to 1988
were made. For paved roads, two separate components of paved road emissions
were considered. These two components were resuspension of dust from paved road
surfaces and emissions from sanding and salting of paved roads during snow/ice
conditions. In addition to developing emissions estimates for these two fugitive
sources, estimates of PM10 emissions from wind erosion (developed as part of the
1985 NAPAP effort) are also presented. Emission estimation techniques, data
sources, assumptions and problem areas encountered in the development of these
estimates are also discussed.
UNPAVED ROADS
Methodology
Estimation of PM10 emissions from unpaved roads was carried out similarly to
the methods outlined by Barnard (1989) and Barnard et al. (1987). This method
requires estimation of unpaved road VMT from unpublished data available from the
U.S. DOT. Data for state-level local functional class unpaved roads (which typically
comprise 80-90% of all unpaved road mileage) is assembled each year in
spreadsheets by DOT. The data is compiled by locale (rural, small urban and urban),
road surface type (paved, gravel/soil surfaced and unimproved) and by ADTV class.
By determining the midpoint of the average daily traffic volume class and multiplying
that number by the number of miles of unpaved road in that class and by 365
days/year, the annual VMT for each ADTV class can be determined. Once the local
functional class unpaved VMT was determined, then the average urban and rural
ADTV on these roads was calculated. These values were used to determine the VMT
on non-local functional class unpaved roads. Mileage for non-local functional class
roads is available from the U.S. DOT Annual Highway Statistics publication.
Information was obtained for the years 1984-1988.
For some states, local functional class unpaved road mileage information was
not available. For these states, the estimation method, developed by Barnard et al.
(1987) was utilized. This method basically utilizes information available for
surrounding states to develop the appropriate information for the state with missing
data.
20
-------
Once the source extent was determined, then appropriate emission factors
needed to be determined. As can be seen from equation 1, the silt content, vehicle
speed, vehicle weight, number of wheels and number of dry days per year are needed
to determine the emission factor. In most cases, state-level information was utilized
for these parameters. Silt content was derived from a database developed during the
1985 NAPAP effort by the Illinois State Water Survey. This database contains the silt
content of over 200 unpaved roads from over thirty states. Average silt content of
unpaved roads in a state were calculated for each state that had three or more
samples for that state. For states that did not have the required number of samples,
the average for all samples from all states was substituted.
The number of dry days per year in each state was determined by averaging (if
two or more stations were available) the number of days with more than 0.01 inches
of precipitation from the most rural meteorological stations available from Local
Climatological Data (LCD) annual summaries. This information was obtained for the
years 1984-1988.
For vehicle speed, the following assumptions were made:
Rural Roads Speed (mph)
Minor arterial 45
Major collector 40
Minor collector 40
Local 35
Urban Roads Speed (mph)
Other principal arterial 50
Minor arterial 45
Collector 40
Local 40
Estimates of vehicle weight and the number of wheels per vehicle were made
using information provided by the U.S. DOT (Jeff Haugh, personal communication).
This data indicated that the following weighted average values were appropriate for
the following vehicle classes:
Vehicle Type Weight (tons) Wheels
Single Trailer Trucks 26.7 18
Multi-Trailer Trucks 31.5 20
Single Unit Trucks 9.55 7
Passenger vehicles 2.5 4
By utilizing the above information, emission factors for each vehicle type and for
21
-------
each road type could be determined. Once this was accomplished, national statistics
provided by U.S. DOT on travel activity by vehicle type (Jeff Haugh, personal
communication) were utilized to allocate the percentage of travel on each road type to
each vehicle type. By doing this, a composite weighted emission factor that reflected
the vehicle mix on each road type was developed. This emission factor was then
multiplied by the source extent (i.e. VMT) on that road type to develop state-level
emissions estimates. The state-level emissions were then compiled into EPA regional
emissions.
Results
The results of the emissions estimates developed using the methodology
described above are presented in Table 2 (all values in short tons). These values
clearly show that unpaved road PM10 emissions are highest in Regions 5 and 6 for the
years examined. Overall U.S. annual emissions vary between 17.9 and 19.6 million
tons for the period 1984-1988. Figure 1 depicts the regional contributions to the
overall U.S. emissions. There is no consistent overall trend for the years examined.
This result is probably related to the differences in the state-level reporting of unpaved
road mileage and from year-to-year differences in the meteorological correction term.
It is important to remember that the states are not required to report local functional
class unpaved road information (the most important class of unpaved roads) to the
U.S. DOT and that the information reported is an estimate and not a true count of
traffic patterns or number of miles on these roads.
As a point of comparison, unpaved road particulate emissions estimates were
determined as part of the 1985 NAPAP Emissions Inventory. The PM10 emissions
reported as part of that effort were 8.9 million tons for the total U.S. The major
differences between the values reported as part of the NAPAP effort and the
estimates reported here are caused by two factors. First, investigators at the Illinois
State Water Survey responsible for preparing the state-level emissions estimates for
the NAPAP inventory utilized a plume depletion factor to reduce the calculated
emissions to 0.1 of their original value. This plume depletion factor was utilized to
account for emissions capable of traveling long distances and interacting with
precipitation (the major emphasis of the NARAP work). Thus, the 1985 NAPAP
estimates should be multiplied by 10 to yield numbers equivalent to those reported
here. If this multiplication is carried out, the total U.S. PM10 emissions estimated for
the 1985 NAPAP inventory would be approximately 89 million tons. This value is
approximately 4 times the values estimated in this study.
The second difference between the NAPAP estimates and the estimates
presented in this study is that the NAPAP inventory did not use the AP-42 emission
factor for gravel/soil surfaced unpaved roads. The ISWS researchers utilized a new
emission factor developed for passenger vehicles traveling on gravel roads, rathe/
than the AP-42 emission factor. This emission factor uses correction parameters'for
22
-------
vehicle speed, road silt content and the fraction of road surface material between
0.425 and 2 mm. No measurements involving vehicles other than passenger cars
were used to develop this emission factor, and its reliability in predicting emissions
from vehicles other than passenger cars has not been verified.
For unimproved surface unpaved roads, the ISWS investigators utilized the AP-
42 emission factor equation, however, they used an older version of the equation that
had the particle size multiplier (k, see equation 1 above) equal to 0.45 instead of 0.36.
Thus, emissions for this road type were over-estimated by at least a factor of 1.25.
Because of these differences (and to be consistent with other TSP emission
inventories), we decided to utilize the AP-42 emission factor for all unpaved road
emissions estimates. Other differences between the NAPAP estimation technique and
that utilized here are the use of only one vehicle speed and a single vehicle weight to
calculate emissions for the NAPAP inventory. Our technique (as described above)
utilizes information on the vehicle distribution on unpaved roads (and the consequent
weight and number of tire differences) to develop a weighted emission factor that
reflects the traffic distribution. Estimates of vehicle speed on the various unpaved
road functional classifications were also utilized in our estimation method. '
Table 2. Regional PM10 Emissions Estimates from Unpaved Roads
REGION
U.S.
EPA1
EPA 2
EPA 3
EPA 4
EPA 5
EPA 6
EPA 7
EPA 8
EPA 9
EPA 10
1984
18,047,498
529,089
852,973
405,987
2,048,798
2,520,738
4,747,981
2,139,928
1,631,882
1,588,718
1,581,403
1985
19,009,713
607,772
819,879
436,440
2,362,001
3,247,627
4,873,525
2,131,620
1,611,792
1,605,203
1,313,854
1986
18,723,489
609,177
802,135
454,937
2,477,899
3,000,998
4,798,563
2,190,986
1,564,683
1,600,053
1,224,059
1987
17,919,076
397,133
560,562
457,229
2,595,136
3,112,027
4,900,560
2,199,667
1,629,575
819,741
1,247,447
1988
19,698,434
409,081
302,572
455,106
2,848,614
2,789,425
6,212,560
2,778,000
1,808,317
1,042,792
1,051,968
23
-------
Figure 1. Regional Contributions to Unpaved Road PM10 Emissions
I I EPA REGION 10
EPA REGION 9
EPA REGION 8
EPA REGION 7
EPA REGION 6
EPA REGION 5
EPA REGION 4
EPA REGION 3
EPA REGION 2
EPA REGION 1
v. 5^ **%*.< "• f JT£?v* * v*y •& ysSSs;?
^rs^At^I^-
^af^™f*7ftf ^/ ^.^
^'' •• 'S-- *' *% J^ r«'
24
-------
PAVED ROADS
Methodology
Resuspended Dust
Regional PM10 emissions estimates were made for two components of paved
road emissions. The first component was paved road resuspension. The second
component was an estimate of the emissions from paved roads as the consequence
of road sanding and salting operations.
Emissions estimates from paved road resuspension were derived by utilizing
the current AP-42 emission factor (see equation 2 above) in conjunction with U.S.
DOT data on VMT on paved roads. AP-42 provides values for the base emission
factor and coefficient for PM10 emissions in Table 11.2.5-1. Section 11.2.5 of AP-42
also provides information on the silt loading (sL) of paved roads, but the amount of
data available is very limited. As a consequence, equation 3 (above) was utilized to
determine the silt loading. In order to determine sL, the ADTV on various functional
classes of paved roads was required. Since the U.S. DOT also provides road mileage
statistics by functional class, ADTV values for the various paved road functional
classes could be calculated by dividing the VMT by the mileage and then dividing by
the number of days per year. These ADTV values were utilized in conjunction with
equations 2 and 3 to determine functional class specific paved road emission factors.
Once these emission factors were derived, then the PM10 emissions could be
calculated for each paved road functional class.
Road Sanding and Salting
Regional PM10 emissions estimates from road sanding and salting operations
were estimated by utilizing the emission factor given in the EPA "gap filling" document
(MRI, 1988). For road sanding, an emission factor of 0.018 Ib/ton was utilized. For
salting operations, an emission factor of 10 Ib/ton was used. In order to utilize these
emission factors, the number of tons of sand and salt typically applied during sanding
and salting operations was needed. Information in the "gap filling" document for Iowa
indicated that 510 Ib/mi of salt and 1000 Ib/mi of sand are applied per snow day.
Information from the North Carolina Department of Transportation (Dennis Carter,
personal communication) indicated that between 200-500 Ibs/two-lane mile are applied
to roads in North Carolina, depending upon the temperature and the type of
precipitation (wet snow, dry snow, sleet, freezing rain, etc.). The information received
also indicated that the sand/salt ratio used in North Carolina was 20:1. This value is
on the high end of the range indicated in the "gap filling" document. For calculating
emissions, values of 500 Ib/mi for salt and 5000 Ib/mi for sand were assumed.
Source extent information was derived using Table 13 of the "gap filling"
25
-------
document. That table lists the number of single lane miles of road treated for various
states. Arizona, Arkansas, Tennessee, South Carolina, Louisiana, Oklahoma, Texas
and Nevada did not have data. For these states, the number of miles of treated road
was approximated by determining the average for several surrounding states. The
states utilized to derive the estimated mileage and the estimated treated miles of road
(in thousands of miles) were as follows:
State Surrounding States Treated Mileage
NM, CA, UT, CO 7.1
NM, KS, MS 12.2
MS, AL, GA, NC, VA, KY, MO 11.9
GA, NC 6.1
MS, AL, GA 2.6
MO, KS, NM 21.9
NM, MS 5.3
CA, OR, ID, UT 11.8
Table 13 of the "gap filling" document also presents data on the mean annual
snow days for each state. However, since this study was designed to analyze the
development of emission trends, we substituted mean annual snow days with year-
specific information on the number of days with > 1 inch of snow for the most urban
meteorological stations that report annual Local Climatological Data information to the
National Climatic Data Center.
Once the above information was obtained, the regional PM10 emissions
estimates for road sanding and salting operation were calculated by multiplying the
sand application rate by the emission factor times the number of miles of treated road
times the number of days with > 1 inch of snow. Salting emissions were calculated in
an identical manner using the salt emission factor and the salt application rate.
Results
The results of these calculations are given in Tables 3 and 4 (all values in short
tons). These regional PM10 emissions estimates show that paved road PM10
emissions attributable to resuspension are between 9.2 and 10.5 million tons.
Additionally, with few exceptions, paved road resuspension PM10 emissions have
tended to increase in each region each year of the five year period examined.
One interesting item to note in Table 3 is that the total U.S. PM10 emissions
from paved road resuspension is approximately equal to the paved road TSP
emissions estimates presented in Table 1. This is the result of a change in the
emission factor between the time that the TSP inventories presented in Table 1 were
calculated and the current emission factor for PM10. The old TSP emission factor was
26
-------
approximately the same as the current PM10 emission factor.
Regional contributions to paved road resuspension PM10 emissions estimates
are shown in Figure 2.
Table 3. Regional PM10 Emissions Estimates from Paved Road Resuspension
REGION
U.S.
EPA1
EPA 2
EPA 3
EPA 4
EPA 5
EPA 6
EPA 7
EPA 8
EPA 9
EPA 10
1984
9,292,335
431,720
661,243
906,597
1,877,832
1,795,007
1,329,015
567,044
381,767
982,266
359,844
1985
9,490,310
"444,629
672,103
928,437
1,918,011
1,819,558
1,347,365
568,828
392,999
1,033,563
364,818
1986
9,724,860
457,005
694,365
958,512
1,952,973
1,888,822
1,355,751
582,428
396,912
1,063,079
375,011
1987
10,019,982
482,957
708,478
971,557
2,043,687
1,946,045
1,396,151
599,398
403,223
1,076,497
391,989
1988
10,455,975
479,923
729,362
996,902
2,183,031
2,015,364
1,457,858
616,722
419,960
1,138,699
418,155
27
-------
Figure 2. Regional Contributions to Paved Road Resuspension PM10 Emissions
I I EPA REGION 10
EPA REGION 9
EPA REGION 8
EPA REGION 7
EPA REGION 6
EPA REGION 5
EPA REGION 4
EPA REGION 3
EPA REGION 2
EPA REGION 1
28
-------
Regional PM10 emissions estimates from paved road sanding and salting
operations are presented in Table 4. The values shown in that table clearly indicate
(at least for the assumptions utilized in these calculations) that paved road sanding
and salting operations are a relatively insignificant source of PM10 emissions compared
to the resuspension component of paved road PM10 emissions.
Information in Table 4 also clearly indicates the importance of meteorology in
estimating these emissions. Since the number of treated miles and the sand and salt
application rates were invariant over the period considered, the only variable left to
influence emissions was the number of days with > 1 inch of snow.
Regional contributions to paved road sanding and salting PM10 emissions are
shown in Figure 3.
Table 4. Regional PM10 Emissions Estimates from Paved Road Sanding and Salting
REGION
U.S.
EPA1
EPA 2
EPA 3
EPA 4
EPA 5
EPA 6
EPA 7
EPA 8
EPA 9
EPA 10
1984
14,984
843
1,075
913
214
5,597
172
1,699
3,939
238
296
1985
19,683
811
1,493
1,306
322
9,007
215
1,812
3,803
327
587
1986
11,034
717
988
801
98
4,682
138
913
2,267
150
279
1987
10,852
915
1,054
1,167
299
3,951
202
897
1,989
211
167
1988
13,300
588
943
809
129
6,698
248
642
2,787
186
269
29
-------
Figure 3. Regional Contributions to Paved Road Sanding/Salting PM10 Emissions
I I EPA REGION 10
EPA REGION 9
EPA REGION 8
EPA REGION 7
EPA REGION 6
EPA REGION 5
EPA REGION 4
EPA REGION 3
EPA REGION 2
EPA REGION 1
30
-------
WIND EROSION
Regional PM10 wind erosion emissions estimates were not calculated in this
study, due to the complications in calculating emissions estimates from this source
indicated in the feasibility section above. However, to facilitate comparison of the
magnitude of this source to the estimates prepared for unpaved roads and paved
roads, regional emissions estimates developed by Barnard (1990) as part of the 1985
NAPAP emission inventory effort are presented in Table 5. It should be realized, that
these emissions estimates are for particles < 20 microns. Particle size distributions
determined using airplane sampling during wind erosion events indicates that
approximately 90% of the particle mass was in particles smaller than 10 microns
(Gillette et al.. 1978), thus multiplication of the values in Table 5 by 0.9 would give an
indication of the magnitude of PM10 emissions from this source.
Table 5. 1985 NAPAP Wind Erosion Emissions Estimates (all values in short tons)
REGION
U.S.
EPA1
EPA 2
EPA 3
EPA 4
EPA 5
EPA 6
EPA 7
EPA 8
EPA 9
EPA 10
Particulate
4,711,539
0
909
1,040
5,462
366,011
2,340,545
395,476
1,516,533
83,398
2,164
The values in Table 5 indicate that PM10 emissions from wind erosion for the
U.S. (as calculated as part of the 1985 NAPAP inventory) are approximately the same
order of magnitude as emissions from paved road resuspension.
31
-------
SECTION 4
RESULTS AND DISCUSSION
RESULTS
Feasibility Study
Table 6 summarizes qualitatively the feasibility of developing and the
development effort required for generating regional PM10 emissions estimates for the
sources considered in the feasibility section above.
Table 6 indicates that regional emissions estimates for the majority of the major
PM10 fugitive dust sources can be developed. Of the largest sources, wind erosion will
require the largest effort to develop regional emissions estimates, at least initially,
since a determination of an appropriate emission factor and development of source
extent information will be required. Several sources (unpaved roads, paved roads,
and agricultural tilling) require a low level of effort. The remaining categories for which
regional emissions estimates are feasible require either a moderate or moderate to
high level of effort. Two categories are considered infeasible at this time (unpaved
parking lots and storage piles). Unpaved airstrips (although feasible) are probably not
worth the effort that would be necessary to develop emissions estimates, given the
estimated magnitude of the source.
Preliminary Emissions Estimates
The preliminary emissions estimates determined for unpaved and paved roads
in this study indicate that it is feasible to produce regional PM10 emissions estimates
from fugitive dust sources. Based on these estimates, unpaved road PM10 emissions
are approximately double those of paved roads. Additionally, PM10 emissions from
sanding and salting operations on paved roads are significantly lower than emissions
from resuspension of material on paved roads. Since the emission factors for tire and
brake wear are much lower than those for resuspension (by approximately a factor of
350), and since the source extent is the same (VMT), these sources also would be
much lower than the resuspension source.
32
-------
Table 6. Feasibility of Developing Regional PM10 Emissions Estimates
Source
Wind Erosion
Unpaved Roads
Agricultural Tilling
Construction
Mining and
Quarrying
Paved Roads
Burning
Feedlots
Unpaved Airstrips
Landfills
Unpaved Parking
Lots
Storage Piles
Regional Emissions
Estimates Feasible?
Yes
Yes
Yes
Maybe (Depends upon
development of
adequate source extent
information or use of
older TSP emission
factor with PM10
correction parameter)
Maybe (Depends upon
development of
adequate source extent
information)
Yes
Yes
Yes
Yes
Yes
No (not at this time)
No (not at this time)
Development Effort Required to
Produce Regional Emissions
Estimates
Moderate to High
Low (preliminary estimates
made)
already
Low
Moderate to High
Moderate to High
Low (preliminary estimates
made)
already
Moderate
Moderate
Moderate (but may not be worth
effort considering size of source)
Moderate to High
High
High
33
-------
Problem Areas
Several potential problems exist with producing regional estimates for fugitive
PM10 emissions. Although emission factors and source extent estimates exist for the
sources estimated here, the validity of these numbers may be open to question. For
instance, information on unpaved roads in not required to be reported to the U.S.
DOT, and examination of the reported data indicates that this information may not
receive a great deal of attention by the states that do report it. As an example,
Indiana evenly distributes the unpaved local functional class mileage between the
ADTV categories. Since there are four categories, 25% of the total unpaved road
mileage is allocated to each. Surrounding states such as Illinois, Ohio and Kentucky
show that the majority of the mileage for unpaved roads in those states is found in the
lowest two ADTV categories, especially for rural roads. However, since this is the
"data" reported by Indiana, and is supposed to represent their "best" estimate, the
reported information was utilized in developing emissions estimates for inclusion in the
Region 5 results.
Emission factors may also lead to potential problems. The PM10 emission
factor utilized to develop the emissions estimates for paved road resuspension
includes no term for reducing emissions as the consequence of precipitation, even
though the emission factor for unpaved roads does. It seems likely that precipitation
would act similarly to water sweeping of paved roads and that such a term should be
included in determining emissions from this source. The effect of such a term would
be to reduce paved road resuspension PM10 emissions, especially in those regions
with significant precipitation levels.
The emission factor for paved road resuspension is also singularly dependent
upon the silt loading term included in equation 2. According to EPA guidance
(Cowherd et al.. 1988), the silt term can be evaluated if you know the ADTV.
However, the ADTV is linearly related to the VMT. Since VMT is also the source
extent for this source, you can perfectly predict the PM10 emissions by establishing a
relationship between VMT and emissions.
Two items are important to remember in considering whether or not to
determine regional PM10 emissions estimates as part of a trends estimation. First, the
numbers developed for fugitive PM10 emissions will have a great deal of error
associated with them due both to the state of the science for emission factors as well
as the procedures utilized to evaluate the source extent. Second, given the first item,
these are the only tools available for determining this type estimate and any other
person or agency trying to produce similar estimates will be forced to utilize identical
or very similar techniques. Thus, EPA's decision as to whether or not to develop
these regional PM10 emissions estimates for trends estimation should give due
consideration to these two items.
34
-------
RECOMMENDATIONS
The following recommendations for development of fugitive dust PM10 regional
emissions trends estimates are based on the results of this study in conjunction with
conversations with EPA Technical Support Division (TSD) personnel.
1. For the 1989 Trends Report, EPA should include emission estimates for
unpaved roads and wind erosion at the national and regional levels. For
the other five categories (agricultural tilling, construction, mining and
quarrying, paved roads, and burning), emission estimates at the national
level only should be included. It is further recommended that for this
single annual assessment, 1985 should be used for the year of
assessment. 1985 is suggested as the year for presentation, since there
is a critical mass of information available (i.e. the 1985 NAPAP emission
inventory) for several of the categories suggested for inclusion.
Of the above categories that should be included at the national level for
the 1989 Trends reports, data on wind erosion and agricultural tilling
should be developed from the 1985 NAPAP emissions inventory.
Appropriate qualifying language detailing the methodology utilized to
develop these estimates should be placed in the Trends reports, so that
methodological differences leading to potential changes in future year
estimates can be readily explained.
Although the 1985 NAPAP emission inventory included emissions
estimates from unpaved roads, it is recommended that the emissions
estimates developed in this report for 1985 be included in the 1989
Trends reports. The methodology used to develop the estimates for this
source presented in this report utilized the current AP-42 emission factor,
rather than the emission factor developed as part of the NAPAP effort.
In addition, this methodology accounts for the distribution of vehicle types
(i.e. passenger vehicles, trucks, etc.) traveling on unpaved roads,
variability in vehicle weights and number of wheels, and in vehicle speed
on various unpaved road types. The emission factor developed as part
of the 1985 emission inventory effort has never been validated for other
than passenger cars, and the emissions estimation methodology
assumed single values for vehicle speed, vehicle weight and number of
wheels. We feel that the methodology utilized in this study is consistent
with EPA guidance on development of emissions estimates from this
source (i.e. utilization of AP-42 emission factors) and that the allocation
of emissions based on the vehicle mix is a better approach than the
single weight/speed/tires approach utilized in the NAPAP estimation
method.
35
-------
2. Starting with the 1990 Trends reports, emission estimates for the seven
categories included in Recommendation 1 (see above) should be
included at the regional level. These estimates should represent a
minimum of a five year period so that a trend can be developed. As
indicated in Recommendation 1, two categories (unpaved roads and wind
erosion) should be presented on a regional level in the 1989 Trends
reports since regional estimates for these two categories are feasible and
are thought to provide a reasonable, initial characterization of regional
emissions from these sources.
3. In order to provide the regional emissions estimates proposed in
Recommendation 2 (see above), development work would be required in
order to develop methods to produce regional emissions estimates for
those sources for which emissions estimates were not developed in this
feasibility study. These sources include: wind erosion, agricultural tilling,
construction, mining and quarrying and burning. Methodology changes
may be required between the production of national numbers for
inclusion in the 1989 Trends report (see Recommendation 1) and
subsequent regional trends emissions estimates. For instance, the 1985
NAPAP method of producing wind erosion estimates would be difficult to
use for production of regional emissions estimates, thus development of
a methodology to produce regional emissions estimates would be
required prior to the 1990 Trends reports publication.
4. Some specific recommendations can be made with regard to the
emissions estimation methods for several of the sources that should be
included in future regional emissions estimates. These include:
A. Addition of a "dry days" term to the emission factor
equation for paved roads. The same term as is utilized for
the unpaved road emission factor should be used. Addition
of this term would help account for washing of paved roads
due to precipitation and, although not a perfect solution to
this problem, would be defensible, since a similar term is
already utilized to reduce emissions from unpaved roads.
B. Utilization of the single-valued construction emission factor
with a PM10 multiplier to develop regional construction
emissions estimates. The new "gap filling" emission
factors, although representing a better method of estimating
site-specific emissions estimates, would pose large
problems in terms of estimating regional emissions
estimates due to the nature of the source extent information
required.
36
-------
C. Wind erosion emissions estimates may require utilization of
the Universal Soil Loss equation, with a multiplier to
evaluate the vertical flux of PM10 material. Other potential
methods of estimating emissions from this source should be
considered, but the Universal Soil Loss equation may
represent the best available method at this time.
D. A single-valued emission factor may be required for
developing emissions estimates for mining and quarrying
operations coupled with a multiplier to ascertain the PM10
fraction of these emissions. Utilization of the AP-42 PM10
emission factors may present considerable difficulties in
developing adequate source extent information.
5. Regarding the other five source categories considered in this report
(feedlots, landfills, unpaved parking lots, unpaved airstrips, and storage
piles), these categories should not be included for development at this
time in future Trends reports. The main reasons for not including these
sources at this time are inadequate or non-existent information on the
source extent (i.e. activity level) at the regional level or the expected
level of emissions is a great deal lower than the seven categories that
are recommended for development, or both.
37
-------
REFERENCES
Barnard, W.R., "Development of County-Level Wind Erosion and Unpaved Road
Alkaline Emissions Estimates for the 1985 NAPAP Emissions Inventory," U.S.
EPA Rept. No. EPA-600/7-90-005, January, 1990.
Barnard, W.R., D.F. Gatz, and G.J. Stensland, "Evaluation of Potential Improvements
in the Estimation of Unpaved Road Fugitive Emission Inventories," Paper 87-
58.1, Proceedings of the 80th Annual Meeting of the Air Pollution Control
Association, New York, NY, 1987.
Cowherd, C., G.E. Muleski and J.S. Kinsey, "Control of Open Fugitive Dust Sources,"
U.S. EPA Rept. No. EPA-450/3-88-008, September 1988.
Cowherd, C. and C. Guenther, "Development of a Methodology and Emission
Inventory for Fugitive Dust for the Regional Air Pollution Study," U.S. EPA Rept.
No. EPA-450/3-76-003, January, 1976.
Dickson, R.J., W.R. Oliver, and S.R. Tate, "Evaluation of Emissions from Selected
Uninventoried Sources in the State of California," Final Report for ARB Contract
No. A5-147-32, April, 1988.
Evans, J.S. and D.W. Cooper, "An Inventory of Particulate Emissions from Open
Sources," Journal Air Pollution Control Association, Vol. 30, #12, pp. 1298-
1303, December 1980.
Gillette, D., R.N. Clayton, T.K. Mayeda, M.L Jackson, and K. Sridhar, "Tropospheric
Aerosols from Some Major Dust Storms of the Southwestern United States," J^
Geophys. Res.. 87, 9003-9015, 1978.
Hagen, L.J., "Wind Erosion Prediction System: Concepts to Meet User Needs," USDA
Contribution Number 89-543-J, 1989.
Heisler, S.L, "Interim Emissions Inventory for Regional Air Quality Studies," Electric
Power Research Institute Report EPRI EA-6070, November 1988.
Midwest Research Institute, "Gap Filling PM10 Emission Factors for Selected Open
Area Dust Sources," U.S. EPA Rept. No. EPA-450/4-88-003, February, 1988.
38
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-450/4-91-Q05b
. TITLE AND SUBTITLE
Feasibility Of Including Fugitive PM-10 Emissions
Estimates In The EPA Emissions Trends Report
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
September 1990
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
William Barnard and Patricia Carlson
8. PERFORMING ORGANIZATION REPORT NO.
}. PERFORMING ORGANIZATION NAME AND ADDRESS
Pechan & Associates, Inc.
Durham, NC 27707
1O. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-4400
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
OAR, OAQPS, TSD, EIB (MD-14)
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
EPA Project Officer: E. L. Martinez
t6. ABSTRACT
This report describes the results of Part 2 of a two part study. Part 2 was
to evaluate the feasibility of developing regional emission trends for PM-10. Part
1 was to evaluate the feasibility of developing VOC emission trends, on a regional
and temporal basis. These studies are part of the effort underway to improve the
national emission trends. Part 1 is presented in a separate report. The categories
evaluated for the feasibility of developing regional emissions estimates were-
unpaved roads, paved roads, wind erosion, agricultural tilling, construction
activities, feedlots, burning, landfills, mining and quarrying unpaved parking
lots, unpaved airstrips and storage piles.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Emission Trends
Trends
PM-10
Regional
Temporal
Unpaved Roads
Paved Roads
Wind Frnsirm
Agricultural Tilling
Construction Activities
Feedlots
Burning
Landfills
Mining and Quarrying
Unpaved Parking Lots
Unpaved Airstrips
18. DISTRIBUTION STATEMENT $tOraQ6
Unlimited
19. SECURITY CLASS (This Report I
Unclassified
21. NO. OF PAGES
44
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (R«v. 4-77) PREVIOUS EDITION is OBSOLETE
------- |